Neuropilin‐1 is up‐regulated by cancer‐associated fibroblast‐secreted IL‐8 and associated with cell proliferation of gallbladder cancer

Abstract We previously demonstrated that cancer‐associated fibroblasts (CAFs) promoted the proliferation of gallbladder cancer (GBC) cells, but the mechanism is not clear. Neuropilin‐1 (NRP‐1) plays an important role in various malignancies as transmembrane glycoprotein. Our goal was to reveal the relationship between CAFs and NRP‐1 and their potential functions in GBC. In this study, we found NRP‐1 was overexpressed in GBC tissue, associated with poor survival and was up‐regulated by CAFs. The cytokine array cluster analysis revealed IL‐8 secreted by CAFs facilitated the up‐regulation of NRP‐1 in tumour cells. NRP‐1 knockdown suppressed tumour growth in vivo. Gene expression microarray analysis showed 581 differentially regulated genes under NRP‐1 knockdown conditions. Ingenuity pathway analysis demonstrated that NRP‐1 knockdown may inhibit tumour progression by affecting cell proliferation. We then confirmed that NRP‐1 knockdown in NOZ and GBC‐SD cells significantly inhibited cell proliferation. Additionally, the IL‐8 mediated MDM2 and CCNA2 expression were affected by NRP‐1 knockdown. Our findings suggested that NRP‐1 was up‐regulated by CAF‐secreted IL‐8, which subsequently promoted GBC cell proliferation, and these molecules may serve as useful prognostic biomarkers and therapeutic targets for GBC.


| INTRODUC TI ON
Gallbladder cancer (GBC) is the most common malignant tumour of the biliary system. 1 GBC is associated with poor treatment response and prognosis because it is difficult to diagnose early and has a high propensity to metastasize to lymph nodes. 2 Biliary malignancies are characterized by interstitial fibrosis, 3 and cancer-associated fibroblasts (CAFs) have been shown to be directly associated with patient outcomes in a variety of tumours, including cholangiocarcinoma. 4 Our previous study demonstrated that CAFs promoted the proliferation and invasion of GBC cells. 5 However, which cytokines are secreted by CAFs and how these cytokines act on GBC cells remain unclear.
Neuropilin-1 (NRP-1) was involved in the progress of many cancers, 6 it is found overexpressed in hepatocellular carcinoma, 7 colorectal cancer, 8 glioblastoma, 9 and lung cancer, 10 and its expression associated with advanced cancer stages and poor prognosis. [11][12][13][14] The mechanisms for the role of NRP-1 in cancer progression rely on its interactions with several key signalling pathways in cancer cells, such as transforming growth factor β1 (TGFβ1), semaphorin (Sema), vascular endothelial growth factor (VEGF), hedgehog (HH), interferon-γ (IFNγ) and GAIP/RGS19-interacting protein (GIPC1). [15][16][17][18][19][20] Therefore, targeting NRP-1 has been shown to be a potential therapeutic target for some cancers. 21,22 However, the role of NRP-1 in GBC remains unclear. This study investigated the interactions between CAFs and NRP-1 and clarified their biological and prognostic role in GBC tumorigenesis.  23 The TNM staging was according to the 7th AJCC criteria. 24 The study was approved by the Ethics Committee of the First Affiliated Hospital of Xi'an Jiaotong University, China.

| Patient tissue samples and follow-up
The data cut-off point of follow-up was August 2016, and overall survival (OS) was defined as the time interval from surgery to death.

| Tissue microarray, immunohistochemistry and immunofluorescence
Tumour microarray (TMA) was constructed using paraffin-embedded GBC tissue and cholecystitis. We performed immunohistochemistry using a two-step method.
The TMA sections were incubated with first antibody overnight at 4°C and then incubated with secondary antibodies for 1 hour. The cell nuclei were counterstained with haematoxylin (H8070, Solarbio Beijing, China) for 30 seconds. Diaminobenzidine (DA1010, Solarbio, Beijing, China) was used for visualization of positive staining. We scored positive cell staining as follows: 0 point: <10% positive cells,

| Cell, antibodies, and reagents
We purchased GBC-SD, NOZ and SGC-996 cell lines from Chinese Academy of Sciences Cell Bank (Shanghai, China) and cultured GBC cell lines with RPMI-1640. All cells used in this study were authenticated by short tandem repeat DNA profiling by Genetica, and mycoplasma contamination was routinely tested using detection kit (CA1080, Solarbio).
We isolated CAFs and normal fibroblasts (NFs) from 4 different GBC patients (all these 4 patients with T3N0M0 adenocarcinoma) and 4 different chronic cholecystitis patients (all these 4 patients were young men less than 35 years old and with a history of gallstone less than one year), using previously described digestion methods. 5 These antibodies and reagents were used in this study: an-

| Determination of cytokines secreted by CAFs
Four cases of CAFs and four cases of NFs were plated in 10mm dishes (3*10 4 cells each dish) and cultured for 72 hours, and cell supernatants were collected. Following the standard operating procedures for Human Cytokine Array G5 (AAH-CYT-G5) protein array kit (RayBiotech), 20 µL supernatant from each tube was used to assess the difference of cytokines expression between CAFs and NFs. Later, we obtained the expression levels by using fluorescence scanning (GenePix 4000B, Axon Instruments, Inc) and analysis with a built-in data transformation tool. After that, we confirmed IL-8 was highly expressed in the CAFs supernatant by ELISA. Furthermore, to investigate whether NRP-1 was upregulated by the CAFs-secreted IL-8, GBC cells were divided into three groups: GBC cells as control group; cells treated with CAFs supernatants for 48 hours; cells treated with CAFs supernatants combined Reparixin (0.1 μmol/L) for 48 hours; and the expression of NRP-1 were tested.

| Tumour formation in nude mice
All mice were housed and handled under the guidelines of Institutional Animal Care and Use Committee of Xi'an Jiaotong University. The sh-NRP-1-GBC-SD or NT sh-Ctrl cells were digested and made into cell suspensions. After counting, the cells were kept at 4°C for further use. Twelve 4-to 6-week-old healthy nude mice (SJA Laboratory Animal Co., Ltd.) were allocated into two groups randomly. 1 × 10 6 cells were subcutaneously injected to the lower limbs. Tumour size was measured every 3 days, and the mice were killed 3 weeks after injection.

| Cellular proliferation assay
We determined cell proliferation rates using MTT assay. GBC cells infected with NRP-1 shRNA or NT shRNA were seeded in 96-well plates (2 × 10 3 cells per well), after incubated at 37°C for 24 and 48 hours, MTT solution was added and incubated for 4 hours, and then, the absorbance values at 490 nm were obtained. For the cell colony formation assays, we seeded 1000 cells of each group per well in a 60 mm dish and incubated for 14 days, fixed the cells with 4% paraformaldehyde and stained with crystal violet, and those colonies with more than 50 cells were count.

| Statistical analysis
We performed statistical analysis using SPSS 14.0 (SPSS Inc IBM).
Independent-sample t test and ANOVA test was used for comparisons between two groups and multiple groups, respectively. Chisquare test of independence was used to determine whether there is a significant difference between categorical variables. Survival was analysed using the Kaplan-Meier method, and the differences were measured by log-rank tests. Cox regression was carried out for prognostic multivariate analysis. A P < .05 was defined as statistically significant.

| High NRP-1 expression associated with poor survival
Kaplan-Meier survival analysis of patients with different expression levels of NRP-1 showed that the median survival time (MST) was 33 months in the low NRP-1 expression group, in contrast, MST of the high NRP-1 expression group was only 4 months (P < .01, Figure 1C). Multivariate analysis showed that surgical margin, NRP-1 expression and pathological differentiation were all independent risk factors for poor prognosis ( Table 2).

| CAFs-secreted IL-8 up-regulated NRP-1
We treated GBC cell lines (GBC-996, GBC-SD and NOZ) with conditioned supernatants from CAFs and NFs, and we found that CAFs supernatant up-regulated NRP-1 expression level in GBC cell lines ( Figure 2A). We used cytokine array cluster analysis to compare the exocrine cytokines of CAFs and NFs, and we found that IL-8 and IGFBP3 secreted by CAFs were higher than that by NFs (P < .05) ( Figure 2B and Data S1). Furthermore, we confirmed that CAFs supernatants contained more IL-8 than NFs supernatants by ELISA (365.7 pg/mL vs 129.3 pg/mL, P < .05) ( Figure 2C), and we also performed the IL-8 immunofluorescence in NFs and CAFs, which confirmed that CAFs secreted more IL-8 than NFs ( Figure 2D).

| Suppression of NRP-1 attenuated tumour formation and growth
Twelve mice were assigned to the NRP-1 shRNA and NT shRNA groups. We found that tumours grew more rapidly in the NT shRNA group as shown in Figure 3B. Mice were killed on the 21st day after injection, as Figure 3C-E shown, tumours from the NT shRNA group were much larger and heavier than tumours in the NRP-1 shRNA group (P < .01).
We performed TUNEL assay and IHC staining to evaluate proliferation and apoptosis. Comparing to the control group, in the NRP-1 shRNA group, the number of TUNEL positive cells was significantly increased and the expression of Ki-67 in tumours was decreased (P < .001). We also found that Caspase-3 protein expression was higher in the NRP-1 shRNA group than that in the control group (P < .001) ( Figure 3F).

| Gene chips analysis of NRP-1 inhibition
Gene chips were used on three samples each from the NT shRNA-GBC-SD and NRP-1 shRNA-GBC-SD, and pathway analysis was performed to explore the regulation and function of NRP-1 in GBC cells.
A total of 48,726 publicly annotated gene clusters (Data S2) were identified in the microarray data, and 581 genes were differentially expressed among them (FC > |1.5|; P < .05) between the NT shRNA and NRP-1 shRNA groups (Data S3). Among them, 182 genes were up-regulated and 399 genes were down-regulated in the NRP-1 shRNA group. An interactive heat map has shown the abundance of these DE genes ( Figure 4A). The top 20 down-regulated and up-regulated genes were listed in Tables S1 and S2.

| Ingenuity Pathway Analysis of NRP-1 related biological functions
We next imported DE genes data sets into IPA to investigate the possible biological interactions and pathways. Among the 299 different canonical pathways identified by IPA, 43 were significantly enriched (P < .05), and 19 showed either positive or negative z-scores (ie these pathways were activated or suppressed, respectively; see Data S4).

Among these 19 pathways, Cell Cycle: G2/M DNA Damage Checkpoint
Regulation was the only pathway with |z-score | > 2, which meant it was significantly changed ( Figure 4B).
IPA also grouped the DE genes into 18 functional categories, which included 500 diseases or bio functions (Data S5). The enrichment status of these 18 categories is presented in Figure 4C. The heat map showed the activated or suppressed state of these 500 diseases or bio functions after NRP-1 knockdown. Twenty-one diseases or bio functions were significantly increased or decreased (|z-score| > 2; Table S3).
Among these significantly changed diseases and bio functions, proliferation of cells was decreased and ranked as the highest network with 206 focus molecules (z-score = 4.357; Figure 4D), including TGFBR1, EGR1, CCNA2, MDM2, BBC3, IGF2, RRas2 and CDK6. Figure 4E presents the specific status of genes in the Proliferation of cells network.

| Suppression of NRP-1 inhibited growth and colony formation in GBC cells
IPA revealed that pathways and bio functions related to cell proliferation and cell cycle were significantly affected by NRP-1 knockdown. To confirm this finding, we conducted MTT assays. We found that cell proliferation was decreased in NRP-1 shRNA GBC-SD and NOZ cells ( Figure 5A). As the tumour colony formation assay showed ( Figure 5B), in both GBC-SD and NOZ cells, the number of colonies was significantly lower in the NRP-1 shRNA group (P < .01). This suggested that NRP-1 expression was associated with GBC cell proliferation.

F I G U R E 3 NRP-1 knockdown impairs effects of GBC
We then analysed the expression of NRP-1-related genes at the protein level in NRP-1 shRNA (KD) and NT shRNA (NC) GBC-SD cells. As shown in Figure 5C, CCNA2 protein expression was decreased and almost undetectable in the KD group, and MDM2 protein expression was also decreased by 66.3% in the KD group.
In contrast, BBC3 protein was increased in the KD group. EGR1

| D ISCUSS I ON
NRP-1 expression has been associated with prognosis in a variety of tumours, such as liver cancer, bladder cancer and colon cancer. 25,26 Although many studies have focused on NRP-1 function in tumour angiogenesis, 27 CCNA2 drives S phase progression by binding to and activating Cdk2 and Cdk1. Then, Cdk/CCNA2 complexes phosphorylate pocket proteins (Rb, p107, p130) and proteins involved in DNA synthesis. 33 Aberrant expression of CCNA2 has been detected in a variety of cancers, and deregulation of CCNA2 was closely related to tumour proliferation and chromosomal instability. 34 Inhibition of CCNA2 complexes also has been shown to impair the proliferation of tumour cell lines. 35,36 MDM2 is an essential regulator of the p53 tumour suppressor, and it is modified at the transcriptional, post-transcriptional and post-translational levels to control p53 activity. Higher MDM2 protein expression also results in an increased risk for spontaneous tumour formation. [37][38][39] Taken together, our study suggests that IL-8 secretion by CAFs can up-regulate NRP-1 in GBC cells and subsequently promote GBC cell proliferation by CCNA2 and MDM2. A new NRP-1 targeting probe has been be used for the grading diagnosis by MRI of gliomas in nude mice. 40 As we found that NRP-1 expression was associated with GBC tumour differentiation, targeting NRP-1 might also be used for the GBC grading diagnosis. Besides, tumour progression may be retarded by targeting CAFs, 41 and a tumour-imaging method targeting CAFs has shown its potential to serve as pantumor agents. 42 We believe that the CAFs/IL-8/NRP-1 axis could be used in developing novel diagnostic and therapeutic strategies for GBC.

ACK N OWLED G EM ENTS
The research was supported by grants from National Natural We thank LetPub (www.letpub.com) for its linguistic assistance during the preparation of this manuscript.

CO N FLI C T S O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
Clinical data have been approved by the Ethics Committee of the First Affiliated Hospital of Xi'an Jiaotong University and approved by the patients. All animal experiments were approved by Animal Care and Use Committee of Xi'an Jiaotong University.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.