METTL3 stabilizes DDX17 mRNA via IGF2BP2-mediated m6A modification to suppress endometrial cancer progression.

To clarify DDX17 expression in EC, we first used GEPIA2 and UALCAN databases. The results showed that compared to normal tissues, the DDX17 level was significantly downregulated in EC tissues (Fig.  1 A–B). DDX17 was found to be significantly downregulated in EC tissues compared to adjacent normal tissues (mRNA: p  < 0.001; protein: p  < 0.001; Fig.  1 C–D), with similar suppression observed across EC cell lines (HEC-1A, HEC-1B, Ishikawa vs. hEEC; Fig.  1 F–G). Critically, clinicopathological analysis of our cohort (n = 80) revealed balanced distributions between DDX17-low and DDX17-high groups across key prognostic parameters including FIGO stage, molecular subtype, treatment modality, and age (Table  1 ). Despite this clinical equipoise, Kaplan–Meier analysis demonstrated significantly worse overall survival in DDX17-low patients ( p  < 0.001; Fig.  1 E), a finding further validated by multivariate Cox regression where low DDX17 expression emerged as an independent predictor of poor prognosis (HR = 2.72, 95% CI = 1.28–5.78, p  = 0.009) after adjusting for stage (HR = 3.05, p  = 0.0009), p53abnormal (p53abn) subtype (HR = 2.41, p  = 0.016), and other covariates (Table  2 ). Fig. 1 DDX17 downregulation in EC correlates with poor clinical prognosis. A The expression of DDX17 in 33 kinds of tumors compared to normal tissues from GEPIA2 website ( http://gepia2.cancer-pku.cn/#general ). B The expression of DDX17 in Uterine Corpus Endometrial Carcinoma (UCEC) tumors (n = 546) and normal samples (n = 35) from UALCAN website ( https://ualcan.path.uab.edu/cgi-bin/ualcan-res.pl ). C RT-qPCR analysis of DDX17 mRNA expression in paired EC samples and adjacent normal samples (n = 80 pairs). Paired t-test. D Western blotting analysis of DDX17 protein expression in paired EC samples and adjacent normal samples (n = 80 pairs). (E) Overall survival comparison between patients with high (n = 40) versus low (n = 40) DDX17 expression. Cohorts were stratified using median expression cutoff. F RT-qPCR analysis of DDX17 mRNA expression in human endometrial epithelial cells (hEEC) and human EC cells (HEC-1A, HEC-1B, and Ishikawa). N = 3 independent experiments. G Western blotting analysis of DDX17 protein expression in hEEC, HEC-1A, HEC-1B, and Ishikawa cells. N = 3 independent experiments. ** p  < 0.01, *** p  < 0.001 DDX17 downregulation in EC correlates with poor clinical prognosis. A The expression of DDX17 in 33 kinds of tumors compared to normal tissues from GEPIA2 website ( http://gepia2.cancer-pku.cn/#general ). B The expression of DDX17 in Uterine Corpus Endometrial Carcinoma (UCEC) tumors (n = 546) and normal samples (n = 35) from UALCAN website ( https://ualcan.path.uab.edu/cgi-bin/ualcan-res.pl ). C RT-qPCR analysis of DDX17 mRNA expression in paired EC samples and adjacent normal samples (n = 80 pairs). Paired t-test. D Western blotting analysis of DDX17 protein expression in paired EC samples and adjacent normal samples (n = 80 pairs). (E) Overall survival comparison between patients with high (n = 40) versus low (n = 40) DDX17 expression. Cohorts were stratified using median expression cutoff. F RT-qPCR analysis of DDX17 mRNA expression in human endometrial epithelial cells (hEEC) and human EC cells (HEC-1A, HEC-1B, and Ishikawa). N = 3 independent experiments. G Western blotting analysis of DDX17 protein expression in hEEC, HEC-1A, HEC-1B, and Ishikawa cells. N = 3 independent experiments. ** p  < 0.01, *** p  < 0.001 Considering HEC-1A and HEC-1B cells had relatively lower DDX17 expression than Ishikawa cells, they were used for further study. We constructed EC cells with stable DDX17 overexpression using DDX17 overexpression lentivirus. As western blotting revealed, OE-DDX17 dramatically upregulated DDX17 protein level in EC cells (Fig.  2 A). Then, as colony formation assays, wound healing assays, and Transwell assays revealed, DDX17 upregulation inhibited EC cell proliferation, migration, and invasion (Fig.  2 B–D). Western blotting analysis revealed that DDX17 upregulation significantly increased E-cadherin protein expression while downregulating N-cadherin and Vimentin protein expression (Fig.  2 E). Subsequently, we investigated the potential mechanism. As western blotting demonstrated, DDX17 upregulation efficiently suppressed PI3K and AKT phosphorylation in EC cells (Fig.  2 F). Moreover, in vivo experiments showed that DDX17 overexpression resulted in a considerable decrease in tumor volume and weight (F i g.  2 G–I). Simultaneously, immunofluorescence and immunohistochemical staining results demonstrated that DDX17 overexpression significantly reduced the number of Ki67- and p-AKT-positive cells in tumor tissues (Fig.  2 J–K). Fig. 2 DDX17 overexpression inhibits EC tumorigenesis in vivo and in vitro. A HEC-1A and HEC-1B cells were transfected with DDX17 overexpression vector and empty vector, and the overexpression efficiency was detected by western blotting. B Colony formation assays evaluated the proliferative capability of EC cells in vector and DDX17 overexpression groups. C Wound healing assays evaluated EC cell migration. D Transwell assays evaluated EC cell invasion. E Western blotting analysis of E-cadherin, N-cadherin, and Vimentin protein levels in EC cells. F Western blotting analysis of p-PI3K, PI3K, p-AKT, and AKT protein levels in EC cells. G Representative images of tumors derived from control and DDX17-treated HEC-1A cells. H Tumor volume. I Tumor weight. J Immunofluorescence staining for Ki67 and immunohistochemical staining for p-AKT. K Quantification of Ki67- and p-AKT-positive cells in tumors. N = 6 mice/group. *** p  < 0.001 DDX17 overexpression inhibits EC tumorigenesis in vivo and in vitro. A HEC-1A and HEC-1B cells were transfected with DDX17 overexpression vector and empty vector, and the overexpression efficiency was detected by western blotting. B Colony formation assays evaluated the proliferative capability of EC cells in vector and DDX17 overexpression groups. C Wound healing assays evaluated EC cell migration. D Transwell assays evaluated EC cell invasion. E Western blotting analysis of E-cadherin, N-cadherin, and Vimentin protein levels in EC cells. F Western blotting analysis of p-PI3K, PI3K, p-AKT, and AKT protein levels in EC cells. G Representative images of tumors derived from control and DDX17-treated HEC-1A cells. H Tumor volume. I Tumor weight. J Immunofluorescence staining for Ki67 and immunohistochemical staining for p-AKT. K Quantification of Ki67- and p-AKT-positive cells in tumors. N = 6 mice/group. *** p  < 0.001 As the SRAMP database reveals, there are many putative m6A modification sites for DDX17, and two m6A sites with a high confidence threshold are identified (Fig.  3 A). RT-qPCR analysis showed that METTL14 knockdown by shRNA#1/2 had no obvious impact on DDX17 expression, while METTL3 knockdown (shRNA#1/2) effectively reduced DDX17 expression in EC cells (Fig.  3 B–C and Fig. S1 A–B). Meanwhile, METTL3 overexpression resulted in DDX17 upregulation in EC cells (Fig.  3 D). As shown by RIP-PCR assays, DDX17 was enriched in the anti-METTL3 group (Fig.  3 E). We also found that METTL3 knockdown dramatically reduced DDX17 protein expression in EC cells (Fig.  3 F and Fig. S1 C). The results of MeRIP-PCR assays revealed that the m6A modification level was significantly reduced in EC cells following METTL3 silencing (Fig.  3 G and Fig. S1 D). We also used actinomycin D to detect DDX17 mRNA stability and found that METTL3 knockdown suppressed DDX17 mRNA stability (Fig.  3 H and Fig. S1 E). Afterward, we overexpressed METTL3 and found that METTL3 overexpression led to significantly increased protein level of DDX17 (F i g.  3 I). Additionally, MeRIP-PCR assays confirmed elevated m6A modification on DDX17 mRNA in HEC-1A cells transfected with METTL3 (Fig.  3 J). It was also found that METTL3 overexpression enhanced the stability of DDX17 (Fig.  3 K). Moreover, the GEPIA database shows that METTL3 is downregulated in EC tissues, and METTL3 expression is positively correlated with DDX17 expression in EC (Fig.  3 L–M). This positive correlation was also verified in our cohort (Fig.  3 N). Collectively, the METTL3-mediated m6A modification regulates DDX17 expression in EC. Fig. 3 METTL3-mediated m6A modification stabilizes DDX17 in EC cells. A SRAMP database of m6A modification on DDX17 mRNA sequence and the high confidence m6A site location (16930 and 19980) on DDX17 mRNA sequence. B RT-qPCR analysis of METTL14 and DDX17 mRNA expression in HEC-1A and HEC-1B cells after transfection with sh-NC or sh-METTL14#1. C RT-qPCR analysis of METTL3 and DDX17 mRNA expression in EC cells after sh-NC or sh-METTL3#1 transfection. D RT-qPCR analysis of METTL3 and DDX17 mRNA expression in EC cells in response to vector or METTL3 transfection. E RIP-PCR assays were performed using anti-METTL3 antibody, and the enrichment of DDX17 was measured by RT-qPCR. F Western blotting analysis of DDX17 protein expression in EC cells after transfection with sh-NC or sh-METTL3#1. G The m6A modification level of DDX17 in EC cells transfected with sh-NC or sh-METTL3#1 was assessed by MeRIP-PCR assays. H Determination of DDX17 mRNA stability in EC cells transfected with sh-NC or sh-METTL3#1 using actinomycin D. I Western blotting analysis of DDX17 protein expression in HEC-1A cells with vector or METTL3 transfection. J MeRIP-PCR assays analysis of the m6A modification level of DDX17 in HEC-1A cells transfected with vector or METTL3. K The mRNA stability of DDX17 in HEC-1A cells with or without METTL3 overexpression was determined using actinomycin D. L GEPIA database showing the mRNA expression of METTL3 in UCEC tissues (n = 174) or normal tissues (n = 91). M The correlation between METTL3 expression and DDX17 expression predicted by GEPIA database. N Pearson’s correlation analysis of METTL3 mRNA expression and DDX17 mRNA expression in 80 EC tissues, showing relative expression (2 −ΔΔCt ) derived from qPCR. Dashed line indicates linear regression (r 2  = 0.4545, p  < 0.001). N = 3 independent experiments. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 METTL3-mediated m6A modification stabilizes DDX17 in EC cells. A SRAMP database of m6A modification on DDX17 mRNA sequence and the high confidence m6A site location (16930 and 19980) on DDX17 mRNA sequence. B RT-qPCR analysis of METTL14 and DDX17 mRNA expression in HEC-1A and HEC-1B cells after transfection with sh-NC or sh-METTL14#1. C RT-qPCR analysis of METTL3 and DDX17 mRNA expression in EC cells after sh-NC or sh-METTL3#1 transfection. D RT-qPCR analysis of METTL3 and DDX17 mRNA expression in EC cells in response to vector or METTL3 transfection. E RIP-PCR assays were performed using anti-METTL3 antibody, and the enrichment of DDX17 was measured by RT-qPCR. F Western blotting analysis of DDX17 protein expression in EC cells after transfection with sh-NC or sh-METTL3#1. G The m6A modification level of DDX17 in EC cells transfected with sh-NC or sh-METTL3#1 was assessed by MeRIP-PCR assays. H Determination of DDX17 mRNA stability in EC cells transfected with sh-NC or sh-METTL3#1 using actinomycin D. I Western blotting analysis of DDX17 protein expression in HEC-1A cells with vector or METTL3 transfection. J MeRIP-PCR assays analysis of the m6A modification level of DDX17 in HEC-1A cells transfected with vector or METTL3. K The mRNA stability of DDX17 in HEC-1A cells with or without METTL3 overexpression was determined using actinomycin D. L GEPIA database showing the mRNA expression of METTL3 in UCEC tissues (n = 174) or normal tissues (n = 91). M The correlation between METTL3 expression and DDX17 expression predicted by GEPIA database. N Pearson’s correlation analysis of METTL3 mRNA expression and DDX17 mRNA expression in 80 EC tissues, showing relative expression (2 −ΔΔCt ) derived from qPCR. Dashed line indicates linear regression (r 2  = 0.4545, p  < 0.001). N = 3 independent experiments. * p  < 0.05, ** p  < 0.01, *** p  < 0.001 As western blotting revealed, METTL3 overexpression increased the protein level of DDX17 and reduced phosphorylated PI3K and phosphorylated AKT protein levels in EC cells, whereas DDX17 knockdown (shRNA#1/2) reversed these alterations in EC cells (Fig.  4 A). Then, the data from functional experiments revealed that METTL3 overexpression significantly suppressed EC cell proliferation, migration, and invasion, which were restrained by DDX17 knockdown (Fig.  4 B–D). Western blotting analysis demonstrated that the METTL3 overexpression-mediated increase in E-cadherin protein level and reduction in N-cadherin and Vimentin protein levels were counteracted by DDX17 knockdown (Fig.  4 E). Fig. 4 DDX17 knockdown reverses the antitumor effect of METTL3 overexpression in EC cells. A Western blotting analysis of DDX17, p-PI3K, PI3K, p-AKT, and AKT protein expression in HEC-1A and HEC-1B cells of vector + sh-NC, METTL3, METTL3 + sh-DDX17#1, and METTL3 + sh-DDX17#2 groups. B Colony formation assays evaluated proliferation. C Wound healing assays evaluated cell migration. D Transwell assays evaluated cell invasion. E Western blotting analysis of E-cadherin, N-cadherin, and Vimentin protein expression. N = 3 independent experiments. *** p  < 0.001 vs. vector + sh-NC group; ### p  < 0.001 vs. METTL3 group DDX17 knockdown reverses the antitumor effect of METTL3 overexpression in EC cells. A Western blotting analysis of DDX17, p-PI3K, PI3K, p-AKT, and AKT protein expression in HEC-1A and HEC-1B cells of vector + sh-NC, METTL3, METTL3 + sh-DDX17#1, and METTL3 + sh-DDX17#2 groups. B Colony formation assays evaluated proliferation. C Wound healing assays evaluated cell migration. D Transwell assays evaluated cell invasion. E Western blotting analysis of E-cadherin, N-cadherin, and Vimentin protein expression. N = 3 independent experiments. *** p  < 0.001 vs. vector + sh-NC group; ### p  < 0.001 vs. METTL3 group Western blotting analysis confirmed that DDX17 knockdown effectively reduced DDX17 expression while increasing PI3K/AKT phosphorylation in Ishikawa cells (Fig.  5 A). Functional assays demonstrated that DDX17 knockdown enhanced proliferation, migration, and invasion capabilities, which were significantly attenuated by either METTL3 overexpression or PI3K inhibition (LY294002 treatment) (Fig.  5 B–D). Consistently, both METTL3 overexpression and LY294002 counteracted DDX17 knockdown-induced EMT progression, as evidenced by restored epithelial marker expression (Fig.  5 E). Fig. 5 METTL3 overexpression offset the oncogenic effect of DDX17 knockdown in EC cells. A Western blotting analysis of DDX17, p-PI3K, PI3K, p-AKT, and AKT protein expression in Ishikawa cells of vector + sh-NC, sh-DDX17#1, METTL3 + sh-DDX17#1, and LY294002 + shDDX17#1 groups. B Colony formation assays evaluated cell proliferation. C Wound healing assays evaluated cell migration. D Transwell assays evaluated cell invasion. Western blotting analysis of E-cadherin, N-cadherin, and Vimentin protein expression. N = 3 independent experiments. *** p  < 0.001 vs. vector + sh-NC group; ## p  < 0.01, ### p  < 0.001 vs. sh-DDX17#1 group METTL3 overexpression offset the oncogenic effect of DDX17 knockdown in EC cells. A Western blotting analysis of DDX17, p-PI3K, PI3K, p-AKT, and AKT protein expression in Ishikawa cells of vector + sh-NC, sh-DDX17#1, METTL3 + sh-DDX17#1, and LY294002 + shDDX17#1 groups. B Colony formation assays evaluated cell proliferation. C Wound healing assays evaluated cell migration. D Transwell assays evaluated cell invasion. Western blotting analysis of E-cadherin, N-cadherin, and Vimentin protein expression. N = 3 independent experiments. *** p  < 0.001 vs. vector + sh-NC group; ## p  < 0.01, ### p  < 0.001 vs. sh-DDX17#1 group To determine whether the functional effect of METTL3 on EC cell phenotypes are specifically mediated by DDX17 or involve other METTL3 targets, we examined the expression of known METTL3-regulated genes, NLRC5 and BIRC5. Western blotting analysis showed that METTL3 overexpression significantly increased the protein levels of NLRC5 and BIRC5 in both HEC-1A and HEC-1B cells (Fig. S2 A–B). Notably, DDX17 knockdown did not affect the upregulation of NLRC5 and BIRC5 induced by METTL3 overexpression, indicating that DDX17 knockdown specifically impairs the METTL3-DDX17 axis without affecting other METTL3 targets. Finally, we detected whether METTL3-mediated m6A modification stabilizes DDX17 through an IGF2BP-dependent mechanism. The IGF2BP family consists of three members including IGF2BP1, IGF2BP2, and IGF2BP3. As RT-qPCR revealed, only silencing IGF2BP2 significantly affected the mRNA level of DDX17 in EC cells (Fig.  6 A and Fig. S3 A). Then, western blotting analysis demonstrated that IGF2BP2 knockdown reduced DDX17 protein level, and IGF2BP2 overexpression elevated DDX17 protein level in EC cells (Fig.  6 B and Fig. S3 B). The actinomycin D experiment demonstrated that IGF2BP2 knockdown significantly inhibited DDX17 mRNA stability in EC cells (Fig.  6 C and Fig. S3 C). RIP-PCR assays demonstrated enrichment of DDX17 in the anti-IGF2BP2 group. This enrichment was dramatically reduced by METTL3 knockdown (Fig.  6 D–E). Moreover, western blotting demonstrated that DDX17 protein expression was markedly reduced following IGF2BP2 silencing, which was rescued by METTL3 overexpression (Fig.  6 F). Crucially, simultaneous knockdown of METTL3 and IGF2BP2 did not further reduce DDX17 expression or stability compared to individual knockdowns (Fig. S3 D–E), demonstrating their functional interaction within a linear regulatory pathway. Fig. 6 METTL3-mediated m6A modification stabilizes DDX17 via IGF2BP2 recognition. A RT-qPCR analysis of DDX17 mRNA expression in EC cells transfected with sh-NC, sh-IGF2BP1#1, sh-IGF2BP2#1, or sh-IGF2BP3#1. B Western blotting assessed IGF2BP2 and DDX17 protein expression in EC cells transfected with sh-NC, sh-IGF2BP2#1, vector, or IGF2BP2. C The stability of DDX17 mRNA following IGF2BP2 knockdown or overexpression was detected using actinomycin D. D RIP-PCR assays were performed using anti-IGF2BP2 antibody, and the enrichment of DDX17 was measured by RT-qPCR. E RIP-PCR assays were performed using anti-IGF2BP2 antibody after METTL3 knockdown, and the enrichment of DDX17 was measured by RT-qPCR. F Western blotting measured DDX17 protein expression in EC cells in sh-NC, sh-IGF2BP2#1, and sh-IGF2BP2#1 + METTL3 groups. N = 3 independent experiments. *** p  < 0.001 vs. sh-NC or Anti-IgG group; ### p  < 0.001 vs. sh-IGF2BP2#1 group METTL3-mediated m6A modification stabilizes DDX17 via IGF2BP2 recognition. A RT-qPCR analysis of DDX17 mRNA expression in EC cells transfected with sh-NC, sh-IGF2BP1#1, sh-IGF2BP2#1, or sh-IGF2BP3#1. B Western blotting assessed IGF2BP2 and DDX17 protein expression in EC cells transfected with sh-NC, sh-IGF2BP2#1, vector, or IGF2BP2. C The stability of DDX17 mRNA following IGF2BP2 knockdown or overexpression was detected using actinomycin D. D RIP-PCR assays were performed using anti-IGF2BP2 antibody, and the enrichment of DDX17 was measured by RT-qPCR. E RIP-PCR assays were performed using anti-IGF2BP2 antibody after METTL3 knockdown, and the enrichment of DDX17 was measured by RT-qPCR. F Western blotting measured DDX17 protein expression in EC cells in sh-NC, sh-IGF2BP2#1, and sh-IGF2BP2#1 + METTL3 groups. N = 3 independent experiments. *** p  < 0.001 vs. sh-NC or Anti-IgG group; ### p  < 0.001 vs. sh-IGF2BP2#1 group

Eighty paired tumors and adjacent normal tissue samples (≥ 1 cm from margins) were collected from 80 patients with histologically confirmed EC. Samples were snap-frozen in liquid nitrogen, with all adjacent tissues confirmed cancer-free by histopathology. Exclusion criteria encompassed non-endometrioid histology, endometriosis, coexisting cancers, BMI > 40, or recent hormone therapy (24 months pre-surgery). Longitudinal follow-up spanned January 2018-December 2024. Patients provided informed consent at First Affiliated Hospital of Bengbu Medical University, with clinicopathological characteristics summarized in Tables 1 and 2 , and ethical approval was granted by the Ethics Committee of First Affiliated Hospital of Bengbu Medical University (Approval number: 2025102). Table 1 Clinicopathological characteristics of EC patients (n = 80) Parameter Overall (n = 80) DDX17-Low (n = 40) DDX17-High (n = 40) p value Age (years) 62.5 ± 8.7 63.2 ± 9.1 61.8 ± 8.3 0.7724 FIGO stage I 48 (60%) 22 (55%) 26 (65%) 0.6592 II 18 (22.5%) 11 (27.5%) 7 (17.5%) 0.5636 III-IV 14 (17.5%) 7 (17.5%) 7 (17.5%) 0.2447 Molecular subtype POLE-mutated 8 (10%) 3 (7.5%) 5 (12.5%) 0.7575 MMR-deficient 22 (27.5%) 13 (32.5%) 9 (22.5%) 0.6056 NSMP 32 (40%) 14 (35%) 18 (45%) 0.6592 p53 abnormal 18 (22.5%) 10 (25%) 8 (20%) 0.8664 Treatment Surgery only 42 (52.5%) 20 (50%) 22 (55%) 0.9046 Surgery + CT/RT 38 (47.5%) 20 (50%) 18 (45%) 0.9046 CT  Chemotherapy, RT  Radiotherapy; p -values from χ 2 test for categorical variables and t-test for age Table 2 Multivariate cox regression analysis of prognostic factors Variable HR (95% CI) p value DDX17 (Low vs. High) 2.72 (1.28–5.78) *0.009 Age (≥ 60 vs. < 60) 1.36 (0.71–2.60) 0.352 Stage (III-IV vs. I-II) 3.05 (1.58–5.89) *0.0009 Molecular subtype p53abnormal vs. NSMP 2.41 (1.18–4.91) *0.016 MMRd vs. NSMP 1.07 (0.52–2.20) 0.854 POLE vs. NSMP 0.45 (0.11–1.87) 0.269 Treatment (CT/RT vs. Surgery only) 0.87 (0.47–1.63) 0.660 HR  Hazard ratio; CI  Confidence interval; CT/RT  Chemotherapy/radiotherapy Clinicopathological characteristics of EC patients (n = 80) CT  Chemotherapy, RT  Radiotherapy; p -values from χ 2 test for categorical variables and t-test for age Multivariate cox regression analysis of prognostic factors HR  Hazard ratio; CI  Confidence interval; CT/RT  Chemotherapy/radiotherapy Total RNA was extracted using TRIzol reagent (Beyotime, Shanghai, China), with cDNA synthesis performed via reverse transcription using a SYBR Premix Ex Taq Kit (Takara, Japan). Quantitative PCR amplification was conducted with SYBR Green Master Mix (Thermo Fisher, USA) using cDNA templates, with GAPDH serving as endogenous control. Primer sequences (designed and synthesized by Sangon Biotech, Shanghai) are detailed in Table  3 . Relative mRNA expression was calculated using the 2 −ΔΔCt method. Table 3 Sequences of primers used for reverse transcription-quantitative PCR Gene (human) Sequence (5′→3’) DDX17 forward GATGAAGAAGGAGGCGGAGAAGAG DDX17 reverse GGGGTCAGTATCAGCCGCTTTCAG METTL14 forward GAACACAGAGCTTAAATCCCCA METTL14 reverse TGTCAGCTAAACCTACATCCCTG METTL3 forward CCAGCACAGCTTCAGCAGTTCC METTL3 reverse GCGTGGAGATGGCAAGACAGATG BIRC5 forward CCACTGAGAACGAGCCAGACTT BIRC5 reverse GTATTACAGGCGTAAGCCACCG GAPDH forward GGTCTCCTCTGACTTCAACA GAPDH reverse GTGAGGGTCTCTCTCTTCCT Sequences of primers used for reverse transcription-quantitative PCR To extract total protein lysates, tissues or cells were lysed in RIPA buffer that contained 1 mM phenylmethylsulfonyl fluoride (Thermo Fisher). Then, the lysates were quantified using a BCA protein assay kit (Beyotime). Protein samples (40 μg) were separated by 10% SDS-PAGE and transferred to PVDF membranes. After being blocked with 5% skimmed milk for 1 h, the membranes were incubated at 4 °C overnight with primary antibodies targeting DDX17 (ab180190, 1:1000; Abcam), E-cadherin (ab40772, 1:1000; Abcam), N-cadherin (ab76011, 1:5000; Abcam), Vimentin (ab92547, 1:1000; Abcam), p-PI3K (#13,857, 1:1000; Cell Signaling Technology), PI3K (#4263, 1:1000; Cell Signaling Technology), p-AKT (#4060, 1:1000; Cell Signaling Technology), AKT (#9272, 1:1000; Cell Signaling Technology), IGF2BP2 (ab124930, 1:2000; Abcam), NLRC5 (DF13672, 1:1000; Affinity), and GAPDH (ab8245, 1:1000; Abcam). On the next day, the membranes were washed thrice with PBS containing Tween 20 and incubated with corresponding horseradish peroxidase (HRP)-linked secondary antibodies at room temperature for 1 h. After incubation, the membranes were washed. Blots were developed using an enhanced chemiluminescence detection reagent (Thermo Fisher) and analyzed by ImageJ software. Human endometrial epithelial cells (hEEC) and EC cell lines (HEC-1A, HEC-1B, Ishikawa) (Procell, Wuhan, China) were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Thermo Fisher) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. All cells were maintained at 37 °C in a humidified 5% CO₂ incubator, with media refreshed every 48 h. Cell authentication was performed via STR profiling, and cultures were routinely tested for mycoplasma contamination. Lentiviral overexpression vectors (OE-DDX17/IGF2BP2/METTL3 in pLVX-puro) and dual independent shRNA knockdown constructs (sh-DDX17#1/#2, sh-IGF2BP2#1/#2, sh-METTL3#1/#2 in pLKO.1-puro) with matched negative controls (vector, sh-NC) were designed and synthesized by OBIO (Shanghai). HEC-1A and HEC-1B cells were transduced at 60% confluency using Lipofectamine 2000 (Thermo Fisher) with 8 µg/mL polybrene, followed by 5 µg/mL puromycin selection for 72 h to establish stable lines. Knockdown efficiency was validated via RT-qPCR and western blotting for both shRNA constructs per target gene prior to functional assays. EC cells were seeded into 6-well plates (500 cells/well) and cultured for 14 days. Then, the cells were fixed in 4% paraformaldehyde (Beyotime) at room temperature for 15 min, followed by staining with 2% crystal violet (Sigma-Aldrich, USA) for 20 min. The plates were photographed, and colonies (> 50 cells) were counted using ImageJ software. EC cells were seeded into 24-well plates (2.5 × 10 5 cells/well) and cultured until they reached 80–90% confluence. Then, a 10-μL micropipette tip was employed to create a scratch through the well center. Subsequently, serum-free medium was used to replace the original medium. At 0 and 24 h, wound images were obtained under a microscope. The formula, [(wound width at 0 h–width at 24 h)/width at 0 h] × 100%, was used to calculate the relative migration rate. Invasion assays were performed using Matrigel-precoated Transwell chambers (Corning; 0.8 μm pores), wherein the upper chamber received 200 μL serum-free DMEM containing 1 × 10 5 EC cells, while the lower chamber contained DMEM supplemented with 10% FBS. After 24-h incubation at 37 °C, migrated cells on the lower surface were fixed with 4% paraformaldehyde, stained with 0.05% crystal violet for 30 min, and quantified by counting cells in five random non-overlapping fields per membrane under a light microscope (Olympus). Female BALB/c nude mice (4–6 weeks of age, 20 ± 2 g; Charles River Laboratories, Beijing, China) were housed under standard conditions (22–26 °C, 45–65% of humidity, a 12 h light/dark cycle), with free access to food and water. Mice were randomly assigned into vector and DDX17 groups (n = 6/group). DDX17 overexpressing or control HEC-1A cells (5 × 10 6 ) were subcutaneously implanted into the mouse left flank. Tumor growth was measured every 4 days. On day 25 after injection, mice were sacrificed by cervical dislocation under deep anesthesia (sodium pentobarbital), and the tumors were photographed, weighed, and processed for paraffin sections. All animal-related procedures were conducted in accordance with the ethical principles outlined in the NIH Guide for Care and Use of the Laboratory Animals and were approved by the Institutional Animal Ethics Committee of First Affiliated Hospital of Bengbu Medical University (Approval number: 2025412). Tumor tissues were fixed in 4% paraformaldehyde (24 h), paraffin-embedded, and sectioned coronally at 5 µm onto glass slides. Sections underwent deparaffinization, antigen retrieval via microwave heating (98 °C × 10 min in 10 mM citrate buffer, pH 6.0), and rehydration through graded ethanol series. Endogenous peroxidase activity was blocked with 3% H 2 O 2 (10 min); non-specific binding was blocked using 3% BSA/PBS (1 h), followed by incubation with anti-p-AKT primary antibody (AF0016, 1:200; Affinity) at 4 °C overnight. After PBS washes, sections were incubated with HRP-conjugated secondary antibody (30 min, room temperature), developed with DAB substrate (Beyotime), and counterstained with hematoxylin. Slides were mounted and imaged using a brightfield microscope (Olympus). Tumor tissues were fixed in 4% paraformaldehyde for 24 h, embedded in paraffin, and coronally cut into 5-μm-thick sections. Afterward, the tumor sections were deparaffinized in xylene, rehydrated, and washed with deionized water. After being blocked with 5% bovine serum albumin at room temperature for 45 min, the sections were incubated with primary antibodies against Ki67 (ab16667, 1:200; Abcam) at 4 °C overnight. On the next day, the sections were rinsed with PBS containing 0.5% Tween-20 and then incubated with corresponding fluorescence-conjugated secondary antibodies at room temperature for 1 h. Nuclei were stained with DAPI (Beyotime) for 3 min. Images were obtained using a fluorescence microscope (Olympus). RNA immunoprecipitation was performed using the Magna RIP Kit (Sigma-Aldrich) with EC cells (1 × 10 7 ) lysed in RIP buffer containing protease/RNase inhibitors for 5 min on ice. Lysates were cleared by centrifugation (12,000 g, 10 min, 4 °C) and incubated overnight at 4 °C with Protein A/G magnetic beads pre-conjugated with 5 μg anti-METTL3 (ab195325, Abcam), anti-IGF2BP2 (ab117809, Abcam), or control IgG. Bead-bound complexes were washed six times with RIP wash buffer, followed by proteinase K digestion (20 μg/mL, 37 °C, 45 min); immunoprecipitated RNA was extracted with TRIzol and analyzed via RT-qPCR. m6A-specific RNA immunoprecipitation was performed using the Magna MeRIP™ m6A Kit (Sigma-Aldrich), beginning with TRIzol-based total RNA extraction from EC cells. RNA purity/concentration was verified by NanoDrop™ (Thermo Fisher), followed by fragmentation of 30 μg RNA to ~ 100 nt fragments using Magnesium RNA Fragmentation Module (Thermo Fisher) with subsequent addition of Stop Solution. Fragmented RNA underwent pre-clearing with Protein A/G magnetic beads for 1 h at 4 °C to reduce non-specific binding, then immunoprecipitated overnight at 4 °C with rotation using 5 μg anti-m6A antibody (ABE572, Millipore) or IgG control in IP buffer. Beads were washed six times with high-salt buffer, followed by proteinase K digestion (2 μg/μL, 45 min, 45 °C). Immunoprecipitated RNA was extracted via acid phenol–chloroform purification with glycogen carrier, ethanol-precipitated, and analyzed by RT-qPCR to quantify m6A enrichment. EC cells were seeded in 6-well plates at 1 × 10 5 cells/well and transfected according to experimental designs. Then, cells were treated with 10 μg/mL actinomycin D (Sigma-Aldrich, dissolved in DMSO) for 0, 3, and 6 h; at each time point, total RNA was extracted using TRIzol, followed by DNase I treatment (RNase-free DNase Set, Qiagen) to eliminate genomic DNA contamination. DDX17 mRNA expression was quantified via RT-qPCR using GAPDH as the endogenous reference control. Patients were stratified into high- and low-DDX17 expression groups based on median expression levels, with Kaplan–Meier survival analysis and log-rank testing used to compare overall survival differences. Pearson’s correlation analysis used relative expression values (2 −ΔΔCt ) derived from qPCR of 80 EC tissue pairs. Expression levels were normalized to GAPDH and calculated using the ΔΔCt method as described in RT-qPCR section. All statistical analyses were performed using GraphPad Prism 9, with quantitative data expressed as mean ± SD of ≥ 3 independent experiments. Paired Student’s t-tests compared tumor versus adjacent normal tissues, unpaired t-tests analyzed two independent groups, and one-way ANOVA with Tukey’s post hoc test assessed multi-group comparisons, with statistical significance defined as p  < 0.05.

DDX17, a DEAD-box RNA helicase, exhibits context-dependent oncogenic roles—promoting hepatocellular carcinoma metastasis via PXN-AS1 isoform generation [ 16 ] and colorectal cancer progression through the miR-149-3p/CYBRD1 axis [ 17 ]—while suppressing ovarian cancer metastasis via PDIA4 upregulation [ 19 ]. In this study, DDX17 restoration suppresses EC proliferation, migration, invasion, EMT, and xenograft growth by inactivating PI3K/AKT signaling. Mechanistically, METTL3-mediated m6A modification stabilizes DDX17 mRNA via IGF2BP2 recognition, establishing DDX17 as both a tumor suppressor in EC. Our findings reveal that DDX17 is significantly downregulated in EC according to TCGA/GEPIA2 analyses. Consistent with database analyses and our functional data, low DDX17 expression correlates with significantly worse overall survival in our patient cohort and serves as an independent prognostic marker. Importantly, the integration of DDX17 expression levels with established TCGA molecular subtypes (e.g., p53abn, which exhibited a significant hazard ratio in our analysis) might refine prognostic stratification beyond current classifications. Future studies with larger cohorts comprehensively characterized for all TCGA markers (POLE, MMR, and p53 status) are needed to validate DDX17’s additive prognostic value and explore potential subtype-specific roles of this pathway. m6A modification critically regulates tumor proliferation, invasion, and EMT [ 20 , 21 ], with established roles in EC progression [ 12 ]. METTL3, the principal m6A methyltransferase, dynamically controls cancer-relevant pathways [ 22 ]; in EC, reduced METTL3 expression enables immune evasion, while its knockdown promotes tumorigenesis via PAPP/IGFBP4 signaling [ 13 , 23 ]. Our data demonstrate METTL3-mediated m6A modification stabilizes DDX17 mRNA through IGF2BP2 recognition, suppressing EC malignancies via PI3K/AKT inactivation—effects abolished by DDX17 knockdown; conversely, METTL3 overexpression rescued DDX17 deficiency-induced oncogenicity. Though METTL3 regulates multiple targets (e.g., immunomodulatory NLRC5), the specific phenotypic reversal by DDX17 knockdown without affecting NLRC5/BIRC5 confirms DDX17 as METTL3’s key downstream effector for PI3K/AKT inactivation, underscoring its unique mechanistic role in EC suppression. The PI3K/AKT signaling pathway—a highly conserved regulator of cell growth, protein synthesis, and metabolism in eukaryotes—is aberrantly activated in multiple malignancies including endometrial carcinoma [ 24 – 26 ]. Critically, PI3K/AKT hyperactivation drives EC progression, while its pharmacological inhibition exerts potent antitumor effects [ 27 , 28 ]. Our findings demonstrate that DDX17 overexpression suppresses PI3K/AKT phosphorylation, consequently inhibiting EC proliferation, migration, and invasion. Notably, the observation that PI3K inhibition (LY294002) phenocopies METTL3 restoration in rescuing DDX17 deficiency-induced oncogenicity highlights the PI3K/AKT pathway as the key downstream effector and underscores the therapeutic potential of targeting this axis in EC. This is particularly relevant for the p53abn molecular subtype, which is characterized by frequent PI3K/AKT pathway alterations [ 29 ] and exhibits a strong association with poor prognosis in our study. Furthermore, while MMR status primarily informs immunotherapy response, our findings suggest that assessment of the METTL3/DDX17/PI3K axis status, potentially integrated with molecular subtyping, could identify a subset of patients (especially p53abn) who might benefit from PI3K/AKT pathway inhibitors (e.g., alpelisib) alone or in combination. IGF2BP2, a well-characterized m6A reader [ 30 ], drives cancer progression through RNA stabilization mechanisms—demonstrated in EC where it activates JAK/STAT signaling via circCHD7 modification [ 31 ] and promotes colorectal/ovarian cancers through METTL3 collaboration [ 32 , 33 ]. Paradoxically, our study reveals IGF2BP2’s tumor-suppressive role in EC via METTL3-mediated m6A modification recruiting IGF2BP2 to stabilize DDX17 transcripts. This IGF2BP2 functional switching specifically inactivates PI3K/AKT signaling, contrasting its oncogenic roles in other malignancies. This context-dependent function might be influenced by the specific molecular landscape of EC. It would be of significant interest to investigate whether this METTL3-IGF2BP2-DDX17 axis exhibits differential activity or clinical impact across the distinct TCGA molecular subtypes (POLEmut, MMRd, NSMP, p53abn). This study has limitations. First, the results were obtained from a limited clinical cohort size. Future validation in larger, independent cohorts with comprehensive molecular characterization (including full TCGA classification: POLE, MMR, p53 status) is essential to confirm the prognostic value of DDX17 and delineate potential subtype-specific associations of the METTL3-IGF2BP2-DDX17-PI3K/AKT axis. Second, while xenografts provide valuable insights, they cannot fully recapitulate human tumor evolution, metastatic niches, or systemic interactions, particularly the complex interplay with the immune microenvironment which is highly relevant for MMRd tumors. To better mimic human tumor biology, future studies could integrate humanized mouse models or organoid-xenograft co-cultures alongside multi-omics profiling. This study establishes that METTL3-mediated m6A modification recruits IGF2BP2 to stabilize DDX17 mRNA, forming a linear regulatory axis that suppresses endometrial carcinogenesis through PI3K/AKT pathway inactivation (Fig.  7 ). Our data suggest that DDX17 loss contributes to poor prognosis and that targeting this axis, particularly in PI3K/AKT-hyperactivated subtypes like p53abn, holds therapeutic promise. Future research integrating comprehensive molecular subtyping will be crucial to fully realize the diagnostic and therapeutic potential of this pathway. Fig. 7 A schematic diagram depicting the METTL3-m6A-IGF2BP2 regulatory axis stabilizing DDX17 mRNA to suppress EC progression A schematic diagram depicting the METTL3-m6A-IGF2BP2 regulatory axis stabilizing DDX17 mRNA to suppress EC progression

Endometrial cancer (EC), a common gynecologic malignancy, ranks as the sixth most prevalent cancer among women worldwide. In recent years, increasing estrogen use, rising obesity prevalence, and dietary alterations have contributed to growing EC incidence [ 1 ]. Annually, approximately 420,368 women are diagnosed with EC globally, resulting in 97,923 deaths, and incidence rates continue to rise [ 2 ]. Current standard treatments include surgery, chemotherapy, radiation, and hormone therapy. While early-stage EC has satisfactory outcomes, limited therapeutic options impair prognosis in advanced EC [ 3 ]. Therefore, investigating effective molecular targets for EC is imperative. N6-methyladenosine (m6A) modification represents a significant epigenetic and epitranscriptomic marker in eukaryotes. This modification occurs in DNA, mRNA, tRNA, and noncoding RNAs [ 4 ]. Among eukaryotic RNA modifications, m6A constitutes the most prevalent internal modification. It is dynamically regulated by methyltransferases (m6A “writers”), demethylases (m6A “erasers”), and reader proteins that recognize m6A marks [ 5 , 6 ]. In this process, METTL3–structurally supported by METTL14 and stabilized by WTAP–catalyzes m6A deposition. Conversely, ALKBH5 and FTO mediate m6A removal [ 7 ]. m6A modification critically regulates mRNA stability, translation efficiency, and alternative splicing, influencing diverse biological processes including tumor progression [ 8 , 9 ]. For example, m6A regulates EPR29 expression and regulates EC cell proliferation [ 10 ]. PEG10 promotes EC progression via IGF2BP1-meidated m6A modification [ 11 ]. The m6A modification of EGR1/PTEN, mediated by WTAP/IGF2BP3, regulates EC cell malignancy [ 12 ]. METTL3, an enzyme that functions as a critical methyltransferase in the m6A modification, has been found to facilitate immunosurveillance in EC [ 13 ]. DEAD-box RNA helicases constitute a versatile group of ATP-dependent enzymes crucial for RNA processing, including alternative splicing, modification, and ribosome assembly [ 14 ]. Among these, DDX17 exhibits the highest evolutionary conservation and performs essential eukaryotic functions encompassing transcriptional regulation, miRNA processing, and pre-mRNA splicing [ 15 ]. Recent studies demonstrate context-dependent roles: in hepatocellular carcinoma, DDX17 promotes carcinogenic PXN-AS1 isoform production [ 16 ]; in colorectal cancer, it influences metastasis via the miR-149-3p/CYBRD1 axis [ 17 ]; and in lung adenocarcinoma, DDX17 regulates proliferation/apoptosis-related gene expression and splicing [ 18 ]. However, DDX17’s role and molecular mechanisms in EC progression remain unexplored. This study investigates DDX17 and m6A modification in EC pathogenesis. We first established differential DDX17 expression in EC versus normal tissues, followed by functional validation experiments. Subsequently, we elucidated DDX17’s regulatory mechanisms, revealing novel connections to m6A signaling. Our findings provide insights into developing targeted EC therapies.

Below is the link to the electronic supplementary material. Supplementary file1 (tif 884 kb) Supplementary file2 (tif 487 kb) Supplementary file3 (tif 1143 kb) Supplementary file1 (tif 884 kb) Supplementary file2 (tif 487 kb) Supplementary file3 (tif 1143 kb)