MicroRNA-19b-3p suppresses gastric cancer development by negatively regulating neuropilin-1

102 paired GC tissues and adjacent nontumorous tissues were collected from patients undergoing surgery at Ganzhou People’s Hospital and the First Affiliated Hospital of Zhengzhou University (GZPH GC cohort) between year 2010 and 2018. Tissue microarray (TMA) were constructed with GC tissue specimen. All patients signed the written informed consent. The Ethics Review Committee of Ganzhou People’s Hospital approved the study.

GC cell lines (SGC-7901, AGS, MGC-823, MKN-45, BGC-823) and control gastric epithelial cell line GES-1 were purchased from the Cell Bank, Chinese Academy of Science (Shanghai, China). HEK293 cells were from ATCC and maintained in the lab. Cells were cultured in Dulbecco’s Modified Eagle’s Medium (Gibco, USA) supplemented with 10% fetal bovine serum (Clark, USA), 100 U/ml Penicillin, and 100 μg/ml Streptomycin at 37 °C under 5% CO2 in a humidified incubator.

SGC-7901 or MGC-823 cells were transiently transfected with si-NRP1 or negative control (NC), miR-19b-3p mimics or NC, miR-19b-3p inhibitor or NC, or pcDNA3.1-NRP1 using lipofectamine 3000 (Invitrogen, USA). SGC-7901 cells were transduced with sh-NRP1 or NC and stably NRP1 knockdown cells were selected using neomycin before cell implantation.

Complementary DNA was prepared from total RNA using First strand cDNA synthesis Kit (Roche, Germany). NRP1 and miR-19b-3p expression were analyzed by quantitative PCR using SYBR Green master mix (Bio-Rad, USA). GAPDH and U6 were used as control. The primers used in the study were listed as following: NRP1, 5′-ACCCAAGTGAAAAATGCGAATG -3′ and 5′-CCTCCAAATCGAAGTGAGGGTT-3′; miR-19b-3p, 5′-AACAGAAGTTTTGCAGGTTTGCATC-3′ and 5′-CAGTGCAGGGTCCGAGGT-3′; U6, 5′-CCAGUUUACCUAACGCAAUTT-3′ and 5′-TTCACGAATTTGCGTGTCAT-3′; GAPDH, 5′-CTGGGCTACACTGAGCACC-3′ and 5′-AAGTGGTCGTTGAGGGCAATG-3′.

Western blot was performed as previously described [24]. The primary antibodies used in the study were listed as following: NRP1 (Abcam, ab81321), E-cadherin (Abcam, ab194982), vimentin (Abcam, ab92547), N-cadherin (Abcam, ab18203), BMP4 (Abcam, ab200796), ICAM1 (Abcam, ab221777), VCAM1 (Abcam, ab134047), GAPDH (Abcam, ab181602).

Cell growth was analyzed by CCK-8 assay, EdU staining assay and Colony formation assay as previously described [25].

Cell invasion was evaluated with transwell chamber (Corning, USA). Briefly, 1 × 10 [5] transfected SGC-7901 or MGC-823 cells in serum-free medium were added to the Matrigel-coated top chamber. The bottom chamber was added with 500 μL DMEM medium containing 10% FBS. After 48 h, the invaded cells were fixed and stained with crystal violet.

SGC-7901 or MGC-823 cells were cultured to form monolayer in 6-well plate. An artificial wound was created using a sterile 200 μL pipette tip. Floating cells were washed awat with PBS and rest cells were cultured in serum-free medium for 48 h. The wound was recorded at 0 h and 48 h to calculate the migration distance.

Luciferase vector containing wild type or mutated 3′-UTR sequence of NRP1 was cloned using pGL3-luc vector (Promega, USA). HEK393 cells were seeded into 24-well plates and transfected with luciferase vector, together with miR-19b-3p mimics or negative control. Relative luciferase activity was analyzed using the Dual-Glo luciferase reporter assay kit (Promega, USA) 48 h later.

Briefly, 5 μm paraffin tumor section slides were de-paraffined and blocked with normal rat serum, and then incubated with primary antibodies overnight at 4 °C. Then, slides were incubated with HRP conjugated secondary antibodies, and the specific protein expression was detected using a DAB kit (Vector Lab, USA). The primary antibodies used in the study were listed as following: Ki-67 (Proteintech, 27309-1-AP), NRP1 (Abcam, ab81321), E-cadherin (Abcam, ab194982), vimentin (Abcam, ab92547), N-cadherin (Abcam, ab18203), BMP4 (Abcam, ab200796), ICAM1 (Abcam, ab221777), VCAM1 (Abcam, ab134047).

BALB/C nude mice (5–6 weeks, 6 mice/group) were obtained from Vital River Laboratory (Beijing, China). 3 × 10⁶ stable transfected SGC-7901 cells with luciferase reporter were subcutaneously inoculated into the nude mice. Tumor growth was monitored every week and tumor bioluminescence was photographed using an IVIS Lumina II system (Caliper Life Sciences, USA) following the manual. Tumor volume was calculated (length × width²/2). Tumor were extracted and weighed at 5 weeks. The animal experiment was approved by the Institutional Animal Care and Use Committee of Ganzhou People’s Hospital.

To explore the expression of NRP1, we examined the NRP1 expression in 30-paired GC and adjacent nontumorous tissues. The mRNA levels of NRP1 were significantly upregulated in GC tissues (Fig. 1a). In addition, GC tissues had higher protein expression levels of NRP1 than that in noncancerous normal tissues (Fig. 1b). We further revealed that the expression of NRP1 was notably enhanced in GC cell lines than that in control cell line GES-1 (Fig. 1c). NRP1 IHC staining was performed using GC TMA (Fig. 1d) and GC tissues showed higher NRP1 IHC staining intensity (Fig. 1e). Intriguingly, higher NRP1 expression was significantly correlated with late TNM stages, distant metastasis and recurrence (Fig. 1f–h). Kaplan–Meier analysis showed that patient with high expression of NRP1 had poor overall survival (OS) and disease-free survival (DFS) compared with that in patients with low NRP1 expression in GZPH GC cohort (Fig. 1i–j). We also performed Kaplan–Meier analysis based on TCGA GC cohort and kmplot GC cohort. The results suggested that high NRP1 expression was associated with poor prognosis in both TCGA GC cohort and kmplot GC cohort (Fig. 2a–d). These findings demonstrated that NRP1 expression was upregulated in GC and was associated with poor prognosis.

To illustrate the function of NRP1 in GC, RNA silencing was performed to knockdown the expression of NRP1 in GC cells by using siRNA (Fig. 3a). Functionally, knockdown of NRP1 inhibited cell growth, DNA synthesis and colony formation of SGC-7901 and MGC-823 GC cells (Fig. 3b–e). In addition, inhibition of NRP1 suppressed GC cell migration and invasion, as demonstrated by wound-healing assay and transwell assay (Fig. 3f–g). Thus, knockdown of NRP1 inhibits the malignant phenotype of GC cells in vitro.

We further investigated the role of NRP1 in xenograft GC tumor model. SGC-7901 cells stably knockdown NRP1 or control cells were subcutaneously inoculated into nude mice to establish the xenograft tumor. As shown in Fig. 4a–b, knockdown of NRP1 significantly repressed GC tumor growth in vivo. Knockdown of NRP1 markedly slowed tumor growth (Fig. 4c), with a much lower tumor weight (Fig. 4d) compared with control group. Moreover, IHC staining showed that knockdown of NRP1 notably down-regulated the expression of NRP1 and proliferation marker Ki-67 in tumor tissues (Fig. 4e–f). In summary, knockdown of NRP1 suppresses GC tumor development and progression in vivo.

To further understand how NRP1 is regulated in GC development, bioinformatics analysis was performed using online database miRBase to predict the potential miRNAs interacting with NRP1. MiR-19b-3p had the putative binding sequences targeting 3′-URT of WT NRP1 (Fig. 5a). In addition, miR-19b-3p expression was down-regulated in GC tissues in comparison with that in adjacent normal tissues in GZPH GC cohort (Fig. 5b). Pearson correlation analysis revealed that miR-19b-3p expression was negatively associated with NRP1 expression (r = −0.474, P = 0.012, Fig. 5c). Overexpression of miR-19b-3p significantly dampened NRP1 expression, while inhibition of miR-19b-3p enhanced NRP1 expression in SGC-7901 or MGC-823 cells (Fig. 5d–e). Luciferase reporter assay was carried out to further validate the regulation between miR-19b-3p and NPR1. MiR-19b-3p mimics specifically decreased the luciferase activity in HEK293 cells transfected with reporter vector containing WT 3′-UTR of NRP1, but not in cells transfected with reporter vector with mutated 3′-UTR of NRP1 (Fig. 5f). Taken together, the data suggest that NRP1 is negatively regulated by miR-19b-3p.

Rescue experiment was conducted to further verify the regulation of NRP1 by miR-19b-3p. Overexpression of miR-19b-3p notably suppressed NRP1 expression while the protein expression of NRP1 was partially restored by NRP1 overexpression in SGC-7901 or MGC-823 cells (Fig. 6a). Functionally, NRP1 overexpression dampened the inhibitory effect of cell growth mediated by miR-19b-3p mimics (Fig. 6b–d). Furthermore, overexpression of miR-19b-3p inhibited GC cell invasion and migration, while the inhibition was reversed by NRP1 overexpression (Fig. 6e–f). These results suggest that miR-19b-3p regulates GC cell growth, migration and invasion via negatively regulating NRP1.

Gene set variation analysis (GSVA) using TCGA GC datasets was performed to further understand the functional mechanism of NPR1 in GC development. As shown in Fig. 7a, epithelial-mesenchymal-transition (EMT) was among the most upregualted pathways in GC tissues with high NRP1 expression compared with that in GC tissues with low NRP1 expression. Gene Set Enrichment Analysis (GSEA) further validated that EMT and focal adhesion were significantly enriched in NRP1 high group (Fig. 7b–c). IHC staining was performed in xenograft and the results found that knockdown of NRP1 significantly elevated E-cadherin expression and repressed the expression of Vimentin and N-cadherin (Fig. 7d). Moreover, IHC staining revealed that knockdown of NRP1 inhibited the expression of focal adhesion/migration related proteins such as BMP4, ICAM1 and VCAM1 (Fig. 7e). The regulation of EMT and focal adhesion by miR-19b-3p/NRP1 was further tested in SGC-7901 cells. Overexpression of miR-19b-3p or knockdown of NRP1 significantly enhanced E-cadherin expression and suppressed Vimentin/N-cadherin expression, while repressed BMP4, ICAM1 and VCAM1 expression in GC cell lines. The functional effect of miR-19b-3p overexpression was partially reversed by NRP1 overexpression (Fig. 8a). In contrast, miR-19b-3p inhibition promoted EMT-related Vimentin and N-cadherin expression and enhanced focal adhesion with higher expression of BMP4, ICAM1, and VCAM1. Knockdown of NRP1 abrogated the regulatory effect of miR-19b-3p inhibition (Fig. 8b).