Cells were established previously from primary PDAC, including PANC10.05, BxPC3, MIA PaCa-2, PANC3.11 and SW1990. The STR reports of PANC10.05, BxPC3, MIA PaCa-2 and SW1990 cell lines are available (identified by the BIOWING company). PANC3.11 cells were established by Elizabeth M. Jaffee from Johns Hopkins University and have been published [35, 36]. Mycoplasma contamination testing was performed for these cell lines.

Primary cancer samples were collected at the Chinese PLA General Hospital, which were comprised of intraductal papillary mucinous neoplasm (IPMN, 43 cases), mucinous cystic neoplasm (MCN, 21 cases), and PDAC (298 cases). All patients were not subjected to chemo-radiotherapy before surgery. Tumors were classified according to tumor-node-metastasis (TNM) staging (AJCC 2019). The procedures were performed in accordance with the Declaration of Helsinki. The study was approved by the Institutional Review Board of the Chinese PLA General Hospital, and informed consent was obtained from all patients (Approval No. 20090701-015, 20160204-26 and S2025-791-01).

DNA was prepared with phenol-chloroform and modified following previous approaches [37]. Total RNA preparation and cDNA synthesis followed the manufacturer’s instructions (#K1691, Thermo Scientific, USA). Demethylating reagent, 5-aza (2 μmol/L), was utilized to induce re-expression of epigenetically silenced genes (#A3656, Sigma-Aldrich, USA). The primers are listed in Table S1.

Antibodies used for IHC are shown in Table S2. Paraffin samples were stained according to a previous description. The staining score was evaluated following the German semiquantitative scoring criteria [38].

The coding region of L3MBTL4 (referenced as NM_173464) was inserted into pCDH-CMV-MCS-puro plasmid. L3MBTL4 expressing or empty vectors were transfected into HEK293 cells using Lipofectamine 3000 Reagent (Invitrogen, USA). Lentiviral supernatant was then added to the RPMI-1640 medium (31800089, Gibco, USA) containing 10% fetal bovine serum (900-108, GEMINI, USA) and 1% penicillin/streptomycin solution (BL505A, Biosharp, China). L3MBTL4 expressing cells were selected by puromycin treatment at a concentration of 2.5μg/mL (BxPC3) and 1.5 μg/mL (MIA PaCa-2) for 3 days. Single-cell clones were screened by limited dilution in 96 well plates and validated using western blotting.

Cells were seeded in 96- well plates for 2 × 10³ BxPC3 and 2 × 10³ MIA PaCa-2 cells per well for MTT assay (#KGT5251, KeyGEN Biotech, China). Six-well plates were used to perform colony formation, with 1 × 10³ cells/well for 12 days. Cell cycle and apoptosis were analyzed by propidium iodide (PI) staining (#KGA512, KeyGEN Biotech, China) and Annexin V-FITC/PI Apoptosis Detection Kit (#KGA108, KeyGEN Biotech, China) with FACScan flow cytometer.

The efficiency of NU7441 was tested by MTT assay. For 50% inhibitory concentration (IC₅₀) analysis, PDAC cells were seeded in 96-well plates at 2 × 10³ per well, and the cell viability was evaluated by the OD value after treatment for 48 h. The sensitivity of PDAC cells to NU7441 (#HY-11006, MCE, USA) was assessed by colony formation assay. Cells were seeded with 2000 each well and treated with 0.01 μM cisplatin (#S1166, Selleck, USA) and 0.5 μM NU7441 for 48 h. Thereafter, the medium was changed to RPMI-1640 containing 10% fetal bovine serum and 1% penicillin/streptomycin solution, and the cells were grown for 10 days.

RNAiMax reagent was utilized for L3MBTL4 knockdown. The procedure was performed according to instructions (#13778150, Invitrogen, USA). The sequences of the siRNAs are shown in Table S1 (JTS Scientific, China). Antibodies used for western blotting and IP are listed in Table S2. Mass spectrometry technique was used to analyze the components of the interacting complex.

HR and NHEJ reporter assays were followed as in a previous study [29]. The U2OS cells were identified by the Procell Life Science company, and the STR report of U2OS cells is available. L3MBTL4 silenced or re-expressed BxPC3 and MIA PaCa-2 cells were grown in 6-well plates and transfected with 1 μg of pCVL Traffic Light Reporter 1.1 (Sce target) Ef1a Puro plasmids and 1 μg of I-SceI plasmids. The cells were harvested for FACS analysis after growing for 48 h, and the data were analyzed using FlowJo (BD Biosciences, USA). The experiments were performed in triplicate.

To evaluate the impact of L3MBTL4 on DDR, the alkaline comet electrophoresis method was employed [29, 39]. DNA damaged PDAC cell models was induced by cisplatin at 1 μM and 0.5 μM for BxPC3 and MIA PaCa-2 cells, respectively. The tail moments were quantified by the Comet Score (version2.0, Tritek Corp.). Each group included 100 cells.

A PDAC cell xenograft mouse model was employed for evaluating the influence of L3MBTL4 methylation in DNA-PK inhibitor sensitivity. L3MBTL4 silenced or re-expressed MIA PaCa-2 cells (4 × 10⁶ cells) were injected subcutaneously into four-week-old female BALB/c nude mice (purchased from SPF company, Beijing, China). The tumor volume was calculated following a previous study [28]. When the tumor volume reached about 100 mm³, the mice were randomly divided into four groups (each group with 6 mice), including the control group, cisplatin group (2 mg/kg), NU7441 group (10 mg/kg), and cisplatin combined with NU7441 group (2 mg/kg of cisplatin and 10 mg/kg of NU7441). Cisplatin and NU7441 were administered twice a week intraperitoneally for two weeks. Tumors were measured for 21 days, 3 days each time. Thereafter, the mice were euthanized using cervical dislocation by experienced personnel in strict compliance with the Guide for the Care and Use of Laboratory Animals. The animal experiments were performed according to the guidelines approved by the Animal Ethics Committee of the Chinese PLA General Hospital (Approval NO. 2021-X17-44).

GraphPad Prism 8.0 software (GraphPad Software Inc., CA, USA) was utilized for statistical analysis. The associations between L3MBTL4 methylation and clinicopathological factors were analyzed by chi-square test. Kaplan–Meier plots and the log-rank test were used to estimate the OS. Univariate and multivariate Cox regression analysis were applied to evaluate the influencing factors of survival time. The difference of the OD values was analyzed by repeated measures ANOVA and Bonferroni’s multiple comparisons test in L3MBTL4 unexpressed and re-expressed PDAC cells. The difference between two groups was assessed by the Student’s t-test. P < 0.05 was regarded as statistically significant.

To explore the possibility of epigenetic regulation of L3MBTL4, the RNA expression and DNA methylation data of PDAC were obtained from GTEx and The Cancer Genome Atlas (TCGA) databases (http://xena.ucsc.edu/). The levels of L3MBTL4 mRNA were significantly lower in PDAC samples than in normal pancreatic tissue (P < 0.001, Figure S1A and Supplementary material). The expression of L3MBTL4 was reversely associated with the methylation of CpG sites in the promoter region (cg14693194, cg26120251, cg08007465, cg02731042 and cg12924825, all P < 0.05, Figure S1B–1C and Supplementary material). These findings hint that L3MBTL4 is possibly regulated by DNA methylation.

To validate the promoter region methylation regulating L3MBTL4 expression, PDAC cells were detected with RT-PCR and methylation specific PCR (MSP). L3MBTL4 was not expressed in BxPC3 and MIA PaCa-2 cells, while reduced expression was exhibited in PANC10.05, PANC3.11 and SW1990 cells (Figure 1A). Complete methylation was observed in BxPC3 and MIA PaCa-2 cells, and partial methylation emerged in PANC10.05, PANC3.11 and SW1990 cells (Figure 1B). The reverse association of L3MBTL4 expression and methylation was shown in PDAC cells. The result was verified by inducing its expression with 5-aza (Figure 1A). The methylation density and the efficiency of MSP primers were verified by bisulfite sequencing (BSSQ) in MIA PaCa-2 and BxPC3 cells (Figure 1C).

Thereafter, the methylation status of L3MBTL4 was detected in IPMN, MCN and PDAC and it was methylated in 16.3% (7/43) of IPMN, 19.0% (4/21) of MCN, and 28.2% (84/298) of PDAC (Figure 1D). L3MBTL4 methylation was significantly associated with tumor differentiation (P < 0.001) and tumor size (P < 0.01, Table 1), indicating that L3MBTL4 methylation is increased with tumor progression. L3MBTL4 methylation was associated with poor OS (P < 0.05, Figure 1E) and was an independent prognostic factor for poor OS in 107 cases of follow up data available patients (P < 0.05, Table 2).

Then, the expression and methylation status of L3MBTL4 were analyzed. L3MBTL4 was stained in both nucleus and cytoplasm. The staining score of L3MBTL4 was lower in cancer tissue samples compared to adjacent normal tissues (Figure 1F and 1G, P < 0.001). The high level of L3MBTL4 was revealed in 23 cases of PDAC tissue samples, among which 8 cases were methylated (34.8%). Reduced level of L3MBTL4 was observed in 25 cases of PDAC, among which 19 cases were methylated (76.0%). Methylation of L3MBTL4 was associated with its low-level expression, implying L3MBTL4 methylation regulates its expression (Figure 1H, P < 0.01).

The viability of PDAC cells with or without L3MBTL4 expression was evaluated using MTT assay. The OD values were 0.92 ± 0.03 vs. 0.61 ± 0.02 and 0.48 ± 0.02 vs. 0.35 ± 0.01 in BxPC3 and MIA PaCa-2 cells before and after restoration of L3MBTL4 expression for 96 hours, respectively (both P < 0.001, Figure 2A). The OD value was reduced by L3MBTL4, indicating an inhibitory effect on cell proliferation.

Colony formation assay was performed to evaluate the role of L3MBTL4 in PDAC cell growth. The clone numbers were 566.0 ± 41.4 vs. 243.7 ± 19.4 and 234.3 ± 10.3 vs. 79.7 ± 14.4 in L3MBTL4 unexpressed and re-expressed BxPC3 and MIA PaCa-2 cells, respectively (both P < 0.001, Figure 2B). L3MBTL4 reduced PDAC cell colon number, demonstrating the inhibitory role of L3MBTL4 in PDAC cells.

The effect of L3MBTL4 on the cell cycle was assessed with flow cytometry. The percentages of G0/G1 phase cells were 34.2 ± 0.3% vs. 48.0 ± 1.2% (P < 0.001) and 40.6 ± 1.3% vs. 52.1 ± 3.1% (P < 0.01) in L3MBTL4 unexpressed and re-expressed BxPC3 cells and MIA PaCa-2 cells, respectively. The percentages were 51.4 ± 1.6% vs. 34.4 ± 3.2% (P < 0.01) and 36.9 ± 1.4% vs. 26.0 ± 4.7% (P < 0.05) in the S phase, respectively. The ratios of G2/M phase cells were 14.5 ± 1.7% vs. 17.6 ± 1.9% and 22.4 ± 1.2% vs. 21.9 ± 1.7% before and after re-expression of L3MBTL4 in BxPC3 cells and MIA PaCa-2 cells, respectively (Figure 2C). These results indicate that L3MBTL4 re-expression induces G1/S arrest in PDAC cells. The levels of G1/S checkpoint related proteins were analyzed. The levels of Cyclin A2, Cyclin D1, Cyclin E1, and CDK2 proteins were reduced by L3MBTL4 expression, validating the effect of L3MBTL4 on cell cycle (Figure 2D).

The role of L3MBTL4 in apoptosis was measured by flow cytometry. The percentages of apoptotic cells were as follows: 5.5 ± 0.8% vs. 11.5 ± 1.2% (P < 0.01) and 3.8 ± 0.2% vs. 6.3 ± 0.3% (P < 0.001) in L3MBTL4 unexpressed and re-expressed BxPC3 and MIA PaCa-2 cells, respectively (Figure 2E). These results suggest that L3MBTL4 induces PDAC cell apoptosis. Representative apoptotic markers were detected by western blot. L3MBTL4 reduces the level of Bcl2 and increases the levels of Cleaved-caspase3 and Bax (Figure 2F), further verifying its role in inducing apoptosis.

To understand the mechanism of L3MBTL4 in PDAC, IP and mass spectrometry techniques were utilized. Ku70 was discovered with the highest score in the complex (Figure 3A). This finding was validated by reciprocal IP and western blot (Figure 3B). KEGG pathway analysis was performed to analyze the binding proteins, and NHEJ components were uncovered to be an important part (Figure 3C). Then, the interaction of L3MBTL4 and Ku70 was tested by AlphaFold3 (Figure 3D).

Ku70 is a key component of NHEJ signaling. L3MBTL1 and L3MBTL2, the family members of L3MBTL4, have been reported to participate in DDR [32, 33, 40]. DDR is a double-edged sword for carcinogenesis and cancer chemoradiotherapy. Next, we evaluated the impact of L3MBTL4 on DDR in PDAC cells. In L3MBTL4 silenced and re-expressed BxPC3 and MIA PaCa-2 cells, the IC₅₀ values of cisplatin were 1.8 ± 0.4 µM vs. 5.3 ± 0.6 µM (P < 0.01) and 7.2 ± 0.4 µM vs. 18.2 ± 2.6 µM (Figure 3E, P < 0.01), respectively. These results demonstrated that L3MBTL4 reduces the sensitivity of PDAC cells to cisplatin, reflecting the potential role of L3MBTL4 in DDR.

Thereafter, the effect of L3MBTL4 on DSB repair was evaluated via a comet assay. In L3MBTL4 silenced and re-expressed BxPC3 and MIA PaCa-2 cells, the tail moment was 83.0 ± 4.4 vs. 55.7 ± 3.1 (Figure 3F, P < 0.001) and 71.7 ± 7.6 vs. 30.3 ± 6.5 (P < 0.001), respectively. These results indicate that L3MBTL4 is involved in DNA DSB repair.

In mammalian cells, DSBs are the most severe DNA damage, as they may cause large chromosomal region loss, and DSBs are repaired predominantly by HR and NHEJ pathways. To better understand the mechanisms of L3MBTL4 in DNA DSB repair, HR and NHEJ efficiency were measured by HR and NHEJ reporter assays. HR efficiency was significantly reduced by L3MBTL4 knockdown in U2OS cells expressing DR-GFP, indicating that L3MBTL4 promotes HR signaling (Figure 3G). The efficiency of NHEJ was significantly increased after knockdown of L3MBTL4 in U2OS cells expressing EJ5-GFP, implying the inhibitory role of L3MBTL4 in NHEJ signaling (Figure 3G). Thereafter, the role of L3MBTL4 in DDR was validated in BxPC3 and MIA PaCa-2 cells (Figure 3H).

To further validate the roles of L3MBTL4 in HR and NHEJ signaling, the important molecules of these pathways were detected. Increased p-ATM and p-CHK2 levels were observed after restoration of L3MBTL4 expression in BxPC3 and MIA PaCa-2 cells, revealing the activating role of L3MBTL4 in ATM pathway (Figure 3I). The impact of L3MBTL4 on ATR/CHK1 signaling was also evaluated. However, no obvious changes were observed for the levels of p-ATR and p-CHK1 (Figure 3I). The inhibitory role of L3MBTL4 in NHEJ pathway was verified by detecting the level of p-DNA-PKcs and an increased level of p-DNA-PKcs was observed in L3MBTL4 unexpressed PDAC cells (Figure 3I).

To acquire more evidence that L3MBTL4 is involved in NHEJ, an MTT assay was utilized to evaluate the IC₅₀ of NU7441, a DNA-PK inhibitor, in PDAC cells. In BxPC3 and MIA PaCa-2 cells, the IC₅₀ values were 8.3 ± 1.6 vs. 22.3 ± 1.9 µM (P < 0.001) and 12.1 ± 1.8 vs. 26.9 ± 2.5 µM (P < 0.01), without the expression or restoration of L3MBTL4, respectively (Figure 4A). These results demonstrate that loss of L3MBTL4 expression sensitizes PDAC cells to DNA-PK inhibitors.

The low dose cisplatin induced DNA damage cell model was subsequently used to test the sensitivity of L3MBTL4 silenced PDAC cells to DNA-PK inhibitors. Under cisplatin and 0.5 μM NU7441 treatment, the normalized colony efficiency was 19.1 ± 2.5% vs. 35.6 ± 5.8% (P < 0.05) and 11.7 ± 1.4% vs. 38.4 ± 1.9% (P < 0.001) without expression or re-expression of L3MBTL4 in BxPC3 and MIA PaCa-2 cells (Figure 4B). These results suggested that loss of L3MBTL4 expression sensitized PDAC cells to NU7441. This effect was further validated by detecting the levels of γ-H2AX (Figure 4C).

To further verify that L3MBTL4 deficiency increased the sensitivity of PDAC to DNA-PK inhibitors, a MIA PaCa-2 cell xenograft mouse model was utilized. For the control group, without cisplatin or NU7441 treatment, the normalized tumor volume and weight were designated as 100%. In L3MBTL4 silenced and re-expressed MIA PaCa-2 cell xenografts, the normalized tumor volume was 84.2 ± 5.5% vs. 85.5 ± 5.0% for cisplatin group, 44.6 ± 4.5% vs. 60.4 ± 5.9% for NU7441 group (P < 0.001), and 17.9 ± 1.9% vs. 38.4 ± 5.5% for combined both cisplatin and NU7441 groups (P < 0.001, Figure 5A–5C). The normalized tumor weights were 82.4 ± 11.4% vs. 86.2 ± 13.8% for cisplatin group, 41.7 ± 7.4% vs. 59.5 ± 9.7% for NU7441 group (P < 0.01), and 19.7 ± 2.0% vs. 43.9 ± 4.2% for combined both cisplatin and NU7441 groups in L3MBTL4 silenced and re-expressed xenografts, respectively (P < 0.001, Figure 5D). These findings suggest that L3MBTL4 deficiency sensitizes PDAC cells to NU7441.

To further explore the effect of L3MBTL4 on DDR in vivo, MIA PaCa-2 cell xenograft tumors were stained by IHC. With cisplatin treatment, the levels of p-ATM and p-CHK2 were increased, while the level of p-DNA-PKcs was decreased in L3MBTL4 re-expressed xenograft tumors, showing the promoting role in ATM/CHK2 pathway and the inhibiting role in NHEJ signaling (Figure 5E). These results suggest that epigenetic silencing of L3MBTL4 sensitizes PDAC cells to DNA-PK inhibitors.

In summary, L3MBTL4 is a new component of DDR, and L3MBTL4 methylation is a potential diagnostic and therapeutic marker of PDAC.