Abstract

   Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease with
   limited treatment methods. Long non-coding RNAs (lncRNAs) have been
   found involved in tumorigenic and progression. The present study
   revealed that LINC01133, a fewly reported lncRNA, was one of 16 hub
   genes that could predict PDAC patients’ prognosis. LINC01133 was
   over-expressed in PDAC tumors compared to adjacent pancreas and could
   promote PDAC proliferation and metastasis in vitro and in vivo, as well
   as inhibit PDAC apoptosis. LINC01133 expression positively correlated
   to secreted phosphoprotein 1 (SPP1) expression, leading to an enhanced
   epithelial-mesenchymal transition (EMT) process. LINC01133 bound with
   actin-related protein 3 (Arp3), the complex reduced SPP1 mRNA
   degradation which increased SPP1 mRNA level, ultimately leading to PDAC
   proliferation. This research revealed a novel mechanism of PDAC
   development and provided a potential prognosis indicator that may
   benefit PDAC patients.

   Subject terms: Pancreatic cancer, Cancer microenvironment

Introduction

   Pancreatic cancer is a lethal disease with a five-year survival rate of
   less than 10% [[36]1]. Pancreatic ductal adenocarcinoma (PDAC) is the
   leading histological subtype and encompasses over 90% of all Pancreatic
   cancers [[37]2]. Despite the evolution of chemotherapies and targeted
   therapies, improvement in overall survival remains unsatisfactory.
   Post-operation relapse and metastasis were common in surgically treated
   patients and eventually led to fatality.

   Recently, many studies have validated the reliability of survival
   prognosis models, which incorporated gene expression or mutation
   status. Combined with traditional prediction markers such as tumor
   stage and differentiation, these models have the potential for
   customizing targeted therapy [[38]3, [39]4]. However, many of these
   models were constructed using bioinformatic techniques and the
   underlying mechanisms through which numerous key factors activate or
   suppress tumorigenesis remain largely uncharted.

   Long non-coding RNAs (lncRNAs) represent a subset of RNA that exceeds
   200 nucleotides in length, and only some lncRNAs with small open
   reading frames could encode functional polypeptides or proteins [[40]5,
   [41]6]. LncRNAs can influence both transcriptional and
   post-transcriptional regulation networks, and dysregulated lncRNAs are
   involved in many human pathological processes including cancer [[42]7].
   Many lncRNAs have been found dysregulated in association with the
   progression of PDAC through a wide array of biological processes. For
   instance, lncRNA MACC1-AS1 elevation has been linked with NOTCH1
   signaling pathway activation, resulting in unfavorable outcomes in PDAC
   [[43]8]. LncRNA PACERR polarized tissue-associated macrophages in PDAC
   by modulating the KLF12/p-AKT/c-myc pathway [[44]9]. Furthermore,
   lncRNAs HIF1A-AS1 and PVT1 have been implicated in fostering
   gemcitabine resistance of PDAC [[45]10, [46]11].

   Positioned on the genetic map at 1q23.2, LINC01133 emerges as a
   noteworthy lncRNA in recent studies. It has diverse expressions and is
   dysregulated in various cancers. Elevated LINC01133 level was observed
   in lung cancer, hepatocellular carcinoma and gynecological cancers, and
   correlated with augmented proliferation potential [[47]12].

   The epithelial-mesenchymal transition (EMT) is a cellular process in
   which epithelial cells relinquish the need for neighboring cell
   contact, adopting mesenchymal attributes. The process is associated
   with tumor initiation, proliferation, and resistance to therapy
   [[48]13]. LncRNAs were found capable of regulating EMT and therefore
   have potential as biomarkers or therapeutic targets [[49]14]. LINC01133
   was found to participate in EMT, but its mechanisms are varied, leaving
   room for further illumination [[50]15–[51]18].

   In the present study, we constructed a novel PDAC gene expression model
   to predict prognosis. This model directed our attention towards a hub
   gene, LINC01133, prompting a comprehensive exploration into its role in
   PDAC. Later, we discovered its unique interaction with secreted
   phosphoprotein 1 (SPP1) which led to an enhanced EMT process in PDAC.
   Our study further demonstrated that LINC01133, which is bound with Arp3
   and the complex regulated the downstream SPP1, could be elevated
   through activation of c-Jun in the Wnt/β-catenin pathway.

Methods

Online data collection

   PDAC RNA sequencing data, clinical and survival information of TCGA
   dataset were downloaded from by Xena database
   ([52]https://xenabrowser.net/datapages/). Additionally, two PDAC
   microarray datasets, [53]GSE32676 and [54]GSE61166, were selected and
   downloaded from the GEO database
   ([55]http://www.ncbi.nlm.nih.gov/geo/).

Construction of a PDAC prognosis prediction model

   Differential expression analysis was carried out using the ‘limma’
   package in R 4.2.2 across the 3 aforementioned datasets. Genes with 0
   expression in over 50% of samples were filtered out. Samples with a
   follow-up duration of less than 30 days were excluded. Differentially
   expressed genes (DEGs) consolidated from 3 datasets were submitted to
   univariable Cox regression analysis and lncRNA with a P value of less
   than 0.1 and met proportional hazard assumption were chosen for
   stepwise multivariable Cox regression using the ‘survival’ package in R
   to construct a prognostic model. The performance of the model was
   evaluated by the receiver operating characteristic curve and area under
   the curve using the R package “survivalROC”. The risk score was
   calculated for each patient using the model, and high-risk and low-risk
   patients were grouped according to the median risk score.

Tissue sample and clinical data collection

   Forty pairs of PDAC tumors and adjacent normal pancreas tissue were
   obtained from patients underwent radical pancreatectomy without
   pre-operation treatment. The study was approved by the Ethics Committee
   of the First Affiliated Hospital of Nanjing Medical University
   (2017-SRFA-049). Written consent was obtained from all patients and all
   procedure was conducted according to the Declaration of Helsinki.

RNA extraction and qRT-PCR

   Total RNA was extracted using TRIzol reagent (TaKaRa, Shiga, Japan).
   Reverse transcription was performed using PrimeScript^TM RT reagent Kit
   with gDNA Eraser (TaKaRa). The TB Green^TM Premix Ex Taq^TM (TaKaRa)
   was used to conduct qRT-PCR. U6 small nuclear RNA was adopted as the
   internal control. Primers were designed and synthesized by Invitrogen
   (Shanghai, China). Primer sequences were listed in Supplementary Table
   [56]S1.

Western blot analysis

   Total proteins were extracted from cells using lysis buffer (RIPA,
   Millipore, Billerica, MA, USA) and separated by sodium dodecyl
   sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Antibodies
   include Arp3, SPP1, vimentin, E-Cadherin, and N-Cadherin were purchased
   from Proteintech (Wuhan, China). β-actin (Beyotime, Shanghai, China)
   was used as an endogenous reference.

Statistical analysis and large language model usage

   R 4.1.2 was used to perform statistical analysis. Continuous variables
   between groups were compared using a two-sided Mann–Whitney U test or
   Student’s t-test depending on normal distribution status. Categorical
   variables between groups were compared using χ2 or Fisher’s exact test
   as appropriate. Kaplan–Meier method and log-rank test were used to
   determine the significance of survival analysis and median overall
   survival. P < 0.05 was considered statistically significant. During the
   preparation of this work, the authors used ChatGPT in order to improve
   language and readability. The authors reviewed and edited the content
   and took full responsibility for the content of the publication.

   Details for cell culturing, transfection, functional experiments,
   animal model construction, RNA-sequencing, immunofluorescence and
   fluorescence in situ hybridization, RNA pull-down and
   immunoprecipitation (RIP) assay, see [57]Supplementary Methods.

Results

Constructing a PDAC prognosis model with gene expression signature and model
assessment

   Differential expression analysis was performed on the TCGA,
   [58]GSE32676, and [59]GSE61166 datasets using the ‘limma’ package. 625,
   297, and 2 183 DEGs between PDAC samples and adjacent normal samples
   were identified, respectively, with a threshold of |log2 Fold
   Change | > 2 and P value < 0.05 (Fig. [60]1A). Sixty-four upregulated
   and 25 downregulated DEGs were selected after integrating the 3
   datasets using the robust rank aggregation method (Fig. [61]1B)
   [[62]19]. Eventually, 16 genes were selected out of 89 integrated DEGs
   by Cox regression, and a prognosis stratification model was constructed
   (Fig. [63]1C). Nine genes showed a P value less than 0.05 and were
   independent prognostic factors ().

Fig. 1. Constructing a PDAC prognosis model based on gene expression
signature.

   [64]Fig. 1
   [65]Open in a new tab

   A Volcano plot of differentially expressed genes (DEGs) in datasets
   [66]GSE32676, [67]GSE61166, and TCGA database. B Integrated DEGs from
   the 3 datasets using the robust rank aggregation method. C
   Identification of 16 genes included in the prognostic model through
   stepwise multivariable Cox regression. D The Kaplan-Meier curve
   illustrated the overall survival in high and low-risk groups. E The
   receiver operating characteristic (ROC) curve of the risk model. F The
   Kaplan-Meier curve of the TCGA cohort stratified according to LINC01133
   expression levels.

   The TCGA cohort was divided into high/low-risk groups based on the risk
   score generated from the prognosis model. Kaplan–Meier analysis
   revealed a significantly shortened survival time in the high-risk group
   compared to the low-risk group (Fig. [68]1D). The AUC of ROC at 5-year
   and 10-year were 0.804 and 0.829, respectively, proving the model’s
   robustness in prognosis prediction (Fig. [69]1E).

   Comparing the clinicopathological features between different risk
   groups found patient’s age, gender, evaluated lymph nodes and positive
   lymph nodes did not differ between groups (Supplementary Table [70]S2).
   The high-risk group contained more higher-grade tumors (P = 0.03).

LINC01133 was upregulated in PDAC tumor tissue and cells

   Among the 16 genes included in the final model, we chose LINC01133 for
   further experiments, as its expression level was steadily upregulated
   in PDAC in three datasets. LINC01133 over-expression was an independent
   risk factor for PDAC and a shorter overall survival time was observed
   in patients with higher LINC01133 expression levels (P = 0.02, Fig.
   [71]1F). LINC01133 expression was validated in 42 pairs of PDAC patient
   samples, and LINC01133 level was higher in PDAC tissues compared to
   matched adjacent normal tissue (Fig. [72]2A). LINC01133 was highly
   expressed in BXPC-3 and ASPC-1, and a relatively low LINC01133
   expression was observed in PANC-1, CFPAC-1, and MIACAPA-2 by qRT-PCR
   analysis (Fig. [73]2B).

Fig. 2. LINC01133 promoted PDAC cell proliferation, migration, and resistance
to apoptosis in vitro.

   [74]Fig. 2
   [75]Open in a new tab

   A LINC01133 expression level in PDAC tissues compared to matched
   adjacent normal tissue. B LINC01133 expression in PANC-1, MIACAPA-2,
   CFPAC-1, ASPC-1 and BXPC-3 cell lines. C-D Si-LINC01133 1# and 3#
   transfection could effectively knockdown LINC01133 expression in BXPC-3
   and ASPC-1 cell lines, while pcDNA-LINC01133 transfection could
   over-express LINC01133 in PANC-1. E-G CCK-8, EdU, and colony formation
   analysis performed on BXPC-3, ASPC-1, and PANC-1 cells, transfected
   with si-LINC01133 and pcDNA-LINC01133 respectively, to determine cell
   proliferation ability. H Transwell analysis on BXPC-3, ASPC-1, and
   PANC-1 cells to determine cell migration ability after transfected with
   si-LINC01133 and pcDNA-LINC01133, respectively. I Flow cytometry
   analysis was used to determine the apoptosis rate of ASPC-1 and PANC-1
   cells after transfected with si-LINC01133 and pcDNA-LINC01133,
   respectively.

LINC01133 promoted PDAC proliferation, migration, and resistance to apoptosis
in vitro

   Three small interfering RNAs designed to silence LINC01133
   (si-LINC01133 1#, si-LINC01133 2#, and si-LINC01133 3#) were
   transfected in BXPC-3 and ASPC-1 cells and transfection efficiency was
   verified by qRT-PCR analysis. Si-LINC01133 1# and si-LINC01133 3#
   displayed a more effective knockdown effect and were used for
   subsequent functional experiments (Fig. [76]2C). PANC-1 cells were
   transfected with LINC01133 over-expression vector (pcDNA-LINC01133),
   and the transfection efficiency was also confirmed by qRT-PCR analysis
   (Fig. [77]2D).

   We then explored the function of LINC01133 in vitro. CCK-8 and EdU
   analysis revealed a decrease in PDAC cell proliferation after LINC01133
   knockdown and an increase after over-expression of LINC01133 (Fig.
   [78]2E, F). Colony formation analysis showed a decrease in PDAC cell
   colonies after knocking down LINC01133, meanwhile, colonies increased
   in LINC01133 over expressed PDAC cells (Fig. [79]2G). An increased
   penetration through the basement membrane was observed in LINC01133
   overexpression cells during transwell analysis, while downregulated
   LINC01133 significantly suppressed PDAC cell migration (Fig. [80]2H).

   Flow cytometry analysis found LINC01133 knockdown could induce cell
   cycle arrest at the S phase in BXPC-3 cell line, but not in ASPC-1 and
   PDAC cell line (Supplementary Fig. [81]S1). An increase in early and
   late apoptosis in LINC01133 knockdown cells suggested that LINC01133
   may lead to PDAC cells resisting apoptosis (Fig. [82]2I).

LINC01133 promoted PDAC tumor growth and metastasis in vivo

   Since LINC01133 could promote PDAC proliferation in cell lines, we
   thought to investigate the functional role of LINC01133 in vivo. Nude
   mice xenograft models were constructed using a BXPC-3 cell line. Tumor
   volume was measured every 2 days (Fig. [83]3A). Before the tumor
   reached 1.5 centimeters in maximum diameter, the mice were euthanized,
   and tumors were harvested. The si-LINC01133 1# group showed decreased
   tumor size and weight compared to the si-NC group (Fig. [84]3B, C). The
   qRT-PCR analysis confirmed LINC01133 expression was lower in the
   si-LINC01133 1# group (Fig. [85]3D). Immunohistochemistry staining
   showed a relatively lower expression of proliferation index Ki-67 in
   the si-LINC01133 group than in the si-NC group (Fig. [86]3E).

Fig. 3. LINC01133 promoted PDAC tumor growth and metastasis in vivo.

   [87]Fig. 3
   [88]Open in a new tab

   A Gross image of nude mice injected with BXPC-3 cells transfected with
   NC and si-LINC01133 1# and resected tumors. B Tumor volume of NC and
   si-LINC01133 1# group measured every 6 days. C Mean tumor weight and
   SD. D Relative expression of LINC01133 in xenograft tumors between 2
   groups. E H&E staining and IHC staining of ki-67 of the xenograft
   tumors. Magnification, ×400. F Lung metastasis tumor nodule stained
   with H&E and IHC using ki-67 antibody. G Metastasis nodule in the heart
   of 1 mouse in NC group, grossly and microscopically. NC: negative
   control. IHC, immunohistochemistry. Points, mean; bars, SD. *P < 0.05,
   **P < 0.01.

   Tail vein metastasis models found metastatic tumor nodules in the lung
   in both groups (Fig. [89]3F), while the number of metastatic nodules
   and calculated size were higher in the NC group than in the
   si-LINC01133 group (42 vs 2, 1.34 mm^2 vs 0.64 mm^2, respectively). We
   also detected a metastatic lesion in the heart of a mouse in the NC
   group which died 12 h before the scheduled ending (Fig. [90]3G), while
   the si-LINC01133 group had no metastasis lesion in organs other than
   the lung. These results indicated that LINC01133 could promote tumor
   growth and metastasis in vivo.

LINC01133 mediating genes were enriched in the EMT pathway and participated
in the PDAC EMT process

   To determine the downstream mechanism of LINC01133 in PDAC development,
   transcriptome sequencing was performed on BXPC-3 cells transfected with
   si-LINC01133 1# and si-NC. 108 genes were differentially expressed
   between the control and LINC01133 knockdown group, and enrichment
   analysis showed that these genes were mostly enriched in extracellular
   space, extracellular region, cornified envelope, and endoplasmic
   reticulum lumen pathways (Fig. [91]4A, B). The gene set enrichment
   analysis hallmark showed that the DEGs were enriched in the
   epithelial-mesenchymal transition pathway (Fig. [92]4C). Previous
   studies have demonstrated that LINC01133 participated in the EMT
   process in colorectal cancer and PDAC tumor-derived exosomal [[93]17,
   [94]18]. Combined with our bioinformatic analysis results, we suspected
   LINC01133 could promote the EMT process in PDAC.

Fig. 4. LINC01133 mediating genes were enriched in the EMT pathway and
participated in the PDAC EMT process.

   [95]Fig. 4
   [96]Open in a new tab

   A mean-centered, hierarchical clustering of RNA expression heatmap of
   BXPC-3 cells transfected with NC and si-LINC01133 with 3 repeats. B
   Bubble plot of enriched pathways in the cellular process using Gene
   Ontology analysis. C GSEA-hallmark enrichment plot of the EMT pathway.
   D Western blot assay and relative protein level of EMT markers
   E-cadherin, N-cadherin, and vimentin in LINC01133 knockdown and
   over-expressed groups, with control. GSEA, gene set enrichment
   analysis. *P < 0.05, **P < 0.01.

   To verify our hypothesis, western blot analysis was performed which
   showed an increase in the epithelial cell marker E-cadherin and a
   decrease in the mesenchymal cell markers N-cadherin and vimentin after
   LINC01133 knockdown. Subsequently, a decrease in E-cadherin and an
   increase in N-cadherin and vimentin was observed after over-expressing
   LINC01133, confirming that LINC01133 could enhance tumor EMT in PDAC
   (Fig. [97]4D).

LINC01133 promoted tumor migration and EMT by targeting SPP1

   To validate the sequencing results, 10 genes with the highest log2 Fold
   Change score were submitted for qRT-PCR analysis in ASPC-1 and BXPC-3
   cells transfected with s-LINC01133 1# and si-NC. SPP1, LOXL4, and
   ALDH3A1 expression significantly decreased in LINC01133 knockdown of
   both cell lines (Fig. [98]5A). SLURP2, HPGD, LOXL2, CSF2, CCL3, ZBTB32,
   and CCL3L1 expression was not significantly differed after LINC01133
   knockdown in one or two cell lines. We found that SPP1 downregulation
   was most prominent after LINC01133 knockdown in both BXPC-3 and ASPC-1
   cells, and submitted SPP1 for further analysis.

Fig. 5. SPP1 was the downstream factor of LINC01133 and could promote tumor
migration and EMT.

   [99]Fig. 5
   [100]Open in a new tab

   A Relative expression of 10 differentially expressed genes selected by
   RNA sequencing in PDAC cell lines after LINC01133 knockdown. B Relative
   expression of SPP1 in PDAC tumor and adjacent tissue. C Relationship of
   SPP1 and LINC01133 expression in PDAC tumor by qRT-PCR. D SPP1 protein
   level measured by western blot after LINC01133 knockdown. Points, mean;
   bars, SD. *P < 0.05, **P < 0.01.

   To determine whether LINC01133 can regulate SPP1 expression,
   quantitative measurement and function analysis were performed. The
   qRT-PCR confirmed that SPP1 expression was positively correlated with
   LINC01133 expression (Pearson correlation coefficient r^2 = 0.154,
   P = 0.024, Fig. [101]5B). SPP1 was also significantly over-expressed in
   PDAC tissues compared to adjacent normal tissue (Fig. [102]5C). Protein
   level of SPP1 was decreased after LINC01133 knockdown (Fig. [103]5D).
   Taken together, we determined that SPP1 could be regulated by LINC01133
   and was its downstream protein.

   To investigate the role of SPP1 in PDAC tumorigenesis, si-SPP1 was
   designed to knockdown SPP1 expression, which displayed an effective
   knockdown effect and was used for subsequent functional experiments
   (Fig. [104]6A). CCK-8 and colony formation analysis showed cell lines
   proliferation with down-regulated SPP1 was decreased (Fig. [105]6B, C).
   Cell migration was also inhibited after SPP1 knockdown as shown in
   transwell analysis (Fig. [106]6D). These results showed that SPP1 had a
   tumor-promoting ability as LINC01133. We co-transfected pcDNA-LINC01133
   and si-SPP1 into ASPC-1 and BXPC-3 cells to see if SPP1 knockdown could
   reverse the tumor-promoting effect after LINC01133 over-expression.
   Down-regulation of SPP1 can partially reverse the pro-proliferation
   effect of over-expressing LINC01133 compared to cells transfected with
   negative control and pcDNA-LINC01133 (Fig. [107]6E–G). These results
   indicated that SPP1 functions as the downstream target of LINC01133 in
   PDAC.

Fig. 6. LINC01133 promoted tumor migration and EMT by targeting SPP1.

   [108]Fig. 6
   [109]Open in a new tab

   A Si-SPP1 transfection could effectively knock down SPP1 expression in
   BXPC-3 and ASPC-1 cell lines. B-C Cell proliferation ability after SPP1
   knockdown measured by CCK-8 and colony formation analysis. D Cell
   migration ability after SPP1 knockdown measured by transwell analysis.
   E-F Cell proliferation was measured by CCK-8 and colony formation
   analysis after co-transfected with si-SPP1 and pcDNA-LINC01133, with
   controls. G Cell migration was measured by transwell assay after
   co-transfected with si-SPP1 and pcDNA-LINC01133, with controls. H SPP1
   protein level of BXPC-3 and ASPC-1 after transfected with si-SPP1 and
   negative control. I Western blot assay and relative protein level of
   EMT markers E-cadherin, N-cadherin, and vimentin in LINC01133
   over-expressed and si-SPP1 groups, with controls.

   Previous studies indicated that SPP1 was an important factor in tumor
   microenvironment formation [[110]20]. So we investigated if SPP1
   participated in the EMT process. Western blot analysis showed the
   expression of E-cadherin was increased, while N-cadherin and vimentin
   expression were decreased in the si-SPP1 group, suggesting SPP1
   knockdown could suppress the EMT process in PDAC cells. (Fig. [111]6H).
   As we demonstrated before, LINC01133 over-expression promoted the EMT
   process in PDAC. We hypothesized that LINC01133-induced SPP1
   over-expression may increase PDAC proliferation by promoting the EMT
   process. Knocking down SPP1 could partially reverse the pro-EMT effect
   after over-expressing LINC01133 (Fig. [112]6I), indicating that
   LINC01133 executed its EMT-promoting function via targeting SPP1.

Arp3 bound with LINC01133 and promoted SPP1 expression

   LINC01133 was predicted located mainly in the cytoplasm using lncATLAS
   (Fig. [113]7A). Subcellular localization confirmed that LINC01133 was
   mainly located in the cytoplasm of PDAC cells (Fig. [114]7B). To
   determine the molecular mechanism of LINC01133 in PDAC, pull-down assay
   was conducted to explore the interacting protein of LINC01133. The Arp3
   protein level was higher in the biotin-labeled LINC01133 sense
   pull-down product than in the anti-sense product, showing that Arp3
   could directly bind with LINC01133 (Fig. [115]7C). We further verified
   the result using RIP assay, in which the LINC01133 was significantly
   enriched in Arp3 antibody treatment unit than in IgG control unit (Fig.
   [116]7D). IF-FISH showed that LINC01133 and Arp3 were colocalized in
   the cytoplasm of tumor cells in PDAC tissue microarray (Fig. [117]7E).

Fig. 7. Arp3 bound with LINC01133 and promoted SPP1 expression.

   [118]Fig. 7
   [119]Open in a new tab

   A LINC01133 was predicted located mainly in the cytoplasm by the online
   bioinformatic analysis website LncATLAS. B Subcellular fractionation
   location showed LINC01133 mainly located in the cytoplasm of 3 PDAC
   cell lines. C Western blot analysis after LINC01133 pull-down showed a
   higher Arp3 protein level in the sense LINC01133 pull-down product. D
   The qRT-PCR result of LINC01133 level after RIP analysis using
   anti-Arp3 antibody. E ARP3 and LINC01133 expression in PDAC tumor
   tissue by immunofluorescence and FISH. F Relative expression of ARP3
   and SPP1 in PDAC cell line BXPC-3 after ARP3 knockdown. G ARP3 and SPP1
   protein levels were measured by western blot after ARP3 knockdown. H
   ARP3 protein level measured by western blot after LINC01133 and SPP1
   knockdowns. I The RNA degradation of SPP1 in ARP3 knockdown and control
   groups. RIP, RNA immunoprecipitation. FISH, fluorescence in situ
   hybridization.

   Both Arp3 and SPP1 were involved in extracellular matrix regulation
   [[120]21, [121]22], but no research was done on their interaction
   mechanism. BXPC-3 cells were treated with si-ARP3, si-SPP1 or negative
   control, and knockdown of ARP3 led to lower expression of SPP1 on both
   RNA and protein levels in BXPC-3 cells, while SPP1 or LINC01133
   knockdown did not affect ARP3 mRNA or protein expression as shown in
   qRT-PCR and western blot (Fig. [122]7F–H). This indicated that Arp3 was
   the upstream regulator of SPP1.

LINC01133 enhanced ARP3 function which increased SPP1 mRNA stability and
enhanced the PDAC EMT process

   To investigate how Arp3 affects SPP1 expression, actinomycin D was
   administrated after ARP3 knockdown. After new RNA synthesis was blocked
   with actinomycin D, the loss of SPP1 was quantified by qRT-PCR assay.
   RNA degradation of SPP1 was significantly increased after ARP3
   knockdown (Fig. [123]7I), suggesting that LINC01133 and ARP3
   co-regulated the expression of SPP1 in PDAC by stabilizing SPP1 RNA.
   Since LINC01133 was mainly located in the cytoplasm of PDAC cells, we
   hypothesized that the LINC01133 may enhance the function of Arp3 after
   binding, leading to increased SPP1 RNA stability and eventually SPP1
   overexpression.

C-Jun activates LINC01133 transcription after β-catenin pathway activation

   Online websites hTFtargetk [[124]23], PROMO, and JASPAR [[125]24] used
   to perform promoter analysis predicted that the LINC01133 promoter
   could bind to the transcription factor Jun proto-oncogene (c-Jun) (Fig.
   [126]8A). C-Jun siRNA was transfected into the BXPC-3 cells and qRT-PCR
   analysis revealed that LINC01133 expression correlated with c-Jun and
   was significantly decreased after c-JUN knockdown (Fig. [127]8B).
   Dual-luciferase reporter assay showed that relative luciferase activity
   after c-Jun knockdown was decreased compared to control (Fig. [128]8C).
   These findings suggested that c-JUN upregulated LINC01133 expression by
   binding to its promotor region and activating LINC0113 transcription.

Fig. 8. C-Jun bound to the promoter region of LINC01133 and activates its
transcription.

   [129]Fig. 8
   [130]Open in a new tab

   A Online websites predicted that transcription factor C-Jun could bind
   to the promoter of LINC01133. B LINC01133 RNA level after c-Jun
   knockdown showed a marked correlation between the two factors. C
   Luciferase activity in BXPC-3 cells after co-transfected si-JUN with
   LINC01133 promoter-mutated or promoter-wild-type vectors. D RNA
   expression of β-catenin, c-JUN, and LINC01133 after WNT/β-catenin
   pathway activation by LiCl treatment in BXPC-3 and PANC-1.

   Because c-Jun was a reported downstream molecule of β-catenin
   [[131]25], we thought to investigate if the WNT/β-catenin pathway could
   affect LINC01133 in PDAC. After being treated with 20 mM/well WNT
   signaling agonist lithium chloride (LiCl), RNA expression of CTNNB1
   (β-catenin), c-JUN, and LINC01133 in BXPC-3 and PANC-1 were all
   increased compared to controls (Fig. [132]8D), suggesting WNT/β-catenin
   pathway activation could promote LINC01133 expression.

Discussion

   In the present study, we constructed a novel PDAC genetic risk model
   and explored the function of one of the key factors, LINC01133, and its
   role in promoting the EMT process in PDAC. LINC01133 was upregulated
   and could promote tumor cell proliferation in PDAC. Pathway enrichment
   analysis, qRT-PCR, and subsequent western blot tests revealed LINC01133
   expression positively correlated to SPP1, their over-expression led to
   enhanced EMT process in PDAC. Pull-down and RIP assays found that
   LINC01133 bound with Arp3, the complex reduced SPP1 mRNA degradation
   which increased SPP1 mRNA level, ultimately leading to PDAC
   proliferation.

   Previously, LINC01133 has emerged as an important cancer-related gene
   and a growing number of studies have gradually uncovered its role in
   cancer. LINC01133 may execute its biological function in a
   tissue-specific manner as it has diverse expressions and effects on
   tumor progression in different cancers. Its role in the EMT process is
   also varied. In gastric cancer and colorectal cancer, LINC01133 acts as
   a tumor suppressor and is downregulated in tumor cells, inhibiting the
   EMT process [[133]15, [134]17, [135]26]. LINC01133 is up-regulated in
   PDAC and is associated with PDAC tumor growth, proliferation,
   migration, metastasis, and invasion [[136]27, [137]28]. Consistent with
   previous studies, our study found LINC01133 was upregulated in PDAC
   tissue and cellular level, and LINC01133 over-expression could promote
   PDAC proliferation and migration in vivo and in vitro.

   The downstream tumor-promoting mechanism of LINC01133 was also explored
   by previous researchers. LINC01133 could sponge with microRNAs and
   promote PDAC development [[138]29, [139]30]. Our RNA sequencing result
   suggested SPP1 expression correlated with LINC01133. SPP1 is a
   protein-coding gene that produces a multi-functional matricellular
   phosphoglycoprotein, also known as osteopontin. SPP1, or osteopontin,
   was found able to guide EMT through multiple cellular signaling
   pathways and by restructuring the tumor microenvironment to modify EMT
   processes [[140]31]. SPP1 promotes fibrosis and is critically involved
   in the fibroblast to myofibroblast transformation, which participates
   in tumor proliferation, angiogenesis, and metastasis [[141]32,
   [142]33]. It also generates cancer-associated fibroblasts from normal
   fibroblasts and mesenchymal stem cells, leading to increased TGF-β and
   IL-6 secretion and inducing EMT [[143]32]. We were the first study to
   link SPP1 and EMT process with LINC01133 over-expression, demonstrating
   a novel SPP1 regulation network that may be beneficial in future
   anti-cancer therapy. Interestingly, another study on the function of
   exosomal LINC01133 found that LINC01133 could be regulated by
   periostin, which was another extracellular matrix protein secreted by
   pancreatic stellate cells, leading to PDAC EMT [[144]18]. Taken
   together, these suggested that LINC01133 could interact with various
   extracellular components and play a key role in the EMT process in
   PDAC.

   Previous studies have indicated that LINC01133 is located in both
   nuclear and cytoplasm and can execute its function in both locations
   [[145]18, [146]29]. We found LINC01133 mainly located in the cytoplasm
   of PDAC cells. RIP analysis showed that LINC01133 could bind with Arp3
   in PDAC. The RNA stability test also confirmed our hypothesis that Arp3
   inhibited SPP1 degradation. Arp3, also known as actin-related protein
   3, was a subunit of the Arp2/3 complex, the activation of which was
   reported associated with an early PDAC carcinogenesis process known as
   acinar-to-ductal metaplasia [[147]34]. A subunit of the Arp2/3 complex
   could regulate cell migration and invasion in PDAC, and a decrease in
   cell migration ability could be observed after silencing the Arp2/3
   complex [[148]35, [149]36]. Various studies have elaborated on the
   direct role Arp2/3 played in the formation of branched actin network.
   Here, we demonstrated a novel Arp3 regulatory pathway in PDAC that
   could potentially be used as a therapeutic target.

   Bioinformatic analysis inferred that c-Jun can bind to the promoter
   region of LINC01133 and promote LINC01133 transcription. Our
   experiments confirmed that c-Jun expression correlated with LINC01133.
   C-Jun is a transcription factor that is closely associated with
   tumorigenesis and tumor metastasis, and a downstream modulator of the
   WNT/β-catenin signaling pathway [[150]37]. Previous studies have
   demonstrated a positive correlation between LINC01133 and c-Jun,
   further corroborating our result [[151]38].

   In summary, we developed a novel prognostic model of PDAC based on gene
   signature and investigated the oncogenic functions of a key gene,
   LINC01133. We found that LINC01133 could bind with Arp3 forming a
   complex that could increase SPP1 RNA stability, leading to enhanced EMT
   process in PDAC. The process could be regulated by WNT/β-catenin
   pathway activation through c-JUN over-expression. Together, we
   demonstrated a novel pathway that could induce EMT in PDAC and
   hopefully could improve the understanding of the interaction between
   lncRNA and cancer for potential diagnostic and therapeutic purposes.

Supplementary information

   [152]Supplementary Materials and Methods^ (22.3KB, docx)
   [153]Supplemental legends^ (12.7KB, docx)
   [154]Supplementary figure S1^ (1.1MB, jpg)
   [155]Supplementary table 1 and 2^ (17.5KB, docx)
   [156]Original western blot data^ (530.8KB, pdf)

Acknowledgements