LINC00261 elevation inhibits angiogenesis and cell cycle progression of pancreatic cancer cells by upregulating SCP2 via targeting FOXP3

Abstract Long non‐coding RNAs (lncRNAs) biological functions and molecular mechanisms associated with pancreatic cancer (PC) remain to be poorly elucidated. We aimed to clarify the role of lncRNA LINC00261 (LINC00261) in PC and confirm its regulatory mechanisms. Bioinformatics analysis, RNA pull‐down and RIP assays were performed to investigate relationship between LINC00261 and forkhead box P3 (FOXP3). Further, dual‐luciferase reporter gene and ChIP assays were employed to confirm the relationship among LINC00261, FOXP3 and sterol carrier protein‐2 (SCP2). PC cells were introduced with a series of vectors to verify the effects of LINC00261 and SCP2 on the viability, cell cycle progression, migration and angiogenesis of PC cells. Nude mice with the xenograft tumour were used to evaluate the effects LINC00261 on the tumourigenicity. LINC00261 was lowly expressed in PC tissues and cells. SCP2 was inhibited by LINC00261 through FOXP3. Functionally, upregulated LINC00261 or downregulated SCP2 led to reduced cell viability, migration, angiogenesis and tumourigenicity potentials. This study demonstrated the inhibitory role of LINC00261 in the angiogenesis and cell cycle progression of PC cells. It acts through the negative regulation of SCP2 via targeting FOXP3. Findings in this study highlight a potentially biomarker for PC treatment.


| Bioinformatic prediction
The PC-related gene expression profiles (GSE16515 and GSE27890) and annotated probe files were downloaded from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm. nih.gov/geo) followed by analysis through the Affymetrix Human Genome U133 Plus 2.0 Array. GSE16515 includes 16 normal samples and 36 tumour samples, and GSE27890 includes 4 normal samples and 4 tumour samples. With the normal samples as controls, we used the Affy installation package of the R software for background correction and normalization of the microarray data. 18 Then, we applied the linear model, empirical Bayesian statistics method of the Limma installation package combined with a traditional t-test for non-specific screening of data to identify differentially expressed lncRNAs 19

| Sample collection
Pancreatic cancer tissues were collected from 57 PC patients (34 males and 23 females, average age: 58 years) who underwent pancreatectomy in the Affiliated Hospital of Jiaxing University from January 2014 to December 2017. Adjacent normal tissues were collected as the controls. Among all patients, 9 cases had tumour size ≤2 cm, 21 had tumour size 2-4 cm, and 27 cases had tumour size >4 cm. According to the tumour node metastasis (TNM) stages, 22 18 cases were classified to stage I, 27 cases were stage II, 12 cases were stage III, and 30 cases were diagnosed with lymph node metastasis (LNM). A serial section from each specimen was stained with haematoxylin & eosin staining for histological evaluation. 23 Histological examination showed that 24 cases were highly differentiated, 20 cases were moderately differentiated, and 13 cases were poorly differentiated. Besides, 24 of the 57 cases were diagnosed with venous invasion and 33 cases without venous invasion. The patients meeting following criteria were included: (1) patients with PC were confirmed both by tumour pathology and genetics; (2) tumours were clinically evaluated to be amenable to surgical resection; (3) patients were fit to tolerate abdominal surgery; (4) patients didn't present with complications of other gastrointestinal tumours or history of other tumours; (5) chemotherapy or radiotherapy had not been performed before surgery. Patients with severe impairment in heart, liver, kidney and other organs, a history of autoimmune disease, in the active stage of a chronic infectious disease or diagnosis of an acute infectious disease were excluded. The serum of PC patients and 57 healthy people was collected, where the expression of LINC00261 in the serum was detected by enzyme-linked immunesorbent assay (ELISA).

| ELISA
According to instructions, Quantikine ELISA Human Galectin-9 immunoassay kit (DGAL90, R&D Systems) was used to determine LINC00261 expression in serum of PC patients and healthy donors.

| Western blot analysis
Western blot analysis was performed with following antibodies (all from Abcam) 24 : anti-rabbit antibodies to FOXP3 (ab215206), SCP2 (ab140126), vascular endothelial growth factor (VEGF; ab10766), Ki67 (ab92742), CyclinD1 (ab16663) and Brachyury (ab209665) as well as horseradish peroxidase (HRP)-labelled goat anti-rabbit antibody (1: 5000). After washing, enhanced chemiluminescence was used to develop images. The optical density (OD) value of the protein bands was measured by a gel imaging analysis system with relative protein content as the ratio of average OD value of the target to that of the internal control.

| Cell culture and transfection
Human pancreatic ductal epithelial cell line H6c7 and human PC cell lines SW1990 and PANC-1 (American Type Culture Collection) were cultured in RPMI 1640 culture medium (22400089, Gibco BRL.) containing 15% foetal bovine serum (FBS) and seeded into 6-well plates (2.5 × 10 5 /well) for incubation. The culture medium was renewed every 2-3 days based on the growing states of cells. Cells were subcultured upon 80%-90% cell confluence.
Cells (logarithmic growth phase) were seeded into a 6-well cell plate (2 × 10 5 cells/well) for further culture and transfected using a Lipofectamine 2000 kit (Invitrogen) upon 90% cell confluence.

| Dual-luciferase reporter gene assay
Binding site of FOXP3 in promoter region of SCP2 was investigated by Website analysis prediction. The SCP2 promoter region was constructed into the pGL3-Basic vector (Promega) as a recombinant vector SCP2-wild type (WT), and the FOXP3 binding site of SCP2 was mutated into the pGL3-Basic vector (Promega) as a recombinant vector SCP2-mutant (MUT). The sequenced luciferase reporter plasmids SCP2-WT and SCP2-MUT were co-transfected into HEK-293T cells with sh-NC, sh-FOXP3, oe-NC and oe-FOXP3, respectively.

| Fluorescence in situ hybridization (FISH)
Fluorescence in situ hybridization was performed as per the instructions of Ribo™ lncRNA FISH Probe Mix (Red) (RiboBio). Five randomly selected fields of view were photographed under a fluorescence microscope (Olympus).

| RNA binding protein immunoprecipitation (RIP) assay
The binding of LINC00261 to the transcription factor FOXP3 was detected using a RIP kit (Millipore) 25 with the following antibodies: anti-FOXP3 (sc-53876, 2 µg per 1 ml of cell lysate, Santa Cruz Biotechnology) and Immunoglobulin G (IgG; ab6785, Abcam, used as NC).

| Chromatin Immunoprecipitation (ChIP) assay
SW1990 cells were fixed with formaldehyde (10 min) to generate DNA-protein cross-linking, followed by ultrasonic disruption. Cells were centrifuged (4℃; 10 min) under 12,000 g, and the collected supernatant was divided into two tubes and incubated with IgG and anti-FOXP3 (sc-53876, 2 µg per 1 ml of cell lysate, Santa Cruz) overnight at 4℃, respectively. The protein-protein complexes were precipitated with Protein Agarose/Sepharose. After centrifugation for 5 min, the supernatant was discarded and the non-specific complexes were washed. De-crosslinking was then performed at 65℃ overnight. DNA fragment was recovered by extraction and purification with phenol/chloroform. Binding in SCP2 promoters were detected by RT-qPCR.

| Flow cytometry
At 48th hour after transfection, cells were trypsinized to adjust the concentration to 1 × 10 6 cells/ml. Then, 1 ml of cells was centrifuged at 1200 g for 10 min. Following supernatant removal, pellets were resuspended in PBS, centrifuged and fixed with pre-cooled 70% ethanol at 4℃ for overnight incubation. After PBS rinsing, 100 μl of cell suspension was incubated with 50 μg of propidium iodide staining solution (40710ES03, Qcbio) containing RNAase for 30 min in the dark. After the cell suspension was filtered by a 100-mesh nylon net, the cell cycle was measured by a flow cytometer (Becton, Dickinson and Company) at the excitation wavelength of 488 nm.

Human umbilical vein endothelial cells (HUVECs; INS-1, Shanghai
Zishi Biological Technology Co., Ltd.) were cultured with SW1990 cell culture solution in a 24-well plate (250 μl/well) pre-coated with pre-chilled Matrigel. Then, the cells were detached with 1 ml trypsin containing EDTA, and the detachment was stopped by addition of serum-containing culture solution. After counting, the cells were resuspended and adjusted (1 × 10 5 cells/ml). After matrigel were solidified, the 24-well plate was removed, followed by the addition of 500 μl/well of the cell suspension. Afterwards, the plate was incubated for 24 h at 37℃ with 5% CO 2 and observed under an inverted microscope. Mean number of tubes formed was calculated in three randomly selected fields.

| Transwell assay
Migration assay was assessed by 24-well plate as described previously. 26 The plates were photographed under an inverted microscope, and the mean number of invaded cells was recorded in 10 randomly selected fields.
After inoculation, mice were reared under standard conditions, and tumour length and width were measured every 7 days. Size, length (L) and width (W) of subcutaneous transplanted tumour model were measured using Vernier callipers. The volume of tumour was calculated as length × width 2 /2. All mice were euthanized at the 28th day, and the tumour was excised. The excised tumour was then fixed in 10% neutral formalin solution (24 h), dehydrated, embedded and stored for future examination.

| Immunohistochemistry (IHC)
Paraffin-embedded tumour was cut into sections at the thickness of 3 ~ 4 μm. The sections were dewaxed, hydrated, placed in 3% H 2 O 2 for 10 min and subjected to high-pressure antigen retrieval for 90 s. Then, the sections were blocked with 5% bovine serum albumin

| Statistical analysis
All data were presented by mean ± standard deviation. SPSS 21.0 software (IBM Corp.) was used for data analysis. Comparison between two groups was analysed by unpaired t-test. Comparisons among multiple groups were analysed using one-way analysis of variance (ANOVA). p < 0.05 was considered statistically significant.

| Downregulation of LINC00261 occurs in PC tissues
After screening PC-related datasets GSE16515 and GSE27890, we found poor expression of LINC00261 ( Figure 1A,B). RT-qPCR showed a notable reduction in the expression of LINC00261 in PC tissues and cells ( Figure 1C). ELISA revealed that the expression of LINC00261 was significantly reduced in the serum of PC patients ( Figure S1). Next, RT-qPCR was performed to measure the expression of LINC00261 in the normal pancreatic ductal epithelial cell line H6c7 and two PC cell lines (SW1990 and PANC-1), which demonstrated that downregulation of LINC00261 also occurred in SW1990 and PANC-1 cells ( Figure 1D).
Moreover, relationships between LINC00261 expression and relevant clinicopathological characteristics of PC patients were explored, revealing that LINC00261 expression had significant correlation with differentiation degree, LNM, venous invasion and TNM stages, whereas no correlation was observed for the gender, age, tumour sites and volume (all p > 0.05) (Table S2).

| LINC00261 overexpression represses oncogenic phenotype of PC cells
To explore the effect of LINC00261 on the biological features of

| LINC00261 regulates the expression of SCP2 through FOXP3
To determine the mechanism by which LINC00261 may be involved in the development of PC, its corresponding localization was observed by FISH assay, which showed most of the LINC00261 and FOXP3 expression was localized in the nucleus of PC cells, and the rest scattered in the cytoplasm ( Figure 3A). Then, as predicted by the MEM database, we found that SCP2 was most likely regulated by LINC00261 ( Figure 3B). Besides, SCP2 is involved in PPAR pathway ( Figure 3C), which has been proven to exert important regulatory functions in PC. 26,27 Thus, SCP2 may be regulated by LINC00261 in PC. Moreover, a role of FOXP3 in pancreatic ductal adenocarcinoma has been reported. 15 Therefore, we selected FOXP3 and SCP2 for subsequent analysis. Subsequently, the sequence logo of the transcription factor FOXP3 was further analysed through the Jaspar   website (http://jaspar.gener eg.net/) as shown in Figure 3D, and the binding site of the target gene SCP2 promoter region was further obtained (Table S3).

| LINC00261 represses angiogenesis and cell growth in PC through FOXP3-madiated SCP2
We next investigated whether the LINC00261/FOXP3/SCP2 axis participated in angiogenesis and cell cycle regulation in PC. Initial RT-qPCR and IHC studies showed that SCP2 level was reduced in PC tissues ( Figure 4A), and Pearson correlation analysis identified a positive relationship between LINC00261 and SCP2 ( Figure 4B). Additionally, elevation of LINC00261 led to elevation of SCP2 expression, which could be reversed by the addition of oe-FOXP3 or sh-SCP2 ( Figure 4C).

| LINC00261 restrains the tumour growth in vivo
We constructed a stably transfected PC cell line PANC-1 expressing oe-LINC00261, sh-LINC00261, oe-FOXP3, sh-SCP2 or Vector and subcutaneously injected into a tumour model in mice. As depicted in Figure 5A,B and Figure

| DISCUSS ION
Angiogenesis, which is a key factor in several biological processes related to development and progression of cancer, may be regulated by lncRNAs. 28 We investigated the effects of LINC00261 on the biological process and angiogenesis of PC cells and provided evidence that LINC00261 enhanced SCP2 expression by suppressing the transcription factor FOXP3, thereby inhibiting the angiogenesis and cell cycle change in PC.
We confirmed decreased expression of LINC00261 in PC tissues, as has been reported in a previous study. 29 In the present study, LINC00261 was further demonstrated to exert an inhibitory role in the ability of proliferation, migration, angiogenesis and tumour formation of PC cells. LINC00261 functions as a tumour suppressor in PC through the miR-222-3p/HIPK2/ERK axis. 30   has been demonstrated in relation to the mechanism concerning the GATA6-mediated DKK3. 34 These findings are in accord with results of our study, thus indicating that LINC00261 is downregulated in several carcinomas and that its overexpression may play a very important role specifically in PC, thus highlighting LINC00261 as a potential molecular diagnostic marker for target for cancer therapy.
Furthermore, we validated FOXP3 as a target gene of LINC00261, which could be directly inhibited by LINC00261, based on present results from PC-related datasets and the luciferase assay. FOXP3 is supposed to play a crucial role in the process of regulation of T cell differentiation. 35 Also, previous work suggests the role of FOXP3 in tumour immunity in cancer patients. 36 The lymphocyte density of FOXP3 + was observed to be related to LNM and further implicated with the development of PC. 37 Meanwhile, it has been reported that FOXP3 was regulated by lncRNA EGFR-AS1 and thus is involved in the regulation of NSCLC progression. 17 Moreover, LINC00261 overexpression decreased the expression of FOXO3 to inhibit the oncogenic phenotype of PC cells in regulation of miR-552-5p. 12 Interestingly, our study showed that LINC00261 could upregulate SCP2 expression by suppressing the expression of FOXP3. As indicated by previous research, SCP2 is able to regulate angiogenesis and tumour migration. 14 Notably, SCP2 exerts function on vascular inflammation of systemic connective tissue diseases in rats. 38 Importantly, SCP2 may be especially associated with diseases involving abnormalities in lipid trafficking. 39 As is widely appreciated, PC is aggressive with dismal prognosis due to its propensity for widespread metastatic disease. 40 In addition, several research projects have reported an increased de novo lipogenesis in PC cells. 41,42 Therefore, SCP2 may be involved in the regulation of lipid trafficking in PC cells.
In conclusion, we identified that LINC00261 regulates SCP2 expression in PC by suppressing FOXP3, thus inhibiting angiogenesis and cell cycle progression ( Figure 6). Our findings present potential clinical therapeutic target for PC treatment, which calls for broader investigations in the clinical setting. However, more samples should be included to better validate the current findings.

ACK N OWLED G EM ENTS
Not applicable.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no conflicts of interest.