Introduction

Adenocarcinoma of the pancreatic ducts is a particularly dangerous cancer, with a survival rate after five years of less than 10%. This low survival rate is primarily due to its propensity for early metastasis and resistance to treatment1,2. Recent findings underscore the critical role of non-coding RNAs, particularly the lncRNAs and microRNAs, in the regulation of pancreatic ductal adenocarcinoma (PDAC) development through complex molecular networks3,4.

The H19, a maternally imprinted gene on chromosome 11, has been associated with several malignancies by functioning as a competitive endogenous RNA (ceRNA) that sequesters tumor-suppressive miRNAs. H19 overexpression is linked to advanced tumor stages and poorer prognoses in various malignancies5. H19 up-regulates IGF2 and its own expression via miR-483-5p-mediated enhancer activation and chromatin loop formation, hence promoting hepatocellular carcinoma development6. H19 functions as a principal regulator in the pathogenesis of PDAC, where its dual role in TAM reprogramming—altering immune-suppressive cytokine networks (IL-6/IL-10/TGF-β) and generating oncogenic splicing variants—creates a conducive microenvironment for tumor advancement, with JAK/STAT inhibitors demonstrating therapeutic potential in H19-high PDAC models7. These findings underscore its carcinogenic potential in pancreatic ductal adenocarcinoma (PDAC).

Among miRNAs, miR-29c-5p has emerged as a significant tumor suppressor in gastrointestinal malignancies. Dysregulation of hsa-miR-29c-5p has been seen in breast cancer precursors, indicating its potential role in initiating early aberrant DNA methylation events in ER-positive breast cancer through the inhibition of DNMT3A8. The stress-induced downregulation of the tumor suppressor miR-139-5p leads to the overexpression of LPAR4 in pancreatic cancer cells, which activates fibronectin-enriched extracellular matrix (ECM) genes through the LPAR4/AKT/CREB pathway, thereby creating a self-sustaining tumor-initiating niche9. Concurrently, integrin-mediated ECM remodeling enhances stress tolerance and chemoresistance in adjacent cells10. Nonetheless, the upstream regulators of miR-29c-5p in pancreatic ductal adenocarcinoma, especially those involving ncRNA-mediated inhibition, are little defined.

The ATF2 gene an important member of the AP-1 transcription factor class, exhibits context-dependent roles in malignancy. ATF2 deficiency promotes colorectal cancer invasion by transcriptionally de-repressing the oncogenic transmembrane glycoprotein TROP2, which facilitates tumor cell de-adhesion and filopodia-mediated migration without triggering epithelial-mesenchymal transition11. The ATF2low/TROP2high signature serves as a prognosis predictor for assessing metastatic risk for those with colorectal cancer11.ATF2 transcriptionally controls cancer development by dimerizing with various partners to influence cell cycle (e.g., cyclins, RB1) and survival pathways (e.g., Bcl2), demonstrating context-dependent oncogenic or tumor-suppressive functions across malignancies12.ECM1 facilitates breast cancer invasion by cyclic vesicle transfer, which is integrin β2-dependent.)13, and in ovarian cancer, it regulates stromal remodeling and alternative splicing through secretory14. This study indicates that increased ECM1 levels are associated with worse clinical outcomes in PDAC patients, positioning it as a potential therapeutic target.

Recent research indicates interaction among H19, miR, and ATF2/ECM1. The miR-29 family is frequently functionally inactivated due to the overexpression of H19 in malignancies associated with smoking. The overexpression of its target genes, DLL4 (a Notch ligand) and VEGFA (a principal regulator of vascular permeability), is highly correlated with chemotherapy resistance and the probability of metastasis in patients15. , while The imbalance of ATF2/miR-545-5p is a core driver of the malignant progression of NSCLC16. Notwithstanding these advancements, the functional axis of H19/miR-29c-5p/ATF2/ECM1 in PDAC remains unexamined. Addressing this gap may reveal new processes behind PDAC metastasis and offer actionable targets for precision treatment.

Materials and methods

Patients and clinical samples

We collected four pairs of pancreatic cancer tumor tissues and corresponding adjacent non-cancerous tissue specimens from patients who underwent surgery at the Department of General Surgery, Digestive Disease Hospital, Affiliated Hospital of Zunyi Medical University within 2023 and 2024. Before sample collection, we secured clearance from the Institutional Review Board of the Affiliated Hospital of Zunyi Medical University. The gathered tissue specimens were immediately preserved in a -80 °C freezer for expression analysis. The pathology specimens were fixed, dehydrated, and embedded, with the wax blocks maintained at room temperature and shielded from light. The pancreatic ductal adenocarcinoma (PDAC) tissue samples included in this study must meet the following criteria: confirmed diagnosis of PDAC by postoperative pathological examination; complete clinicopathological data available (including patient age, TNM stage, differentiation grade, and survival follow-up data, etc.); samples obtained from surgical resection, fixed with formalin, embedded in paraffin, and quality-controlled to ensure tumor cell content > 50%, with nucleic acid quality meeting the requirements for subsequent experiments; patients must not have received neoadjuvant chemotherapy, radiotherapy, or other antitumor treatments before surgery; the study must be approved by the ethics committee, and informed consent must be obtained from the patients. Exclusion criteria include: pathological types other than PDAC (e.g., neuroendocrine tumors, acinar cell carcinoma, etc.); missing clinical data or samples unsuitable for analysis due to improper fixation or severe degradation; patients with other malignancies or significant organ dysfunction.

RNA extraction and PCR detection

Total RNA extraction from tissues and cells was conducted utilizing an RNA extraction kit (QIAGEN, Germany) in accordance with the manufacturer’s instructions. RT-qPCR was conducted using a Bio-Rad CFX96 real-time PCR system with SYBR qPCR Mix (Takara, Dalian, China). Lncrnas and mrnas were standardized to GAPDH levels, while mirnas were normalized to U6 levels.

Cell transfection

H19-specific shRNAs, together with miRNA mimics and inhibitors, were manufactured and procured from Ribobio (Guangzhou, China). All goods were evaluated against their respective negative controls. The mature shRNA mimic was transfected into pancreatic cancer cells with LipofectAMINE 2000 reagent (Invitrogen, USA).

RNA-FISH

The specified pancreatic cancer cells were inoculated on confocal culture plates. Cy3-labeled H19 probes were produced and acquired from Ribobio (Guangzhou, China). Cells were fixed using 4% paraformaldehyde, and fluorescence staining was conducted following the manufacturer’s guidelines. Images were captured and examined with a CAIS laser scanning confocal microscope manufactured by Nikon Instruments (Japan).

Transwell migration/invasion assay

The Transwell experiment was performed to assess cellular migration and invasion capacities utilizing 8 μm pore chambers (Corning, USA). Invasion experiments involved pre-coating the top chambers with 50–100 µg/well of Matrigel (BD Biosciences) diluted in serum-free media and polymerized at 37 °C for 2–4 h, whereas migration assays employed uncoated chambers. Cells were collected using 0.25% trypsin-EDTA, resuspended in serum-free media, and adjusted to a concentration of 1 × 10⁵–1 × 10⁶ cells/mL. The bottom chamber contained 500µL of complete medium with 10% FBS as a chemoattractant, while 100–200 µL of cell suspension was introduced into the top chamber, guaranteeing the absence of air bubbles. Following incubation at 37 °C with 5% CO₂ for 12–48 h (modified according to cell motility), non-migratory/invasive cells on the top membrane were eliminated using a cotton swab. Migrated or invaded cells were fixed with 4% paraformaldehyde for 30 min and subsequently stained with 0.1% crystal violet or DAPI for visualization. Images of 5 to 10 arbitrary fields per chamber were obtained using a light/fluorescence microscope (Nikon, Japan), and cell counts were analyzed with ImageJ software (NIH, USA). Data were presented as mean ± SD from triplicate experiments and evaluated using Student’s t-test or one-way ANOVA (GraphPad Prism v9.0), with p < 0.05 being statistically significant. Essential controls comprised serum-free media (negative control) and high-motility cell lines (positive control), while methodological concerns highlighted the homogeneity of Matrigel and the prevention of edge effects during enumeration.

Wound healing assay

The wound healing experiment was conducted to assess cell migration capability utilizing 24-well plates. Cells were inoculated at a density of 2 × 10^5 cells per well and grown in full media (DMEM supplemented with 10% FBS) until achieving 90–100% confluency. A consistent scratch was made in the monolayer with a sterile 200 µL pipette tip, followed by two washes with PBS to eliminate detached cells. Serum-free media (or medium with ≤ 2% FBS) and experimental treatments (e.g., 10 µM LY294002 or 20 ng/mL EGF) were administered to the corresponding wells. Plates were incubated at 37 °C in a 5% CO₂ atmosphere, and scratch closure was seen at 0 and 24 h using a phase-contrast microscope (Nikon, Japan). Images of three arbitrary fields per well were examined using ImageJ with the “MRI Wound Healing Tool” plugin to determine the percentage of wound closure:

Data from triplicate trials were presented as mean ± SD and statistically analyzed using one-way ANOVA with Tukey’s post hoc test (GraphPad Prism v9.0), with p < 0.05 deemed significant. Essential controls comprised mitomycin-C (5 µg/mL, 2-hour pretreatment) to suppress proliferation and serum-free medium as a baseline reference. Methodological rigor guaranteed consistent scratch width (about 1 mm) and the elimination of edge artifacts during measurement.

Dual luciferase reporter assay

PC cells were inoculated in 6-well plates at a density of 5 × 10^5 cells per well. Thereafter, miR-29c-5p mimics were introduced into pancreatic cancer cells. Following the removal of the supernatant, the lysates were employed to assess firefly luciferase and renilla luciferase activity. Absorbance was measured using a microplate reader following the manufacturer’s procedure (Guangzhou Ribobio Technology Co., LTD.). The relative luciferase activity was subsequently adjusted to the internal control of firefly luciferase.

Tumor formation experiment in nude mice

Tumor xenograft studies were conducted with six-week-old male BALB/c nude mice and PANC-1 cells. Initially, the relevant knockdown and overexpression lentiviruses from various groups were transfected into PANC-1 cells. Subsequently, the specified PANC-1 cells were proliferated and cultured, and 5 × 106 cells (200 µL) were subcutaneously administered into the pancreas of nude mice. The tumor xenografts’ volume was assessed weekly. The mice were euthanized, and photos were captured. Immunohistochemistry was employed to identify the expressions of ATF2 and ECM1 in the tumor xenograft tissues. The Animal Care Committee at Zunyi Medical University authorized the animal trials.According to the institutional ethics approval (approval number: zyfy-an-2023-0332), euthanasia was performed using isoflurane inhalation anesthesia (induction concentration 4%, maintenance concentration 2%) to ensure a humane and painless endpoint of the experiment, in accordance with the guidelines of the American Veterinary Medical Association. Tumor progression was monitored twice a week by calculating the volume with a digital caliper. Growth was strictly terminated when the maximum diameter reached 15 mm or the volume reached 2000 cubic millimeters to avoid unnecessary suffering.

Western blotting

The specified cells were rinsed three times with pre-chilled PBS. RIPA lysis buffer (Beyotime Biotechnology) was applied to the culture plates and incubated for 30 min to induce lysis. Thereafter, the protein lysates underwent centrifugation at 12,000 rpm, and the protein concentrations were quantified with the BCA Protein Assay Kit (Beyotime Biotechnology). The lysates were combined with 5× loading buffer and heated at 95 °C for 5 min. The total proteins were isolated by electrophoresis and subsequently transferred to PVDF membranes. The membranes were treated with the appropriate primary antibodies for 16 h. The membranes were treated in 5% TBST milk for 1 h, followed by a 2-hour incubation with specific HRP-conjugated secondary antibodies at room temperature. All membranes were imaged with the Bio-Rad imaging equipment.

TCGA database analysis

The ECM1, ATF2, H19, and hsa-mir-29c-5p models underwent a survival study. An examination of the differential expression of ECM1, ATF2, and H19 was conducted. Association studies were conducted for H19, ATF2, and ECM1, as well as for ECM1 and ATF2.

Statistical analysis

All statistical analyses in this study were performed using GraphPad Prism 8.0 software. Normally distributed continuous variables were expressed as mean ± standard deviation (x̄ ± s), and comparisons among multiple groups were conducted using one-way analysis of variance (ANOVA), while comparisons between two groups were performed using independent samples t-test. Continuous variables that deviated from a normal distribution were described by their median and interquartile range, and the group differences were assessed using the appropriate non-parametric tests (Mann-Whitney U test for two groups; Kruskal-Wallis H test for three or more groups). Comparisons of categorical variables were performed using χ² test or Fisher’s exact test. Kaplan-Meier methodology was employed to estimate survival probabilities, with group comparisons performed via the log-rank test.A Spearman correlation analysis was performed to measure the monotonic relationship between the variables.The discriminatory power of the candidate biomarkers was determined using receiver operating characteristic (ROC) analysis.

Results

The overexpression of H19 is associated with the progression and poor prognosis of pancreatic cancer, as well as its subcellular localization

H19 levels were considerably higher (P < 0.05) in pancreatic cancer tissues (n = 179) compared to normal tissues (n = 171) based on the TCGA database of tissue samples (Fig. 1B).In the research investigation of nine pancreatic lines of cancer cells via qRT-PCR, H19 expression levels in five cell lines (PANC-1, BxPC-3, etc.) were shown to surpass those of HPDE normal cells by over fourfold (p < 0.01, Fig. 1D). Patients exhibiting elevated H19 expression demonstrated a markedly reduced median overall survival (18.3 months compared to 32.6 months, HR = 2.17, p = 0.008), hence confirming its prognostic significance (Fig. 1C). The RNA FISH assay demonstrated that H19 was predominantly localized in the cytoplasm, corroborating the predictions made by LncLocator (Fig. 1A, E).

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Expression Dynamics and Subcellular Localization of H19 in Pancreatic Cancer. (A) Quantitative analysis of H19 subcellular localization based on LncLocator, indicating its predominant localization in the cytoplasm and exosomes (exosomes: 80%, nucleus: 10%, cytoplasm: 10%). (B) B.Box plot comparison of H19 expression levels in pancreatic cancer tissues (n = 179) and adjacent normal tissues (n = 171) from the TCGA database, *p < 0.05, **p < 0.01, ***p < 0.001. (C) Kaplan-Meier survival curve stratified by H19 expression in the TCGA database. High H19 levels are associated with poor prognosis (hazard ratio = 1.75, 95% confidence interval: 1.10–2.78; p = 0.018). (D) Relative expression of H19 in pancreatic cancer cell lines (PANC-1, MIA PaCa-2, Capan-2, BxPC-3, CFPAC) and normal HPDE cells. The expression of H19 is upregulated by more than 4-fold in PANC-1 and MIA PaCa-2 cells, *p < 0.05, **p < 0.01, ***p < 0.001. (E) Immunofluorescence staining confirmed the enrichment of H19 (red) in the cytoplasm of PANC-1 and MIA PaCa-2.

LncH19 promotes proliferation, migration, and invasion in pancreatic cancer cells

According to RT-qPCR tests, pancreatic cancer cell lines (HPAF-II, MIA PaCa-2) had considerably greater H19 expression than normal pancreas ductal epithelial cell lines (HPDE) (p < 0.01, Fig. 2A). Functional experiments demonstrated that the inhibition of H19 significantly inhibited cell proliferation (CCK-8 assay, decreased OD450 value, p < 0.01 vs. NC) and reduced the Transwell migration rate by approximately 60% (ImageJ quantification, p < 0.01). Conversely, the overexpression of H19 markedly enhanced proliferation (p < 0.05)(Fig. 2C) and migration (increased by 80%, p < 0.01) (Fig. 2B, D).

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

LncH19 Facilitates Proliferation, Migration, and Invasion of Pancreatic Cancer Cells. (A) The expression level of LncH19 in pancreatic cancer cells was determined by RT-qPCR *p < 0.05, **p < 0.01, ***p < 0.001. (B) In the Transwell assay, compared with the NC group, the migration/invasion capacity of the shLncH19 group was decreased, while that of the OE-LncH19 group was enhanced, *p < 0.05, **p < 0.01, ***p < 0.001. (C) The CCK-8 assay revealed that, compared with the NC group, cell viability was enhanced in the OE-LncH19 group, while proliferation was inhibited in the shLncH19-1/2 group, *p < 0.05, **p < 0.01, ***p < 0.001. (D) The scratch assay demonstrated that, in comparison with the NC group, the migration ability was strengthened in the OE-LncH19 group, while it was weakened in the shLncH19 group cells. Nuclei were counterstained with DAPI (blue). Scale bar: 20 μm, *p < 0.05, **p < 0.01, ***p < 0.001.

miR-29c-5p exhibits tumor suppressive effects in pancreatic cancer, correlates with patient prognosis, and exerts an antagonistic mechanism against H19

The transcript levels of miR-29c-5p in cancer cell lines was found to be 3.1 times lower compared to normal cell lines (Fig. 3A). Furthermore, the CCK8 experiment demonstrated that the overexpression of miR-29c-5p enhanced tumor cell proliferation (Fig. 3B). Wound-healing experiments demonstrated that the suppression of miR-29c-5p led to enhanced proliferation (Fig. 3C). Miranda program projected target associations, indicating that the 3’ UTR of H19 contains functional binding sites for miR-29c-5p. The RNAhybrid algorithm predicted and confirmed the complementary binding sites of miR-29c-5p and H19 at a mechanistic level. The dual-luciferase reporter enzyme experiment demonstrated that wild-type H19 substantially lowered luciferase activity (***P < 0.001 vs NC); however, this effect was negated upon mutation of the binding sites (H19-MUT) (Fig. 3D).Clinical survival study indicated that patients exhibiting elevated levels of miR-29c-5p in the TCGA database demonstrated significantly improved overall survival compared to those with diminished expression (HR = 0.56, p = 0.006, log-rank test), implying its prognostic significance (Fig. 3E).

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

The Tumor-Suppressive Function of miR-29c-5p in Pancreatic Cancer and Its Regulatory Mechanism with H19. (A) The expression levels of miR-29c-5p in diverse pancreatic cell lines: In contrast to normal pancreatic ductal epithelial cells (HPDE), the expression of miR-29c-5p was conspicuously downregulated in pancreatic cancer cell lines (HPAF-II, MIA PaCa-2), *p < 0.05, **p < 0.01, ***p < 0.001. (B) The regulation of cell proliferation by miR-29c-5p: Transfection with miR-29c-5p mimics markedly inhibited cell proliferation (with a decreased OD450 value), while the inhibitor promoted proliferation (data presented as mean ± standard deviation, n = 3, *p < 0.05, **p < 0.01, ***p < 0.001.) (C) Transwell migration assay: The migration rate of the miR-29c-5p mimics group was approximately 60% lower than that of the negative control (NC) group, while the migration rate of the inhibitor group increased by approximately 80% (quantitative analysis by ImageJ, *p < 0.05, **p < 0.01, ***p < 0.001.) (D) The targeted interaction between miR-29c-5p and H19, prediction of the binding sites and secondary structure alignment (via RNAhybrid algorithm); the dual-luciferase reporter gene assay affirmed that the wild-type H19 (H19-WT) could suppress luciferase activity, while the mutant binding site (H19-MUT) abolished this effect, *p < 0.05, **p < 0.01, ***p < 0.001.) (E) Clinical survival analysis: Data from the TCGA database indicated that pancreatic cancer patients with a high expression of miR-29c-5p had a significantly longer overall survival than those with a low expression (HR = 0.56, p = 0.006, log-rank test).

H19 promotes pancreatic cancer progression via ceRNA-mediated miR-29c-5p suppression

CCK-8 experiments showed that either H19 overexpression or miR-29c-5p inhibition significantly enhanced cell proliferation, with OD450 values at 48 h increasing by 30% to 45% compared to the NC group (p < 0.001). Conversely, H19 knockdown or miR-29c-5p overexpression notably reduced proliferation, as evidenced by a 40% to 55% decrease in OD450 values at 72 h (p < 0.01) (mean ± standard deviation, n = 3, Fig. 4A). Scratch assays further indicated that the scratch healing rate in the H19 overexpression group was 50% greater than that of the NC group at 24 h (p < 0.01), whereas the healing rate in the shH19 group diminished by approximately 65% (p < 0.001). Additionally, the miR-29c-5p inhibitor could partially mitigate the migration inhibition induced by H19 knockdown (Fig. 4B). Transwell invasion assays demonstrated that the quantity of invasive cells in the H19 overexpression cohort was 2.1 times greater than that of the NC cohort (p < 0.001), whereas the invasion capacity of the shH19 cohort diminished by 70% (p < 0.01), and the miR-29c-5p mimic was able to counteract the pro-invasive influence of H19 (Fig. 4C). In summary, H19 enhances the proliferation, migration, and invasion of pancreatic cancer cells by sequestering miR-29c-5p, hence alleviating its tumor-suppressive role.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

H19 promotes pancreatic cancer progression by ceRNA-mediated inhibition of miR-29c-5p. (A) CCK-8 assay shows the proliferation rate (OD450) of pancreatic cancer cells transfected with H19 overexpression plasmid, shH19, miR-29c-5p inhibitor or mimics at 0, 24, 48 and 72 h (mean ± SD, n = 3, *p < 0.05, **p < 0.01, ***p < 0.001.) (B) Quantitative migration ability of each group in the scratch assay (scratch healing rate at 0 and 24 h, *p < 0.05, **p < 0.01, ***p < 0.001.) (C) Transwell invasion/migration assay of each group (crystal violet staining, 5 × 10⁴ cells per well, *p < 0.05, **p < 0.01, ***p < 0.001.)

ECM1 and ATF2 are the downstream targets of miR-29c-5p and are regulated by H19

Analysis of the TCGA database indicates that the mRNA expression levels of ATF2 and ECM1 in pancreatic cancer tissues are markedly elevated compared to normal tissues (p < 0.01, Fig. 5A). Immunohistochemistry further substantiates that the protein expression levels of ATF2 and ECM1 in cancer nests of clinical patients are markedly elevated compared to surrounding tissues (Fig. 5A). In conjunction with bioinformatics predictions, miR-29c-5p possesses a conserved binding site within the 3’UTR region of ATF2 (as predicted by the RNAhybrid algorithm with a binding energy of ΔG = -28.5 kcal/mol), and ATF2, functioning as a transcription factor, has a distinct binding sequence in the promoter area of ECM1 (refer to Fig. 5B). Clinical correlation analysis reveals that miR-29c-5p expression in TCGA data has a strong negative connection with ATF2 (r = -0.62, p < 0.01) and ECM1 (r = -0.57, p < 0.01) (Fig. 5C). Functional verification indicates that the knockdown of H19 or the overexpression of miR-29c-5p mimics significantly diminishes the protein expression of ATF2 and ECM1, as evidenced by Western blot analysis with β-actin serving as an internal reference control. Specifically, ATF2 expression in the shH19 group is reduced by approximately 65% (p < 0.001), while ECM1 is decreased by 55% (p < 0.01). The inhibitory effect observed in the miR-29c-5p mimics group aligns with that of H19 knockdown (Fig. 5D). The aforementioned findings indicate that H19 facilitates the malignant advancement of pancreatic cancer by sequestering miR-29c-5p, thereby alleviating its suppression of ATF2 and subsequently activating ECM1-mediated tumor microenvironment remodeling.

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Alterations in the H19/miR-29c-5p/ATF2/ECM1 Axis during Pancreatic Cancer Progression. (A) The differential mRNA expression of ATF2 and ECM1 in pancreatic cancer in the TCGA database; in pathological immunohistochemistry, the protein expression of ATF2 and ECM1 was significantly increased in the cancer nests of clinical patients compared to the adjacent tissues, *p < 0.05, **p < 0.01, ***p < 0.001.). (B) The predicted binding sites of miR-29c-5p on the 3’UTR of ATF2 and the sequence illustration of the ATF2 binding site of the transcription factor. (C) In the TCGA database, miR-29c was negatively correlated with the expressions of ATF2 and ECM1. (D) Western blot verification of the protein levels of ATF2 and ECM1; knockdown of H19 decreased ATF2 and ECM1, while miR-29c-5p mimics exhibited a similar inhibitory effect. β-actin was utilized as a loading control, *p < 0.05, **p < 0.01, ***p < 0.001.)

H19 drives pancreatic cancer growth in xenograft models via the miR-29c-5p/ATF2/ECM1 axis-mediated tumor microenvironment remodeling

This Research has demonstrated that the H19/miR-29c-5p/ATF2/ECM1 axis substantially enhances the proliferation of pancreatic cancer xenografts in nude mice by modulating the remodeling of the tumor microenvironment. A subcutaneous injection of pancreatic cancer cells was utilized to produce an in vivo model (Fig. 6A). Live imaging illustrated that the intensity of the fluorescence of tumors in the H19 overexpression or miR-29c-5p inhibition groups was 2.3 times higher in comparison to that of the control group (NC) (p < 0.0001). In contrast, the fluorescence signal intensity in the H19 knockdown or miR-29c-5p mimic overexpression groups decreased by approximately 60% (p < 0.001). At the end of the experiment, measurements of tumor masses indicated that the H19 overexpression group exhibited an 85% increase in tumor weight compared to the normal control group (p < 0.0001), while the shH19 group showed a 55% reduction in masses (p < 0.001). Additionally, the miR-29c-5p inhibitor was able to partially counteract the suppressive effect of H19 knockdown on tumor growth (Fig. 6B). The immunohistochemical analysis revealed that the positive expression rates of ATF2 and ECM1 in the tumor tissues of the H19 overexpression group were 92% and 88%, respectively, significantly surpassing those in the NC group (45% and 52%, respectively) (p < 0.01). In contrast, the positive rates of ATF2 and ECM1 in the shH19 group diminished to 30% and 28%, respectively (p < 0.001), while the miR-29c-5p mimics group further reduced their expression (positive rates: ATF2 22%, ECM1 18%, p < 0.0001) (Fig. 6C). The aforementioned observations indicate that H19 facilitates tumor proliferation in an in vivo pancreatic cancer model by sequestering miR-29c-5p, which alleviates its targeted suppression of ATF2, thus activating ECM1-mediated extracellular matrix remodeling.

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

The H19/mir-29c-5p/ATF2/ECM1 axis promotes the growth of pancreatic cancer in a nude mouse model. (A) Tumor models were established by subcutaneous tumor formation experiments, and in vivo imaging of subcutaneous tumors in nude mice was performed; (B) Comparison of tumor weights in each group, *p < 0.05, **p < 0.01, ***p < 0.001. (C) Immunohistochemical detection of ATF2 and ECM1 expression in tumor tissues of each group; The effects of H19 and miR-29c-5p on the nude mouse PC cell mode, *p < 0.05, **p < 0.01, ***p < 0.001.

The ATF2/ECM1 axis drives pancreatic cancer progression via transcriptional regulation and microenvironment remodeling: clinical validation as diagnostic and prognostic biomarkers

Studies indicate that ATF2 and ECM1 collaboratively influence tumor advancement in pancreatic cancer and possess considerable diagnostic and prognostic significance. The comprehensive review of various clinical databases and experimental data revealed that ECM1 and ATF2 have distinct expression patterns and clinical associations in pancreatic cancer. The Human Protein Atlas database indicates that pancreatic cancer patients exhibiting low ECM1 protein expression experience a markedly extended overall survival period (p < 0.01), implying that reduced ECM1 expression may function as a beneficial prognostic molecular marker(Fig. 7A). TCGA data further corroborate that the median survival duration of patients exhibiting low ECM1 mRNA expression is 6.2 months longer than that of the control group (HR = 0.52, p < 0.001), and its diagnostic performance is exceptional (AUC = 0.978), markedly superior to the conventional marker CA19-9 (AUC = 0.82)(Fig. 7B).This aligns with the pro-cancer mechanism of ECM1 facilitating tumor microenvironment fibrosis and immune evasion. Analysis of TCGA data indicates that ATF2 mRNA exhibits elevated expression in pancreatic cancer tissues, demonstrating considerable diagnostic value (AUC = 0.949)(Fig. 7B). Immunohistochemical examination of four clinical pancreatic cancer samples reveals that both ECM1 and ATF2 proteins exhibit elevated expression within cancer nests(Fig. 7C). TCGA data indicate a substantial positive association between ATF2 and ECM1 mRNA expression (r = 0.73, p < 0.0001), implying that ATF2 may directly modulate its transcription by interacting with the ECM1 promoter region(Fig. 7D). The Human Protein Atlas database indicates that ATF2 is predominantly localized in the nucleus, aligning with its role as a transcription factor, whereas ECM1 protein is found in the cytoplasm and extracellular matrix, suggesting its involvement in tumor progression through the regulation of extracellular matrix remodeling and intercellular communication(Fig. 7E).

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

Expression and clinical significance of ATF2 and ECM1 genes in pancreatic cancer patients. (A) In the Human Protein Atlas database, low expression of ECM1 protein in pancreatic cancer indicates a better prognosis; in the TCGA database, it was found that low expression of ECM1 mRNA has a better survival prognosis; ECM1 mRNA has a good diagnostic effect, with an AUC area under the ROC curve of 0.978; (B) In the TCGA database, the AUC area under the ROC curve of ATF2 mRNA is 0.949, indicating a good diagnostic effect on pancreatic cancer; (C) In the immunohistochemistry of clinical sample pancreatic cancer pathological sections, high expression of ECM1 and ATF2 proteins was found; (D) Correlation between the mRNA expression of ATF2 and ECM1; (E) Subcellular localization of ATF2 and ECM1 in the Human Protein Atlas database.

Discussion

The results of transcript sequencing studies have recently discovered a variety long non-coding RNAs (lncRNAs), majority of which at first are associated with cancer and have demonstrated significant biological applications in the regulation of malignancy and progression17,18,19,20,21. The expression of lncRNAs in a tissue-specific manner and their regulatory roles in essential signaling pathways position them as significant biomarkers and potential therapeutic targets. In this study, we identified a PDAC-promoting lncRNA, H19, through public database analysis. The expression of H19 has been considerably overexpressed in pancreatic ductal adenocarcinoma cell lines compared to normal pancreatic ductal epithelial cell lines. Furthermore, patients with high H19 expression exhibit poor prognosis. Previous studies have also shown that H19 functions as an oncogene in various cancers, including esophageal, gastric, colorectal, liver, and gallbladder cancers22,23,24,25.Significantly, functional investigations revealed that H19 enhances the growth, invasion, and migration of pancreatic cancer cells within vitro alongside in animals. These findings are consistent with prior research, further confirming the oncogenic function of H19 in PDAC.

This research presented evidence that H19 impacts the ECM1 protein in pancreatic cancer. ECM1 has been validated as an oncogenic associated with multiple malignancies, including ovarian and carcinoma of the pancreas26,27. In this study, we found that H19 can positively regulate the expression of ECM1.

As competing endogenous RNAs (ceRNAs), a core activity of many lncRNAs is to act as molecular sponges for microRNAs (miRNAs). They competitively bind to common miRNAs, hence indirectly modulating mRNA stability and translation, and extensively affecting gene expression at the post-transcriptional stage of transcription28,29,30,31. The physiological function of lncRNAs is significantly influenced by their cytoplasmic localization: nuclear lncRNAs predominantly govern transcriptional activities and epigenetic alterations, whereas cytoplasmic lncRNAs primarily operate at the post-transcriptional stage by serving as ceRNAs to sequester miRNAs, thereby modulating the integrity of mRNA and translation, among additional mechanisms. This work determined that lncRNA H19 is localized in the cytoplasm by the integration of bioinformatics database analysis and RNA fluorescent in situ hybridization, also known as FISH, investigations. This distribution pattern provides experimental evidence for its potential to act as a ceRNA, adsorbing miRNAs and participating in post-transcriptional regulation. The research results revealed that miR-29c-5p manifestation is profoundly decreases in PDAC patients and corresponds with a greater likelihood of survival. In vitro studies demonstrated that the miR-29c-5p may decrease the proliferation and expansion of pancreatic ductal adenocarcinoma cell lines. More importantly, subsequent luciferase reporter enzyme research validated that H19 functions as a sponge for miR-29c-5p in pancreatic cancer cells by directly binding to complementary regions. This is the inaugural research detailing the tumor-suppressive function of miR-29c-5p in pancreatic ductal adenocarcinoma, although miR-29c-5p has previously been identified as a tumor suppressor in other malignancies, including gallbladder cancer32. In this study, increased expression of miR-29c-5p restricted the proliferation and migration of pancreatic malignancy cells. Rescue tests suggest the increased levels of miR-29c-5p might partially counteract the pro-tumor effects induced by H19 overexpression. Conversely, knockdown of miR-29c-5p in pancreatic cancer cells produced effects opposite to those of H19 knockdown, proving that miR-29c-5p is a functional mediator of H19 in pancreatic cancer. These data indicate that part of the oncogenic mechanism of H19 in pancreatic cancer is achieved by sponging miR-29c-5p.

Generally, as a ceRNA, the biological activity of an lncRNA depends on its targeted miRNAs. To anticipate the possible targeting of genes of miR-29c-5p in PDAC along with investigate the downstream molecules that may mediate the H19/miR-29c-5p axis, we conducted an integrative bioinformatics database study.The results predicted that ATF2 could be a potential target of miR-29c-5p. Moreover, Western blot analyses demonstrated the regulatory interaction of the H19/miR-29c-5p pathway with both ATF2 and ECM1. ECM1 is a secreted glycoprotein with multiple functions, including stimulating cell proliferation, promoting angiogenesis, negatively regulating endochondral ossification, and promoting tumor progression33,34.Consistent with expectations, we observed that the expression levels of both ECM1 and ATF2 are raised in PDAC tissues relative to normal tissues, and heightened ECM1 expression is substantially correlated with adverse prognosis in PDAC patients. Moreover, an examination of the TCGA database revealed that ATF2 expression levels have an upward trend with ECM1 and a negative correlation with miR-29c-5p. Notably, in pancreatic cancer cells, H19 positively regulates ECM1 expression, while miR-29c-5p shows the opposite regulatory effect.

This study demonstrates that H19 functions as an oncogenic long non-coding RNA in pancreatic ductal adenocarcinoma. Demonstrable overexpression is observed in pancreatic malignancy. Elevated H19 expression is strongly correlated with unfavorable clinicopathological characteristics and diminished prognosis in patients with PDAC. Functional experiments indicate that H19 promotes pancreatic cancer cell proliferation and migration both in vitro and in vivo. Mechanistically, H19 directly binds to miR-29c-5p and acts as a competing endogenous RNA to upregulate ECM1 expression. This study systematically elucidated the tumor-promoting mechanism of the H19/miR-29c-5p/ATF2/ECM1 axis in pancreatic cancer; however, several limitations remain. First, the clinical sample size was limited (only 4 cases of matched tumor and adjacent normal tissues), which restricted the statistical power and may affect the generalizability of the conclusions. Continued study necessitates bigger samples for substantiation. Second, direct experimental evidence for ECM1-mediated extracellular matrix (ECM) remodeling—such as morphological or biochemical validation—is lacking, leaving its specific functional mechanism largely inferential. The interaction between this axis and the tumor microenvironment (TME), especially with crucial elements like macrophages associated with the tumor (TAMs) and fibroblasts associated with tumors (CAFs), remains inadequately understood, warranting additional research into this cross-talk. Future research should aim to deepen the mechanistic understanding and promote clinical translation by expanding sample sizes, employing advanced imaging techniques, and utilizing models such as organoid co-culture system.

Conclusions

This study provides evidence for the oncogenic role of H19 in pancreatic ductal adenocarcinoma (PDAC) through multidimensional validation. At diagnosis, H19 overexpression in PDAC tissues and cell lines is strongly associated with aggressive pathological features and significantly shortened patient survival, highlighting its potential as a biomarker for stratifying high-risk patients requiring intensified therapy. In advanced stages, H19 promotes metastasis by functioning as a ceRNA that sponges tumor-suppressive miR-29c-5p, thereby derepressing downstream targets ATF2 (a transcription factor) and ECM1 (an extracellular matrix protein). This molecular axis contributes to remodeling of the tumor microenvironment and confers resistance to therapy. Notably, ECM1 exhibits superior diagnostic accuracy compared to conventional markers, and co-expression of ATF2 and ECM1 serves as a strong predictor of poor prognosis. Collectively, targeting the H19/miR-29c-5p/ATF2/ECM1 axis offers the potential for refined patient stratification and innovative therapeutic strategies against this highly lethal malignancy.