LINC00909 up-regulates pluripotency factors and promotes cancer stemness and metastasis in pancreatic ductal adenocarcinoma by targeting SMAD4

Abstract

Background

Pancreatic cancer stem cells are crucial for tumorigenesis and cancer metastasis. Presently, long non-coding RNAs were found to be associated with Pancreatic Ductal Adenocarcinoma stemness characteristics but the underlying mechanism is largely known. Here, we aim to explore the function of LINC00909 in regulating pancreatic cancer stemness and cancer metastasis.

Methods

The expression level and clinical characteristics of LINC00909 were verified in 80-paired normal pancreas and Pancreatic Ductal Adenocarcinoma tissues from Guangdong Provincial People’s Hospital cohort by in situ hybridization. RNA sequencing of PANC-1 cells with empty vector or vector encoding LINC00909 was experimented for subsequent bioinformatics analysis. The effect of LINC00909 in cancer stemness and metastasis was examined by in vitro and in vivo experiments. The interaction between LINC00909 with SMAD4 and the pluripotency factors were studied.

Results

LINC00909 was generally upregulated in pancreatic cancer tissues and was associated with inferior clinicopathologic features and outcome. Over-expression of LINC00909 enhanced the expression of pluripotency factors and cancer stem cells phenotype, while knock-down of LINC00909 decreased the expression of pluripotency factors and cancer stem cells phenotype. Moreover, LINC00909 inversely regulated SMAD4 expression, knock-down of SMAD4 rescued the effect of LINC00909-deletion inhibition on pluripotency factors and cancer stem cells phenotype. These indicated the effect of LINC00909 on pluripotency factors and CSC phenotype was dependent on SMAD4 and MAPK/JNK signaling pathway, another downstream pathway of SMAD4 was also activated by LINC00909. Specifically, LINC00909 was localized in the cytoplasm in pancreatic cancer cells and decreased the stability the SMAD4 mRNA. Finally, we found over-expression of LINC00909 not only accelerated tumor growth in subcutaneous mice models, but also facilitated tumorigenicity and spleen metastasis in orthotopic mice models.

Conclusion

We demonstrate LINC00909 inhibits SMAD4 expression at the post-transcriptional level, which up-regulates the expression of pluripotency factors and activates the MAPK/JNK signaling pathway, leading to enrichment of cancer stem cells and cancer metastasis in pancreatic cancer.

Introduction

In terms of aggressiveness, pancreatic cancer (PC) is among the worst diseases of the digestive system; most patients present with vague symptoms until the disease reaches an advanced stage [1]. PC mortality has been rising in recent years, with a survival rate of 10% in the United States after 5 years [2]. High aggressiveness and early metastasis are crucial factors linked to inferior prognosis in pancreatic ductal adenocarcinoma (PDAC) patients [3, 4].

Tumor initiation and development are driven by a few colony of differentiated cells called cancer stem cells (CSCs). According to recent studies, pancreatic CSCs assume a crucial role during tumor recurrence, metastasis, and resistance to chemotherapy in PDAC [5]. There is evidence that the activation of epithelial-to-mesenchymal transition (EMT), a process associated with cancer cell metastasis, is linked to the properties of CSCs [6, 7]. Researchers have identified several pluripotent transcription factors, including c-Myc, KLF4, OCT4, SOX2, and NANOG, which are responsible for regulating stemness [8]. An abundance of CSCs and/or high expression of pluripotency factors in tumors may forebode worse prognosis in PDAC patients, indicating that repression of cancer stemness could provide a prospective means to improving the PDAC treatment. SMAD4 is one of the most important driver genes in PDAC. This cancer suppressor gene is inactivated due to mutation or homozygous deletion in 60% of PDAC cases [9]. Inhibition of JUN N-terminal kinase (JNK) by SMAD4 suggests that deletion of SMAD4 would enhance JNK activity [10] and be required in the epithelial-to-mesenchymal transition response [11]. Importantly, SMAD4 was found to affect the expression of pluripotency factors and regulate the CSC phenotype.

The term long non-coding RNA (LncRNA) refers to RNAs whose length exceeds 200 nt but lack the capability of coding proteins [12]. It has been demonstrated that lncRNAs mediated multi-faceted biological progression, such as immune response, cell differentiation, apoptosis, and cancer stemness [13]. For example, lncRNA-HOTTIP upregulated the expression of HOXA9 as a result of activating the Wnt/β-catenin pathway, which is essential for maintaining the CSC phenotype in PDAC [14]. The level of long intergenic non-protein coding RNA 909 (LINC00909), a lncRNA with 15,407 bases length, is increased in glioma. Knockdown (KD) of LINC00909 inhibited the proliferation of glioma cells, indicating that LINC00909 drives tumor progression in glioma [15]. Nevertheless, the biological function of LINC00909 in PDAC, in particular whether LINC00909 contributes to regulating the CSC phenotype, is largely undiscovered.

As a result of this study, we identified that LINC00909 was markedly upregulated in PDAC tissues. Moreover, LINC00909 enhanced stemness and promoted metastasis in PDAC. Additionally, the increased expression of pluripotency factors induced by LINC00909 was dependent on the downregulation of SMAD4. Our study suggests that LINC00909 plays a vital role in controlling the CSC phenotype as well as cancer metastasis in PDAC.

Materials and methods

Identification of lncRNA

The Cancer Genome Atlas (TCGA) database (http//gdc.cancer.gov/) was used to obtain data on the expression of lncRNAs in 171 patients with PDAC.

Patients and clinical samples

Eighty pairs of primary PDAC specimens were collected from patients undergoing surgical resection of PC at the Guangdong Provincial People’s Hospital (GDPH) between February 2008 and June 2021. After specimen collection, a temperature of 80 °C was used for the storage of all PDAC tissue samples. A summary of the clinicopathological characteristics of the patients is presented in Table S1. This study was authorized by the Ethics Committee of GDPH.

Next-generation sequencing (NGS)

RNA was collected from PANC-1/Vector and PANC-1/LINC00909-overexpressing (PANC-1/LINC00909-OE) cells. Next-generation sequencing (NGS) was conducted as previously described [16].

Functional enrichment analysis

According to the NGS results, we determined DEGs associated with LINC00909 (adjusted P < 0.05 and Log₂|fold change| >1.0). These DEGs were integrated into DAVID 6.8 (https://david-d.ncifcrf.gov/) and SangerBox (http://sangerbox.com/) for Gene Ontology (GO) analyses and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses, respectively. Gene set enrichment analysis (GSEA) indicated the functional pathways and gene sets that exhibited obvious diversity between the high- and low-LINC00909 expression groups.

Materials

Gemcitabine (abs814679) was obtained from Univ (Shanghai, China). Antibodies against CD133 (#64,326, 1/1,000), c-Myc (#5605, 1/1,000), NANOG (#4903, 1/2000), SOX2 (#14,962, 1/1,000), CD133 (#53,276, 1/50), and aldehyde dehydrogenase 1 family member A1 (ALDH1A1; #65,583, 1/50) were acquired from Cell Signaling Technology (Boston, MA, USA). Antibodies against KLF4 (ab215036, 1/1,000), OCT4 (ab181557, 1/1,000), BCL-2 (ab32124, 1/1,000), p53-upregulated modulator of apoptosis (PUMA; ab33906, 1/2,500), β-actin (ab8226, 1/5,000), and GAPDH (ab8245, 1/2,000) were obtained from Abcam (Cambridge, UK). Antibodies against SMAD4 (AF5247, 1/1,000), p-JNK (Thr183 + Thr185) (AF3318, 1/1,000),XXXJNK (AF6318), p-JUN (Ser73) (AF3095), c-JUN (AF6090, 1/1,000), TWIST (AF4009, 1/1000), SNAIL (AF6032, 1/1,000) were obtained from Affinity (Jiangsu Province, China). Actinomycin D (SBR00013) was obtained from Sigma–Aldrich (USA).

Fluorescence in situ hybridization (FISH)

PDAC cells were collected, fixed with paraformaldehyde at 80–90% density, and pre-hybridized. Next, the Cy3-labeled LINC00909 probe 5’-CTTTATCCACTCGTTGGAATGATTTTTTTGAGAC-3’ (Servicebio, Wuhan Province, China) was hybridized with hybridization buffer overnight at 37 °C. All pictures were taken by laser scanning confocal microscope (Olympus,Tokyo, Japan).

Chromogenic in situ hybridization (CISH)

CISH were conducted using a PDAC and normal tissue micro-array. CISH was utilized for detecting LINC00909 expression in tissue using a specific 5′- and 3′-digoxigenin-labeled probes. Using an organizational slicing digital scanner or imaging system, capture scan files or images of immunohistochemically stained tissue sections. With the help of an image analysis system, automatically extract measurements from the specified tissue areas. Classify positive staining into levels (i): Negative (0 score), Weakly positive (score 1, pale yellow), Moderately positive (score 2, brown-yellow), and strongly positive (score 3, brown-brown). Analyze weak, moderate, and strong positive areas, total tissue area, cumulative optical density (IOD) value of positive staining, and positive area within the measurement regions. Calculate these results to reflect the strength of positivity.

To assess positive staining intensity, we use the H-score method, tailored for specific slide characteristics. The H-score, short for Histochemistry score, converts positive staining quantity and intensity into numerical values, providing a semi-quantitative tissue staining assessment. The H-Score (H-SCORE = ∑(pi×i) = (percentage of weak intensity area ×1) + (percentage of moderate intensity area ×2) + (percentage of strong intensity area ×3)) employs pi for the positive signal pixel percentage and i for the positive staining level. H-score ranges from 0 to 300, with higher values indicating stronger overall positivity.

Immunohistochemistry (IHC)

IHC were conducted using a PDAC and normal tissue micro-array. The tumor tissue was stained for SMAD4, c-Myc, KLF4, and c-JUN. The critical process was blocked 30 min in normal goat serum, then incubated with a primary antibody and a secondary antibody.

PDAC cell lines and culture

The human PDAC cell lines (PANC-1, BxPC-3, CAPAN-2, AsPC-1, SW1990), and normal pancreas epithelium cell line (HPED-6) used in our study were obtained from Procell (Wuhan, China).

AsPC-1 and the other cells were respectively cultured in RPMI 1640 medium (Gibco, USA) and high-glucose DMEM (Gibco) with 10% fetal bovine serum (FBS) (Gibco) at 37 °C and 5% CO₂.

Lentivirus infection and siRNA KD

To establish stable OE cell lines, PANC-1 cells were infected with lentivirus [NR_024484.1] containing the EFS-hLINC00909-CMV-EGFP/T2A/Puro. After puromycin selection, we acquired stable OE cells. Following the instructions provided by manufacturer, we knocked down LINC00909 and SMAD4 by transfecting the siRNAs into PANC-1 or AsPC-1 cells. The siRNA sequences are shown in the Table S1.

Proliferation assays

Cell Counting Kit-8 (CCK-8) and colony formation assays were utilized to measure cell proliferation. For the CCK-8 assays, a 96-well plate with 2*10³ PANC-1 and AsPC-1 cells per well was resuspended and cultured for 0, 24, 48, and 72 h. The CCK-8 reagent was added, and the cells were incubated for 2 h at 37 °C. Thereafter, the number of PDAC cells was determined by optical density at OD 450.

For the colony formation assays, PANC-1 or AsPC-1 cells were seeded into six-well plates (3,000 cells per well). These cells were cultivated in an incubator at 37 °C and 5% CO₂ for 7 days. After rinsing with PBS, we stained the cells were with 0.1% crystal violet to observe the formation of colonies.

Wound-healing and Transwell assays

When the cell density in the six-well plate reached 85%, we used a sterile pipette tip to make a scratch in the center of each well. Wound-healing results were observed at 0 and 24 h. The size of the wound was measured at least thrice.

For Transwell assays, 5 × 10⁴ PDAC cells were totally seeded in the upper chamber of Transwell plates with DMEM medium containing 10% FBS, while DMEM (800 µl) containing 20% FBS was injected into the lower chamber. After 24–48 h of incubation at 37 °C, the cells migrated from the upper chamber to the lower chamber. Subsequently, the upper chambers were taken out for washing thrice with PBS, then the cells were stained with 0.1% crystal violet. The amount of PDAC cells was calculated using microscope (Olympus,Tokyo, Japan).

Sphere-formation assays

A total of 5 × 10³ PANC-1 or AsPC-1 cells were cultured in six-well ultra-low attachment plates (Corning, NY, USA) containing DMEM/F12 medium with B-27 supplement (1:50, Gibco), fibroblast growth factor (FGF; 20 ng/ml, Invitrogen, California, USA) and epidermal growth factor (EGF; 20 ng/ml, Invitrogen). Under a microscope (Olympus, Tokyo, Japan), tumor spheres with a diameter greater than 50 mm were counted and photographed after 7 days.

RT-qPCR assays

The TRIzol reagent (Invitrogen, California, USA) was utilized to extract total RNA from PDAC tissues, human PDAC cell lines, and HPDE-6 cells. Total RNA (1–2 µg) was reverse-transcribed by High Capacity RNA-to-cDNA Kit (TakaRa, Tokyo, Japan). The primer sequences used for RT-qPCR were shown in the Table S2.

Western blotting

For western blotting, these cells were collected and dissolved in RIPA buffer with 1% phenylmethylsulfonyl fluoride, 1% protease inhibitor, and 1% phosphatase inhibitor in an ice bath for 25 min to extract the protein. The BCA protein kit (#23,235; Thermo Fisher Scientific) was utilized to determine the protein concentration. On an 10% SDS-PAGE gel, the same amount of 30 µg protein was separated and transferred onto a polyvinylidene fuoride (PVDF) membrane. After blocking for 1 h in 5% skim milk, primary antibodies were incubated overnight at 4 °C with the PVDF membrane. Whereafter, the membranes were rinsed with TBST buffer five times (5 min per wash) and incubated with horseradish peroxidase-labeled goat anti-rabbit IgG secondary antibody. Finally, the enhanced chemiluminescence reagent was added onto the membranes to visualize the proteins using a Tanon-5200 Chemiluminescent Imaging System (Tanon, China).

Results

LINC00909 was highly expressed in human PDAC and associated with worse clinicopathologic features and outcome

We sought to assess the role of LINC00909 in cancer progression. For this purpose, we used the GEPIA database (http://gepia.cancer-pku.cn/) to investigate whether LINC00909 was obviously expressed in PDAC (Fig. 1A). We found that LINC00909 was overexpressed in most types of cancer (Fig. 1B). To further verify the findings from the GEPIA database, the expression of LINC00909 in 20 pairs of fresh PDAC tissues and adjacent normal tissues was subsequently examined by RT-qPCR analysis. Comparing with adjacent normal tissues, the results exhibited higher LINC00909 mRNA expression in PDAC tissues (Fig. 1C). Moreover, CISH analysis and the H-score of CISH revealed higher expression of LINC00909 in the glandular epithelium tissues of PDAC than in the adjacent normal tissues (Fig. 1D, E). CISH staining of these 80 pairs of PC and adjacent normal tissues was performed to evaluate the clinical role of LINC00909. Importantly, Kaplan–Meier analysis for the GDPH cohorts demonstrated that PDAC patients with high LINC00909 expression were associated with poor overall survival (P = 2.996e-03) and disease-free survival (P = 7.059e-04) (Fig. 1F). Furthermore, patients with high LINC00909 expression in other types of tumors (e.g., liver hepatocellular carcinoma, brain lower grade glioma, adrenocortical carcinoma, and bladder urothelial carcinoma) also had poor prognosis (Fig. S1).

Fig. 1

LINC00909 is highly expressed in human pancreatic ductal adenocarcinoma (PDAC) and associated with worse clinicopathological features and outcome. (A) Upregulation of LINC00909 in PDAC was verified using TCGA and GTEx data. (B) Analysis of data from the GEPIA database demonstrated that LINC00909 expression was overexpressed in lymphoid neoplasm diffuse large glioblastoma multiforme (GBM), B-cell lymphoma (DLBC), brain lower grade glioma (LGG), acute myeloid leukemia (LAML), pheochromocytoma and paraganglioma (PCPG), thymoma (THYM) and skin cutaneous melanoma (SKCM). (C) RT-qPCR revealed high expression of LINC00909 in human PDAC tissues compared with normal pancreas tissues (n = 20). (D) Representative images of LINC00909 CISH analysis in the normal pancreas tissues (n = 80) and PDAC tissues (n = 80). (E) The H-Score of CISH analysis confirmed that the expression of LINC00909 was upregulated in PDAC tissues. (F) Kaplan–Meier analyses representing the overall survival (OS) and disease-free survival (DFS) of patients with PDAC. These patients with high expression of LINC00909 in the GDPH cohort were associated with poor prognosis. (G) LINC00909 expression in PDAC was strongly correlated with advanced TNM stage, larger tumor size, poorer differentiation, and more lymph node metastasis. (H) Multivariate Cox regression analyses of 80 patients with PDAC from the GDPH cohort showed that high levels of LINC00909 served as an independent risk factor for OS and DFS. All *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. P-values were caculated by the non-parametric Mann–Whitney U-test in (A) and (B). P-values were calculated by two-tailed t-tests in (C) and (E). P-values were assessed by χ² tests or Fisher’s exact tests in (G). P-values and the hazard ratios (HR) by the log-rank (Mantel–Cox) test are calculated in (H)

Additionally, the correlation between LINC00909 expression and the clinicopathological features of patients with PDAC indicated that LINC00909 OE was correlated with advanced TNM stage (P < 0.05), larger tumor size (P < 0.05), poorer differentiation (P < 0.05), and more lymph node metastasis (P < 0.01) (Table S3 and Fig. 1G). Thereafter, the multivariate Cox regression analyses revealed that LINC00909 served as a risk factor for poor prognosis in 80 GDPH patients with PDAC (Tables S4, S5 and Fig. 1H).

LINC00909 promoted PDAC cell proliferation and migration in vitro

We performed RT-qPCR to monitor the expression of LINC00909 in each cell line. The results showed that LINC00909 had higher expression in PDAC cell lines (particularly PANC-1 and AsPC-1) compared with HPDE-6 cells (a normal pancreatic cell line) (Fig. 2A). FISH analysis was conducted to investigate the localization of LINC00909 in PDAC cells. LINC00909 RNA was principally located in the cytoplasm of PDAC cell lines (e.g., AsPC-1, PANC-1, and CAPAN-2) (Fig. 2B). Subsequently, a stable LINC00909-OE cell line (PANC-1/LINC00909-OE) was constructed to further analyze the gain-of-function of LINC00909 (Fig. 2C). In contrast, six siRNAs of LINC00909 were transiently transfected into PANC-1 and AsPC-1 cells to obtain LINC00909-KD cell lines (PANC-1/LINC00909-KD and AsPC-1/LINC00909-KD) for deletion-of-function analysis by gene KD (Fig. 2E, F). Subsequently, proliferation assays were performed using these three cell lines to evaluate the proliferative capability of these cells. As shown in Fig. 2D and I, upregulation of LINC00909 significantly increased the proliferation of PANC-1 cells. On the contrary, downregulation of LINC00909 prohibited the proliferative capacity of PANC-1/LINC00909-KD and AsPC-1/LINC00909-KD cell lines (Fig. 2G, H, J, K). We also performed Transwell and wound-healing assays to assess the functions of LINC00909 on cell migration. These migration assays indicated that OE of LINC00909 augmented the migration of PANC-1/LINC00909-OE cells (Fig. 2L and S2A), whereas silencing of LINC00909 weakened the migratory ability of tumor cells in KD cell lines in vitro (Fig. 2M, N and S2B, C).

Fig. 2

LINC00909 promotes the proliferation and migration of pancreatic ductal adenocarcinoma (PDAC) cells. (A) RT-qPCR showed that LINC00909 was highly expressed in PANC-1 and AsPC-1 cells among PDAC cell lines. (B) Identification of the cytoplasmic location of LINC00909 in AsPC-1, PANC-1, and CAPAN-2 cells using FISH. (C) The efficiency of LINC00909 overexpression in PANC-1 cells was assessed using RT-qPCR. (D) The effects of LINC00909 overexpression on cell proliferation and viability were verified using CCK-8 assays. (E, F) The efficiency of LINC00909 knockdown in PANC-1 and AsPC-1 cells was detected by RT-qPCR. (G, H) The effects of LINC00909 knockdown on cell proliferation and viability were verified using CCK-8 assays. (I–K) The effect of LINC00909 overexpression and knockdown on colony formation was investigated in PDAC cell lines. (L–N) Representative images of Transwell assays after LINC00909 overexpression and knockdown in PDAC cell lines. All *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. P-values were assessed using ANOVA and two-tailed t-tests, followed by Dunnett’s tests for multiple comparisons in (A) and (C–N). All data represent the means ± SD from 3 independent assays

LINC00909 was associated with the MAPK/JNK pathway and pluripotency factors in PDAC

To further explore the function of LINC00909, 2265 DEGs (Log₂|fold change| >1.0 and adjusted P-value < 0.05) obtained from the NGS database of PANC-1/Vector and PANC-1/LINC00909-OE cells were inputted into DAVID 6.8 and SangerBox for functional enrichment analyses, respectively. The DEGs data was stored in Table S6. The GO analysis showed that LINC00909 and the DEGs were principally enriched in “extracellular matrix organization”, “nuclear division”, “extracellular structure organization”, “regulation of metaphase/anaphase transition of cell cycle”, and “kinetochore” (Fig. 3A). The KEGG pathway enrichment analysis indicated that LINC00909 was likely participate in “MAPK signaling pathway”, “cell adhesion molecules”, “cell cycle”, and “P53 signaling pathway” (Fig. 3B). Meanwhile, GSEA was also carried out to identify the DEGs and pathways distinguishing between the -high and -low expression of LINC00909. No surprise, biological processes were significantly enriched in “MAPK signaling pathway”, “NF-κB signaling pathway”, “JAK-STAT signaling pathway”, “cell adhesion molecules”, and “apoptosis” (Fig. 3C).

Fig. 3

LINC00909 is associated with the MAPK/JNK pathway and pluripotency factors in pancreatic ductal adenocarcinoma (PDAC). (A) Gene Ontology (GO) enrichment analysis of LINC00909 based on the NGS database. (B) Top 15-enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways of LINC00909 based on the NGS database. (C) Gene set enrichment analysis (GSEA) of the NGS database. (D) The expression of factors associated with stemness and metastasis was analyzed by NGS using PANC-1/Vector and PANC-1/LINC00909-OE cells. These factors are shown in the heatmap. (E–G) Western blotting analysis was conducted to detect the protein expression of JNK phospho-JNK, c-JUN, phospho-JUN, TWIST, and SNAIL in LINC00909-overexpressing cells (E) and LINC00909-knockdown cells (F, G). All *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. P-values were assessed using ANOVA and two-tailed t-tests, followed by Dunnett’s tests for multiple comparisons in (D)

Hence, the mRNA expression of essential molecules associated with the MAPK signaling pathway and other potential pathways were analyzed according to the NGS. We found that the expression of JUN and FOS in the MAPK pathway was upregulated. Interestingly, pluripotency factors c-Myc, KLF4, and stem cell markers (EPCAM and ALDH1) were also upregulated (Fig. 3D). Next, we utilized western blotting to verify the protein expression of phospho-JNK, as well as its downstream and metastasis-related molecules to explore the effect of LINC00909 on the MAPK/JNK pathway. And the results disclosed that the protein expression of phospho-JNK, c-JUN, phospho-JUN, and tumor metastasis-related molecules (e.g., TWIST and SNAIL) was obviously enhanced in PANC-1/LINC00909-OE cells (Fig. 3E), whereas it was decreased in the PANC-1/LINC00909-KD and AsPC-1/LINC00909-KD cell lines (Fig. 3F, G). These results demonstrated that LINC00909 was associated with the MAPK/JNK pathway and pluripotency factors in PDAC.

LINC00909 upregulated the expression of pluripotency factors and promoted the pancreatic CSC phenotype

We performed sphere-formation assays in three PDAC cell lines to investigate whether LINC00909 regulates the pancreas CSC phenotype. From the results, LINC00909-OE significantly augmented the amount of PANC-1 cell spheres (Fig. 4A), whereas LINC00909-KD seriously impaired the sphere-formation capability of PANC-1 and AsPC-1 cells (Fig. 4B, C). We also measured the stem cell populations of CD133 and ALDH1 double-positive cells using flow cytometry. OE of LINC00909 markedly augmented the double-positive rate of CD133⁺ and ALDH1⁺ (Fig. 4D), whereas KD of LINC00909 reduced the expression of these markers (Fig. 4E, F). Next, we explored the effect of LINC00909 on the expression of pluripotency factors. The GEPIA database analysis suggested that LINC00909 was positively correlated with several stemness factors (Fig. S4A). According to RT-qPCR and western blotting results, the mRNA and protein levels of these stemness factors, particularly c-Myc, KLF4, and CD133, were drastically enhanced in PANC-1/LINC00909-OE (Fig. 4G, J). In contrast, they were noticeably reduced after the deletion of LINC00909 in PANC-1 and AsPC-1 cells (Fig. 4H, I, K, L). In summary, these data manifested that LINC00909 affects the expression of pluripotency factors and is necessary for maintaining CSC properties in PDAC.

Fig. 4

LINC00909 promotes the pancreatic CSC phenotype and upregulates the expression of pluripotency factors. (A) Sphere-formation assay in PANC-1 cells transfected with empty vector (EV) or vector encoding LINC00909. (B, C) Sphere-formation assay in PANC-1 and AsPC-1 cells transfected with siNC, siLINC00909-3, or siLINC00909-6. (D, F) CD133 + and ALDH1 + expression in PDAC cells was measured by flow cytometry. (G–I) RT-qPCR analysis was conducted to detect the mRNA levels of CD133, OCT4, SOX2, NANOG, KLF4, and c-Myc in LINC00909-overexpressing cells (G) and LINC00909-knockdown cells (H, I). (J–L) Western blotting analysis was conducted to detect the protein expression of CD133, OCT4, SOX2, NANOG, KLF4, and c-Myc in PDAC cells. (M) CCK-8 assays were conducted to detect the cell viability of PANC-1 cells transfected with EV or vector encoding LINC00909 under treatment with different concentrations of gemcitabine. (N, O) CCK-8 assays were conducted to detect the cell viability of PANC-1 or AsPC-1 cells transfected with siNC, siLINC00909-3, or siLINC00909-6 under treatment with different concentrations of gemcitabine. All *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. P-values were assessed using ANOVA and two-tailed t-tests, followed by Dunnett’s tests for multiple comparisons in (A–I) and (M–O).

CSCs contribute to the development of chemo-resistance in cancer. We next investigated whether LINC00909 affects the chemo-sensitivity of PDAC cells. Gemcitabine, a pyrimidine anti-tumor agent commonly used in patients with advanced PC, was employed in the in vitro experiments. The results showed that LINC00909-OE lessened the sensitivity of PANC-1 cells to gemcitabine, as evidenced by the elevated cell viability following treatment with gemcitabine (Fig. 4M). Conversely, suppression of LINC00909 sensitized PANC-1 and AsPC-1 cells to gemcitabine, as reflected by the declining cell viability (Fig. 4N, O). These data indicated that OE of LINC00909 leads to chemo-resistance in PDAC.

LINC00909 enhanced the stemness of PDAC by inhibiting SMAD4 expression

We next investigated the underlying mechanism through which LINC00909 regulates the expression of pluripotency factors. LINC00909 is primarily distributed in the cytoplasm, while pluripotency factors (e.g., KLF4 and c-Myc) are transcription factors localized in the nucleus. Therefore, we hypothesized that LINC00909 may not physically interact and directly regulate these pluripotency factors.

The results of RNA sequencing showed that SMAD4 was downregulated in PANC-1/LINC00909-OE cells, and SMAD4 regulates the expression of pluripotency factors. Therefore, we investigated the levels of SMAD4 to further identify correlations with LINC00909. RT-qPCR and western blotting analysis confirmed that both the mRNA and protein levels of SMAD4 were declined in LINC00909-OE cells (Fig. 5A, D), while SMAD4 was upregulated after deletion of LINC00909 (Figs. 5B and C and 6E and F). To further investigate whether SMAD4 deficiency contributes to the LINC00909-mediated enhancement of stemness in PDAC cells, we knocked down SMAD4 in PANC-1/LINC00909-KD and AsPC-1/LINC00909-KD cells and performed sphere-formation assays. The KD efficiency of SMAD4 in these two cell types was confirmed by RT-qPCR, and siSMAD4-2 was selected for subsequent experiments (Fig. 5G, H). As shown in sphere-formation assays, the downregulation of SMAD4 transparently enhanced the reduction of spherical formation mediated by LINC00909 deficiency in PANC-1 and AsPC-1 cells (Fig. 5I, J). Intriguingly, the western blotting analysis revealed that SMAD4 deficiency markedly augmented the decrease in KLF4, c-Myc, p-JNK and p-JUN expression mediated by LINC00909-KD (Fig. 5K, L). These consequences confirmed that the regulatory effect of LINC00909 on pluripotency factors, cancer stemness and the activation of the MAPK/JNK pathway were dependent on the inhibition of SMAD4.

Fig. 5

LINC00909 enhances the stemness of pancreatic ductal adenocarcinoma (PDAC) by inhibiting SMAD4 expression. (A–C) RT-qPCR analysis was performed to detect the mRNA levels of SMAD4 in LINC00909-overexpressing cells (A) and LINC00909-knockdown cells (B, C). (D–F) Western blotting analysis was conducted to examine the protein expression of SMAD4 in PDAC cells. (G, H) The efficiency of SMAD4 knockdown in PANC-1 (G) and AsPC-1 cells (H) was detected by RT-qPCR. (I, J) Representative images of sphere-formation assays in PANC-1/KD and AsPC-1/KD cells transfected with siSMAD4-2. (K, L) The protein levels of SMAD4, KLF4, c-Myc, JNK, p-JNK, JUN and p-JUN in PANC-1/LINC00909-KD (K) and AsPC-1/LINC00909-KD cells (L) transfected with siSMAD4-2 were analyzed by western blotting. (M) PANC-1 cells transfected with EV or vector encoding LINC00909 were treated with actinomycin D (1 mg/ml) at the indicated time point. (N, O) PANC-1 or AsPC-1 cells transfected with siNC, siLINC00909-3, or siLINC00909-6 were treated with actinomycin D (1 mg/ml) at the indicated time point. Total RNA was extracted and analyzed by RT-qPCR to examine the relative levels of SMAD4 mRNA (M–O). All *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. P-values were assessed using ANOVA and two-tailed t-tests, followed by Dunnett’s tests for multiple comparisons in (A–O).

LncRNAs localized in the cytoplasm act as a momentous role in post-transcriptional gene regulation (e.g., mRNA stability and protein modification). We hypothesized that LINC00909 affects SMAD4 mRNA stability since both the mRNA and protein levels were changed in LINC00909-KD or -OE cells. We treated cells with actinomycin D to inhibit general gene transcription and performed RT-qPCR at the indicated time point to explore the turnover of SMAD4 mRNA. The results showed that OE of LINC00909 accelerated the turnover of SMAD4 mRNA (Fig. 5M), while KD of LINC00909 exerted the opposite effect (Fig. 5N, O). These results imply that LINC00909 inhibits SMAD4 expression by decreasing SMAD4 mRNA stability.

LINC00909 inhibited apoptosis in PC cells.

According to the GSEA analysis, the biological process “apoptosis” was significantly enriched (Fig. 3C); four genes associated with apoptosis were selected to analyze their correlation with LINC00909 in TCGA dataset using the GEPIA database. The analysis predicted that the levels of LINC00909 were positively correlated with BCL2 (correlation [Cor] = 0.41, P = 7.8e-09), and negatively correlated with BCL2 binding component 3 (BBC3; Cor = − 0.33, P = 7.8e-06), BCL2 antagonist/killer 1 (BAK1; Cor = 0.3, P = 3.4e-05), and BH3 interacting domain death agonist (BID; Cor = − 0.26, P = 0.00055) (Fig. S3B). Subsequently, the consequences of flow cytometry confirmed that LINC00909-OE distinctly decreased the amount of apoptotic PANC-1 cells (Fig. 6A); however, LINC00909-KD obviously increased the quantity of apoptotic PANC-1 and AsPC-1 cells (Fig. 6B, C). Furthermore, we conducted a western blotting analysis to verify that LINC00909-OE upregulated the expression of BCL-2 and suppressed the levels of PUMA in PANC-1 cells (Fig. 6D). Inversely, downregulation of LINC00909 increased the levels of PUMA and inhibited BCL-2 expression (Fig. 6E, F). These data showed that LINC00909 repressed apoptosis in PDAC.

Fig. 6

Overexpression of LINC00909 promotes apoptosis in PDAC cells. (A) Flow cytometry was conducted to detect apoptosis after overexpression of LINC00909 in PANC-1 cells. (B, C) Flow cytometry was utilized to detect apoptosis after knockdown of LINC00909 in PANC-1 and BxPC-3 cells. (D) Western blotting was performed to examine the expression of BCL-2 and PUMA after overexpression of LINC00909 in PANC-1 cells. (E, F) Western blotting was performed to examine the expression of BCL-2 and PUMA after knockdown of LINC00909 in PANC-1 and BxPC-3 cells. P-values were assessed using two-tailed t-tests in (A–C). (***P < 0.001; ****P < 0.0001). BCL-2: B-cell lymphoma-2; LINC00909: Long intergenic non-protein coding RNA 00909; PDAC: Pancreatic ductal adenocarcinoma

LINC00909 accelerated tumor growth and metastasis in vivo

In order to investigate whether LINC00909 functions as an oncogene in vivo, PANC-1/LINC00909-OE cells and PANC-1/Vector cells were injected subcutaneously into the abdomen of BALB/c-nude mice (number = 5 per group) (Fig. 7A). The results showed an apparent increase in tumor volume as well as tumor weight in the LINC00909-OE group compared with the empty control group. The findings affirmed that LINC00909-OE enhances tumor growth in PDAC (Fig. 7B–D).

Fig. 7

Overexpression of LINC00909 markedly increases tumor burden and promotes spleen metastasis in pancreatic ductal adenocarcinoma (PDAC). (A, B) Overall appearance of the subcutaneous xenograft models in the Vector and LINC00909-OE groups (n = 5 per group). (C) Tumor volume curves of the subcutaneous xenografts in the PANC-1/Vector and PANC-1/LINC00909-OE groups (n = 5). Tumor volume was measured once every 10 days. (D) Tumor weights (n = 5). (E) Representative appearance of orthotopic xenograft models in the Vector and LINC00909-OE groups (red arrow: orthotopic pancreatic tumor; blue arrow: metastasis in the spleen). (F) The spleen metastasis amount in the orthotopic xenograft model was computed for the indicated group (n = 6 per group). (G) Pie chart revealing the percentage of mice with spleen metastasis in these two groups. (H) H&E staining of the spleen section. (I) IHC staining of SMAD4, KLF4, c-Myc, and c-JUN in orthotopic pancreatic tumors obtained from the control and LINC00909-OE groups. All *P < 0.05; **P < 0.01. P-values were assessed using ANOVA and two-tailed t-tests, followed by Dunnett’s tests for multiple comparisons in (C, D, and F)

We also aimed to mimic the clinical physiology of PDAC. Thus, we established an orthotopic xenograft model to investigate the effects of LINC00909 on tumor metastasis. By analyzing the pancreas and spleens 6 weeks after cancer cell injection, we found the incidence of spleen metastasis was 83.33% (5/6) in the LINC00909-OE group and was 16.67% (1/6) of mice in the control group, indicating that LINC00909 accelerated PDAC metastasis in vivo (Fig. 7E-H). The STRING-protein–protein interaction analysis (https://cn.string-db.org/) suggested interactions between SMAD4, KLF4, c-Myc, and c-JUN (Figs. S4A, B). Subsequently, IHC assays were conducted to detect these protein levels (Fig. 7I). As expected, the levels of SMAD4 were distinctly decreased, and the factors associated with stemness and metastasis were significantly elevated in tumor tissues obtained from LINC00909-OE mice. These results imply that LINC00909-OE enhances the pluripotency of stem cell and promotes PDAC tumor metastasis in vivo.

Discussion

To some extent, developing therapeutic strategies for PDAC depends on the discovery of novel molecules that regulate stemness and progression. Recently, accumulating evidence revealed that lncRNAs play essential roles in regulating stemness properties and cancer metastasis [17]. In this research, we found that LINC00909 was upregulated in PDAC in both TCGA + Genotype-Tissue Expression Project (TCGA + GTEx) and GDPH cohorts. Higher expression of LINC00909 was related to poorer clinicopathological features and patient survival, demonstrating it is an unfavorable biomarker in PDAC. Importantly, OE of LINC00909 induced the expression of pluripotency factors and promoted the CSC phenotype. This discovery supported a novel role of LINC00909 in regulating pancreatic CSC phenotype.

In the present research, we found that the effect of LINC00909 on pluripotency factors was dependent on the inhibition of SMAD4. SMAD4, also termed DPC4 (deleted in pancreatic cancer) [18], acts as a vital role in embryonic stem cell pluripotency, tumor progression, etc. Of note, loss of SMAD4 is a major contributor to PDAC tumorigenesis. Recently, Hoshino et al. [19] discovered that KD of SMAD4 enhanced the expression of NANOG, SOX2, and OCT4 in PC cells, indicating that SMAD4 affects cancer stemness. Consistently, we showed that OE of LINC00909 downregulated SMAD4, leading to the induction of pluripotency factors. It is well established that loss of SMAD4 in PDAC may be due to genetic mutation of deletion. However, our results indicated that the decrease or loss of SMAD4 may be caused by LINC00909. These findings revealed that impaired SMAD4 signaling could be observed in both mutated or wild-type cases, and restoration of SMAD4 was a critical strategy for eradicating CSCs and inhibiting PDAC progression.

By performing bioinformatics analysis and immunoblotting, we found that LINC00909 also activated the MAPK/JNK pathway. Similarly, Ozawa et al. [20] claimed that SMAD4 deletion was associated with activation of the MAPK/JNK pathway and resistance to cetuximab in head and neck squamous cell carcinoma. Moreover, Tan et al. [21] found that loss of SMAD4 induced the activation of p21 (RAC1)-activated kinase 3 (PAK3)-JNK-JUN by regulating miRNAs in lung cancer, thereby contributing to the development of metastasis. Hence, in addition to affecting CSCs, we hypothesized that LINC00909 may promote PC metastasis via MAPK/JNK signaling pathway, which is also due to reduction of SMAD4. In the future, it will be of interest to investigate whether MAPK/JNK activation is required for maintaining PC stemness.

It is also important to determine the mechanism through which LINC00909 downregulates SMAD4 in PC cells. The localization of lncRNAs is closely associated with their biological functions [22]. LncRNAs localized in the nucleus are associated with the regulation of chromatin and transcription [23]. LncRNAs localized in the cytoplasm may affect post-transcriptional regulation, translational regulation, and signal transduction [24]. This study showed that LINC00909 was located in the cytoplasm and downregulated the mRNA and protein expression of SMAD4, which suggests LINC00909 regulates SMAD4 at the post-transcriptional level. Specifically, we found that LINC00909 decreased the stability of SMAD4 mRNA. It will be of great significance to investigate the regulation of mRNA stability, in particular whether LINC00909 binds to SMAD4 mRNA or other RNA-binding proteins. Unfortunately, it was not possible to perform this analysis in the present study due to difficulty in designing an appropriate probe for LINC00909 with a large base size.

In addition, it is reported that LncRNAs can cross paths with exons within the protein-coding locus site on the opposite strand, antisense LncRNAs possess an intrinsic capability to control their associated sense genes at the transcriptional and/or post-transcriptional tiers [25, 26]. Similarly, Min Zhou et al. also reported that LncRNA FAM83H-AS1 has the potential to enhance FAM83H expression through the stabilization of its mRNA and then enhanced WNT/β-catenin signal pathway [27]. The alias of LINC00909 is recognized as ZNF407-AS1, which represents the antisense strand of ZNF407. However, our study did not explore the potential impact of LINC00909 on the stability of ZNF407 mRNA and whether it could consequently facilitate the cancer stemness of tumorigenicity of pancreatic ductal adenocarcinoma (PDAC). Consequently, this manuscript exhibits certain limitations, as it omits the discussion of this aspect. Future investigations can be directed towards this facet to further elucidate the prospective functions of this biological molecular marker.

We constructed an orthotopic xenografts model to validate the effect of LINC00909 on PC metastasis. Furthermore, in vitro experiments suggested that LINC00909 affected the viability of PC cells under treatment with gemcitabine, indicating that LINC00909 was associated with the development of chemo-resistance in PC. Future studies utilizing pre-clinical models should be conducted to investigate whether combined treatment with chemotherapeutic agents and inhibition of LINC00909 will exert a synergistic anti-cancer effect.

We demonstrated that LINC00909 is upregulated in PDAC and associated with inferior prognosis. LINC00909 enhances PDAC cells stemness characteristics and promotes metastasis both in vitro and in vivo. In the mechanism, LINC00909 inhibits SMAD4 expression at the post-transcriptional level, thereby upregulating the expression of stemness factors and activating the MAPK/JNK signaling pathway which is associated with metastasis (Fig. 8). Thus, LINC00909 may be a novel biomarker and target in the treatment of PC.

Fig. 8

Schematic diagram of the mechanism. By inhibiting SMAD4 at the post-transcriptional level, LINC00909 upregulates stemness factors and activates metastasis-related MAPK/JNK signaling pathways