High expression of AFAP1-AS1 is associated with poor survival and short-term recurrence in pancreatic ductal adenocarcinoma

Pancreatic ductal adenocarcinoma (PDAC) is still a lethal malignancy. Long noncoding RNAs (lncRNAs) have been shown to play a critical role in cancer development and progression. Here we identified overexpression of the lncRNA AFAP1-AS1 in PDAC patients and evaluated its prognostic and functional relevance. The global lncRNA expression profile in PDAC was measured by lncRNA microarray. Expression of AFAP1-AS1 was evaluated by reverse-transcriptase quantitative polymerase chain reaction (RT-qPCR) in 90 PDAC tissue samples and adjacent normal tissues. The impact of AFAP1-AS1 expression on cell proliferation, migration, and invasion were evaluated in vitro using knockdown and ectopic expression strategies. Microarray analysis revealed that up-regulation of AFAP1-AS1 expression in PDAC tissues compared with normal adjacent tissues, which was confirmed by RT-qPCR in 69/90 cases (76.7%). Its overexpression was associated with lymph node metastasis, perineural invasion, and poor survival. When using AFAP1-AS1 as a prognostic marker, the areas under ROC curves were 0.8669 and 0.9370 for predicting tumor progression within 6 months and 1 year, respectively. In vitro functional experiments involving knockdown of AFAP1-AS1 resulted in attenuated PDAC cell proliferation, migration, and invasion. Ectopic expression of AFAP1-AS1 promoted cell proliferation, migration, and invasion. AFAP1-AS1 is a potential novel prognostic marker to predict the clinical outcome of PDAC patients after surgery and may be a rational target for therapy.


Background
Pancreatic ductal adenocarcinoma (PDAC) remains one of the most aggressive human cancers [1]. Despite substantial efforts, PDAC is associated with a short survival that has been declining steadily since the early 1990s [2]. PDAC is characterized by a highly malignant phenotype that is associated with early metastasis and resistance to chemotherapy and radiation therapy [3]. There is an urgent need for further understanding of the mechanism of PDAC development and new innovative therapeutic approaches. Identifying the underlying molecular mechanisms of invasion and metastasis in PDAC will be essential for the identification of effective drug targets.
In recent years, it has become increasingly apparent that the noncoding portion of the genome is of crucial functional importance in both normal physiology and diseases [4]. Long noncoding RNAs (lncRNAs), which are defined as those longer than~200 nucleotides but lacking protein coding capacity [3], have recently been shown to play a key role in regulating vital cellular functions including cancer progression [5]. To date, thousands of lncRNAs have been discovered through chromatin signature analysis and large-scale sequencing, and functional studies have shown that many of them exhibit diverse biological functions and have clinical significance [6]. Importantly, many lncRNAs have been identified as being cancer-specific [5,7]. For example, aberrant expression of lncRNA HOTAIR was associated with various cancers such as breast, hepatocellular, gastric, colorectal, and pancreatic, and its expression was associated with survival and prognosis of cancer patients [8]. MALAT1 was discovered as a prognostic marker for lung cancer metastasis but also been linked to several other human malignancies [9]. Other examples include HULC in hepatocellular carcinoma [10] and PCGEM1 in prostate cancer [11,12]. In pancreatic, a number of lncRNAs were found to exhibit pro-oncogenic or tumor-suppressive activities, such as ENST00000480739 [13], LOC285194 [13], HULC [14], HOTAIR [15], and MALAT1 [16], suggesting an important of lncRNAs in progression of pancreatic cancer. Therefore, identification of lncRNAs involved in PDAC progression might help yield novel prognostic biomarkers or therapeutic targets.
In this study, we observed that a lncRNA, AFAP1-AS1, was substantially overexpressed in PDAC tissues. Knockdown of AFAP1-AS1 could inhibit cell proliferation, migration, and invasion of PDAC cells. Moreover, AFAP1-AS1 expression correlated with lymph node metastasis, perineural invasion, and poor survival in PDAC patients. Our results suggest that AFAP1-SA1 may represent a novel indicator of poor prognosis and a potential therapeutic target in PDAC.
RNA Isolation, quantitative real-time reverse-transcription polymerase chain reaction (PCR), and microarrays Quantitative real-time PCR (RT-qPCR) was performed for AFAP1-AS1 and the epithelial-mesenchymal transition (EMT) markers E-cadherin, N-cadherin, Vimentin, Snai1, and Slug. β-Actin was used as an internal control. RNA was extracted from frozen pancreatic cancer tissues and their corresponding non-neoplastic tissues and pancreatic cell lines using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The total RNA was then converted to cDNA by reverse-transcription using oligodT primers and Super-Script II reverse transcriptase (Invitrogen). For real-time quantitative PCR, three replicates of each sample were amplified and analyzed using a Roche Light-Cycler (Roche, Basel, Switzerland). The 20 μl reaction mixtures contained SYBR Green reaction mix (Qiagen Sciences) and 0.5 mM of primer. Relative gene expression was determined using the comparative delta-delta CT method (2-ΔΔCt). The primer sequences for PCR were provided in the supplementary materials (Additional file 1).
Transcriptomic analysis was performed using Arraystar human lncRNA microarrays, V3 (Agilent, USA), which target 27958 Entrez protein-coding genes and 7419 lncRNAs. Total RNA was extracted and mRNA was purified using the mRNA-ONLYTM Eukaryotic mRNA Isolation Kit (Epicentre). Total RNA was fragmented, labeled (One-Color, Cy3, Agilent), purified, and hybridized with probes in Hybridization Chamber gasket slides (Agilent). The slides were then washed and scanned with an Agilent Microarray Scanner. The raw data were extracted with Agilent Feature Extraction software (Agilent Technology). This software uses the robust multi-array average algorithm to adjust the background signals. Normalized data were obtained after performing the quantile method of intra-microarray normalization and the median method of baseline transformation between the microarrays. Differentially expressed genes with a raw expression level of >400 in more than 4 out of the 12 samples used for profiling were extracted and ordered by p-value. Genes with the highest top 10 p-values were selected for validation. The microarray platform and data were submitted to the Gene Expression Omnibus public database at the National Center for Biotechnology Information (accession number: GSE61166, http://www.ncbi.nlm.nih.gov/ geo/query/acc.cgi?acc=GSE61166).
Horseradish peroxidase-conjugated secondary antibodies (Cell Signaling Technology), and an ECL chemiluminescence kit (Pierce) were used to detect bound antibody.

Cell growth and cell cycle assays
Cell proliferation analysis 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) substrate (Sigma-Aldrich) was used to assay cell proliferation according to the manufacturer's instructions. Briefly, a total of 3 × 10 3 cells were seeded into 96-well dishes and allowed to adhere overnight. The growth curves of cells, covering a total of 3 days of culturing, were determined through measuring absorbance at 570 nm. After transfection, 5 × 10 4 SW1990, MIAPaCa-2 cells, or PANC-1 cells were collected and washed three times with PBS. After RNase digestion and PI dyeing, the cells were subjected to FACS analysis.

Migration and invasion assays
The cell migration assay was performed using BD Transwell chambers. Cell invasion assays were performed with chambers uniformly covered with Matrigel (BD Biosciences) diluted with Dulbecco's modified Eagle's medium (DMEM) to a certain percentage and incubated at 37°C for 30 minutes. Cells (1 × 10 6 ) were suspended in 200 μl serum-free DMEM medium and seeded into the upper chamber of each insert. Then, 600 μl of DMEM containing 10% FBS was added to a 24-well plate. After incubation at 37°C for 12 h, the cells that migrated were fixed and stained for 20 min in a dye solution containing 0.4% crystal violet and 20% methanol. The cells in the upper layer of the membrane were removed and the cells in the lower layer were washed off with 33% acetic acid (500 μl per chamber). The migrated cells was quantified by measuring the absorbance of eluent at 570 nm. The relative migration fold change of the experimental group was calculated by normalizing to that of the control group.

Patient samples
All samples were obtained from patients when undergoing resection of the pancreas at Sun Yat-Sen Memorial Hospital between 2009 and 2014. Informed consent was obtained from the patients before sample collection. All patients had a clear histological diagnosis. Patients' specimens and the related clinicopathological data, including complete follow-up, were obtained from the Institute of Pathology and from the Department of Pancreaticobiliary, Sun Yat-Sen Memorial Hospital. All patients in this study met the following criteria: 1) PDAC diagnosis was verified by pathological examination; 2) paraffinembedded tissues were well stored and qualified for serial section; 3) the corresponding tumor tissues and the paired non-tumor tissues were stored in liquid nitrogen immediately following surgical removal; 4) no anticancer treatments given before biopsy collection; and 5) availability of exhaustive clinicopathologic and follow-up data.

Tumorigenicity assays in nude mice
All experiments involving animals were conducted according to the institutional guidelines of Guangdong Province and were approved by the institutional guidelines of Guangdong Province and by the Use Committee for Animal Care. BALB/c nude mice (5 weeks old) were randomly separated into the shControl group or the shAFAP1-AS1 group (5 mice per group). SW1990 cells (3 × 10 6 cells/ mouse) stably transfected with shAFAP1-AS1 or control shRNA were injected subcutaneously into the right axilla of each mouse. Tumor volume was calculated using the following formula: volume = (L × W 2 )/2, where L and W are the longest and shortest diameters, respectively. The mice were sacrificed two weeks after injection.

Statistical analysis
Statistical analyses were performed using SPSS Statistics 17.0 (SPSS lnc®). The chi-square test (X 2 test), Fisher's exact test for nonparametric variables, and Student's t test for parametric variables (two-tailed) were used. Differences in patient survival were assessed using the Kaplan-Meier method and analyzed using the log-rank test in univariate analysis. All tests were two tailed, and results with P = 0.05 were considered statistically significant.

AFAP1-AS1 is aberrantly overexpressed in human PDAC cell lines and cancerous tissues
As a first attempt to identify differentially expressed long noncoding RNAs (lncRNAs) in two subtypes of PDAC tissues (PDAC patients with diabetes versus PDAC patients without diabetes), we conducted microarray analysis utilizing a microarray targeting 7419 LncRNAs using eight cases of PDAC tissues and four cases of chronic pancreatitis tissues (CP) (accession number: GSE61166). All of the differentially expressed (> or <2 fold change) lncRNAs were listed in Additional file 2. We noticed that the long noncoding RNA AFAP1-AS1 is one of the most up-regulated lncRNAs in both subtypes of PDAC tissues (Figure 1a), suggesting a potentially important role for AFAP1-AS1 in PDAC development. Therefore, we next examined the expression of AFAP1- AS1 in multiple PDAC cell lines. We observed that the expression level of AFAP1-AS1 in each PDAC cell line was dramatically up-regulated compared with the HPDE6 cell line (human pancreatic ductal epithelium cell; Figure 1b). Finally, we sought to identify whether AFAP1-AS1 was upregulated in PDAC cell lines and in a large sample size of PDAC tissues. As shown in Figure 1c, we discovered that AFAP1-AS1 was widely upregulated in PDAC tissues compared their paired adjacent non-tumor tissues.
High AFAP1-AS1 expression predicts poor prognosis in PDAC patients with surgical resection We next asked whether the expression of AFAP1-AS1 correlated with the clinical outcome in patients with PDAC. As shown in Figure 1c, 90 cases of PDAC patients received surgical resection were divided into two groups based on APAF1-AS1 expression with 45 patients in each group. Log-rank analysis indicated that the overall survival and progression-free survival was significantly worse in patients with higher AFAP1-AS1 expression in their tumor tissues (Figure 2a, b). Statistical analysis also revealed that AFAP1-AS1 overexpression correlated with lymph node metastasis and perineural invasion (Table 1). No statistical correlation with gender, age, tumor stage and tumor grade was observed. To further determine whether and how AFAP1-AS1 can serve as a biomarker to predict tumor  progression (local recurrence and/or metastasis) after surgery, we constructed a ROC (receiver operating characteristic) curve analysis (Figure 2c). For predicting progression within 1 year, the area under the ROC curve was 0.8669 (p < 0.0001) with an optimal cutoff point of 8.797 (tumor/ para-tumor; sensitivity = 69.81%, specificity = 94.59%) and for predicting progression within in 6 months, the area under the ROC curve was 0.9370 (p < 0.0001) with an optimal cutoff point of 8.797 (tumor/para-tumor; sensitivity = 83.33%, specificity = 91.67%). These findings suggest that AFAP1-AS1 has potential diagnostic value in predicting early recurrence of PDAC.

Inhibition of AFAP1-AS1 in PDAC cells leads to reduced proliferation
To further examine whether AFAP1-AS1 has a causal role in PDAC progression, in vitro functional studies were conducted. We knocked down AFAP1-AS1 expression in MIAPaca-2 and SW1990 cells via stable transfection, and the efficiency of knockdown of the two shRNAs was evaluated (Figure 3a), the most effective shRNA #2 was chose for the following study. AFAP1-AS1 depletion resulted in decreased tumor cell proliferation both in PDAC cell line MIAPaca-2 and SW1990, as determined by MTT assay (Figure 3b, c). We also performed cell cycle assays after shRNA transfection using flow cytometry (Figure 3d, e). Results showed that suppression of AFAP1-AS1 significantly induced G2/M phase arrest. Taken together, these findings suggest that the AFAP1-AS1 modulates cell proliferation partly through regulating cell cycle.

AFAP1-AS1 regulates cell migration and invasion
Enhanced cell migration and invasion abilities are key features associated with cancer metastasis. We therefore examined whether AFAP1-AS1 knockdown affects these functions in PDAC cells. As shown in Figure 4a and b, AFAP1-AS1 knockdown significantly decreased cell motility. Similarly, a matrigel invasion assay showed that AFAP1-AS1 knockdown significantly inhibited invasiveness in MIAPaca-2 and SW1990 cells (Figure 4c, d). Since epithelial-mesenchymal transition (EMT) is closely related with the cell motility and invasiveness, we next examined whether the knockdown of AFAP1-AS1 affects the expression of EMT-related genes. Both PCR (Figure 4e and f) and Western blot analyses (Figure 4g) showed that  Inhibition of AFAP1-AS1 impaired pancreatic cancer cell tumorigenicity in vivo To evaluate the effect of AFAP1-AS1 on the efficiency of xenograft formation of pancreatic cancer cells, we analyzed the in vivo tumorigenicity of SW1990 cells in nude mice following the shRNA-mediated knockdown of AFAP1-AS1. As expected, both tumor volume (Figure 6a, b) and tumor weight (Figure 6c) were significant decreased with pancreatic cancer cells when AFAP1-AS1 expression was inhibited.

Discussion
Although thousands of lncRNAs have recently been identified, investigation of their respective roles in modulating gene expression is relatively incomplete. Functional studies have indicated that some lncRNAs are involved in human cancer tumorigenesis, progression, and adjuvant therapy resistance, acting as oncogenes or tumor suppressors [4,18]. In the current study, by using high-throughput microarrays, we found that AFAP1-AS1 is markedly upregulated in PDAC cell lines and in primary material, and its overexpression correlates with lymph node metastasis, perineural invasion, and poor prognosis of PDAC patients. These observations suggest pro-oncogenic activity of AFAP1-AS1, a notion that is further supported by our functional studies showing that AFAP1-AS1 knockdown attenuates PDAC cell proliferation, migration, and invasion.
Lymph node metastasis and perineural invasion are the strongest indicators of short overall survival in PDAC patients. Efforts have recently been made to identify molecular predictive factors in pancreatic cancer patients [19][20][21]. In the current study, we observed that AFAP1-AS1 overexpression is associated with lymph node metastasis, perineural invasion, and overall survival after surgical treatment, raising the possibility that this lncRNA may provide a means of identifying high-risk patients for more intensive therapy. Importantly, ROC curve analysis revealed that AFAP1-AS1 has great potential in predicting tumor progression after surgery. In the data set of our present study, we observed that an over 8-fold increase in AFAP1-AS1 expression in PDAC tissues compared with adjacent non-tumor tissues was associated with an extremely high risk of short-term recurrence. Whether AFAP1-AS1 alone, or in combination with other markers, could predict PDAC shortterm recurrence required further research through well-designed studies with larger sample size.
In this study, the clinical value of AFAP1-AS1 in PDAC was supported by functional analysis, which showed that AFAP1-AS1 suppression diminished migration, invasion, and expression of EMT-related genes in PDAC cells, and AFAP1-AS1 ectopic expression promoted these malignant behaviors vice versa. This long noncoding RNA was first reported by Wu et al. [22], who showed that AFAP1-AS1 is overexpressed in primary esophageal adenocarcinoma tissues and regulates esophageal adenocarcinoma cell proliferation, migration, and invasion. Consistent with the latter report, the findings of our study support a similar oncogenic role for AFAP1-AS1 in PDAC. Importantly, our data showed that AFAP1-AS1 was one of the most intensely and frequently overexpressed lncRNA in PDAC, further highlighting this transcript to be of significant biological interest in the study of pancreatic cancer pathogenesis.
The AFAP1-AS1 is derived from the antisense strand of the AFAP1 (Actin Filament Associated Protein) gene, the sense strand of which encodes the protein AFAP1. (e-f) Relative mRNA expression levels of EMT-related genes (normalized to β-actin) in PDAC cells after AFAP1-AS1 knockdown were determined by RT-qPCR. (g) Western blot analysis of E-cadherin, N-cadherin, and Vimentin. Data are represented as the mean ± s.d. from three independent experiments, shControl denotes shRNA having no homology to any known mammalian genes as a negative control. ***: P < 0.01, Student's t-test.  Nude mice were subcutaneously injected into the right axilla with 3 × 10 6 cells infected with shControl lentiviral vector (containing scrambled control shRNA) or shAFAP1-AS1 lentiviral vector (containing shRNA targeting AFAP1-AS1). Tumor growth was then monitored using calipers, and the mice were killed two weeks after injection. (c) The tumors were weighed, and compared between the groups. ***: P < 0.01, Student's t-test.