Lipocalin-2 negatively regulates epithelial-mesenchymal transition through matrix metalloprotease-2 downregulation in gastric cancer

Abstract

Background: Although the role of Lipocalin-2 (LCN2) in cancer development has been focused on recent studies, the molecular mechanisms and clinical relevance of LCN2 in gastric cancer (GC) still remain unclear.

Methods: Transcriptome analysis of GC samples from public human data was performed according to Lauren's classification and molecular classification. In vitro, Western blotting, RT-PCR, wound healing assay and invasion assay were performed to reveal the function and mechanisms of LCN2 in cell proliferation, migration and invasion using LCN2 knockdown cells. Gene set enrichment analysis (GSEA) of GC samples from public human data was analyzed according to LCN2 expression. The clinical significance of LCN2 expression was investigated in GC patients from public data and our hospital.

Results: LCN2 was downregulated in diffuse-type GC, as well as in Epithelial-Mesenchymal Transition (EMT) type GC. LCN2 downregulation significantly promoted proliferation, invasion and migration of GC cells. The molecular mechanisms of LCN2 downregulation contribute to Matrix Metalloproteinases-2 (MMP2) stimulation which enhances EMT signaling in GC cells. GSEA revealed that LCN2 downregulation in human samples was involved in EMT signaling. Low LCN2 protein and mRNA levels were significantly associated with poor prognosis in patients with GC. LCN2 mRNA level was an independent prognostic factor for overall survival in GC patients.

Conclusions: LCN2 has a critical role in EMT signaling via MMP2 activity during GC progression. Thus, LCN2 might be a promising therapeutic target to revert EMT signaling in GC patients with poor outcomes.

Keywords: Epithelial–mesenchymal transition; Gastric cancer; Lipocalin-2.

Fig. 1

Transcriptome analysis of gastric cancer patients from public data. a Heatmap of RNA-seq data from 130 patients with gastric cancer according to Lauren classification representing significant differential expressed gene. Blue bar indicates upregulated genes. Red bar indicates downregulated genes. b GSEA of transcriptome data from RNA-seq of DGC patients as compared to IGC patients using HALLMARK gene sets. FDR, false discovery rate. NES, normalized enrichment score. c The top five upregulated and downregulated genes in DGC samples as compared to IGC samples from GSE 113,255. Adj Pval, adjusted p-value. d LCN2 mRNA expression levels in IGC tissues (n = 23) and paired corresponding normal tissues (n = 10). e LCN2 mRNA expression levels in DGC patients compared with IGC patients in ACRG (n = 292) and TCGA (n = 294) cohort studies. f LCN2 mRNA expression level in DGC patients from ACRG cohort study (n = 146) according to molecular classification. EMT, Epithelial–Mesenchymal Transition. The results are presented as a 10–90 percentile plot. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 2

Downregulation of LCN2 induces cell proliferation, migration, invasion and EMT phenotype. a Endogenous LCN2 were expressed by western blot analysis in the gastric cancer cell line. b The efficiency of LCN2 knockdown using two different siRNAs was analyzed in OCUM-12 and NUGC-3 cells by western blot analysis. c The viability of OCUM-12 and NUGC-3 cells was measured by MTT assay 48 h after siRNA transfection (n = 8). d Representative Images of cell migration in OCUM-12 and NUGC3 cells at different time points after scratching. LCN2 knockdown and control cells were analyzed. Wound confluency was quantified (n = 6). e Representative image of invading OCUM-12 and NUGC-3 cells using a two-chamber Matrigel invasion assay. LCN2 knockdown and control cells were analyzed. The number of cells were counted (n = 6). f The number of polygonal or spindle-shaped cells, indicating epithelial–mesenchymal transition (EMT), increased in both OCUM-12 and NUGC-3 cells after siLCN2 treatment. Scale bar, 100 μΜ. Results are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared with siControl

Fig. 3

LCN2 inactivation is associated with EMT signaling via MMP2 activation. a LCN2 related pathway was analyzed by Reactome analysis. FDR, false discovery rate. NES, normalized enrichment score. IGF, Insulin-like Growth Factor; IGFBPs, Insulin-like Growth Factor Binding Proteins. b Western blot analysis revealed several types of Matrix Metalloproteases in OCUM-12 and NUGC-3 cells transfected with siLCN2 and siControl, respectively. c The expression level of LCN2 was significantly decreased under hypoxic condition in OCUM-12 cells. d GSEA of transcriptomic data of GC patient sample from GSE 62,254 using HALLMARK gene sets and stratified on the basis of LCN2 expression. e GSEA plot of “MYOGENESIS”, “EPITHELIAL_MESENCHYMAL_TRANSITION” in the transcriptomic data of GC patient samples from TCGA using HALLMARK gene sets and stratified on the basis of LCN2 expression

Fig. 4

Association between LCN2 and SAA1 in GC cells and patients. a STRING predicted protein interactions derived from significantly downregulated DEG in GES113255. LCN2 interaction with SAA1 was targeted in this study (red arrow). b SAA1 mRNA level was detected by RT-PCR in gastric cancer cell lines. c qPCR analysis of mRNA of SAA1 in OCUM-12 and NUGC-3 cells transfected by siRNA targeted to LCN2 (n = 3). d A positive association of mRNA levels between LCN2 and SAA1 was detected in GC patients from GSE113255 and ACRG cohort studies. e SAA1 phenotypes are distinguished by different amino acids in each of the variants, which are highlighted in red. f Sanger sequencing of SAA1 in OCUM-12 and NUGC-3 cells. Amino acid changes at positions 70, 75 and 90 are highlighted in red. Results are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 compared with siControl

Fig. 5

Clinical relevance of LCN2 in GC patients. a Kaplan–Meier curves for 7-year overall survival and disease-free survival of GC patients from ACRG cohort study according to LCN2 expression (n = 292 and 275, respectively). HR hazard ratio. b Disease-free survival of IGC patients from the ACRG cohort study stratified on the basis with LCN2 expression (n = 139). c Multivariate Cox multiple regression analysis for overall survival of GC patients in ACRG cohort study (n = 292). d LCN2 staining using the LCN2 antibody was mainly found in the cytoplasm of the cancer cells. All fields were analyzed at a magnification of × 200. Scale bar, 100 μΜ. e Overall survival of 590 patients with GC according to LCN2 expression by immunohistochemistry