Identification of BANF1 as a novel prognostic biomarker in gastric cancer and validation via in-vitro and in-vivo experiments

Abstract

Gastric cancer (GC) is a widespread malignancy characterized by a notably high incidence rate and an unfavorable prognosis. We conducted a meticulous analysis of GC high-throughput sequencing data downloaded from the Gene Expression Omnibus (GEO) repository to pinpoint distinctive genes associated with GC. Our investigation successfully identified three signature genes implicated in GC, with a specific focus on the barrier to autointegration factor 1 (BANF1), which exhibits elevated expression across various cancer types, including GC. Bioinformatic analysis has highlighted BANF1 as a prognostic indicator for patients with GC, with direct implications for immune cell infiltration. To gain a more comprehensive understanding of the significance of BANF1 in GC, we performed a series of in vitro experiments to confirm its high expression in GC tissues and cellular components. Intriguingly, the induction of BANF1 knockdown resulted in a marked attenuation of proliferation, migratory capacity, and invasive potential in GC cells. Moreover, our in vivo experiments using nude mouse models revealed a notable impediment in tumor growth following BANF1 knockdown. These insights underscore the feasibility of BANF1 as a novel therapeutic target for GC.

Keywords: GEO; TCGA; barrier to autointegration factor 1; gastric cancer; overall survival.

Figure 1

Identification of differentially expressed genes in gastric cancer and paracancerous tissues. (A) Volcano plot of 157 differentially expressed genes (|Log2FC| > 2, p < 0.01). (B) Heatmap of 93 down-regulated genes expressed in gastric cancer tissues. (C) Heatmap of 64 expressed up-regulated genes in gastric cancer tissues.

Figure 2

Identification of hub genes in gastric cancer using the WGCNA algorithm. (A) The soft threshold is determined by function. The left panel shows the relationship between the scale-free network evaluation metric R² and the soft threshold, and the right panel shows the relationship between average connectivity and the soft threshold. (B) Dendrogram of gene clustering and different colored modules. (C) Gene clustering dendrogram obtained by merging similar modules. (D) Correlations between merged modules and clinical traits, correlation coefficients and p-values are shown in the corresponding modules of different colors. (E) Scatterplot of MM and GS in the black module.

Figure 3

Machine learning screens for disease characterizing genes. (A, B) LASSO regression analysis screening variables. (C) Cross-validation error rate graph based on SVM-RFE. (D, E) Genes were scored using a random forest algorithm to rank genes by importance algorithm. (F) VEEN graph to obtain the intersection of key genes screened by the 4 methods.

Figure 4

Expression and diagnostic value of characterized genes. (A) The box plots showed the expression of three GC characteristic genes (BANF1, ADH7, TMEM27) in GSE54129-GSE118916 dataset. (B) ROC curves were used to evaluate the diagnostic efficacy of three GC characteristic genes (BANF1, ADH7, TMEM27) in GSE54129-GSE118916 dataset. (C) The box plots showed the expression of three GC characteristic genes (BANF1, ADH7, TMEM27) in GSE65801 dataset. (D) ROC curves were used to evaluate the diagnostic efficacy of three GC characteristic genes (BANF1, ADH7, TMEM27) in GSE65801 dataset. (E) The box plots showed the expression of three GC characteristic genes (BANF1, ADH7, TMEM27) in TCGA dataset. (F) ROC curves were used to evaluate the diagnostic efficacy of three GC characteristic genes (BANF1, ADH7, TMEM27) in TCGA dataset. ^*p < 0.05, ^***p < 0.001.

Figure 5

The prognostic value of three characteristic genes. (A–C) OS, PPS, FP km survival curves between high and low expression groups of BANF1. (D–F) OS, PPS, FP km survival curves between high and low expression groups of ADH7. (G–I) OS, PPS, FP km survival curves between high and low expression groups of TMEM27.

Figure 6

Single-cell expression of BANF1 in gastric cancer and its immune correlation. (A, B) Validation of BANF1 expression in different cell types in two gastric cancer single cell datasets (GSE167297, GSE134520). From left to right, cell clustering plot, cell annotation plot, violin plot of cellular signature genes expressed in different cell clusters, and BANF1 expression plot in different cells. (C) Lollipop plot of BANF1 expression correlating with immune cells. (D) Stromal score, immune score, ESTIMATE score violin plots between high and low BANF1 expression groups. (E) Box line plot showing immune infiltration between high and low BANF1 expression groups. ns means no statistical difference, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 7

Expression and localization of BANF1. (A) BANF1 expression among pan-cancer unpaired samples in the TCGA database. (B) BANF1 expression among paired samples of pan-cancer in the TCGA database. (C) RT-qPCR detection of BANF1 mRNA expression in 23 pairs of GC tissues and paired paracancerous tissues. (D) WB detection of BANF1 protein expression in 10 pairs of GC tissues, and the difference of gray values between gastric cancer tissues (n = 10) and adjacent tissues (n = 10) were compared. (E) RT-qPCR was used to detect the expression of BANF1 mRNA in GC cell lines (AGS, HGC-27, MKN-45, MGC-803, BGC-823) and gastric mucosal epithelial cell line (GES-1), respectively. (F) The expression of BANF1 protein in 6 cell lines was detected by WB, and the gray values of gastric cancer cell lines (AGS, HGC-27, MKN-45, MGC-803, and BGC-823) and gastric epithelial cells (GES-1) were compared. (G) Immunofluorescence staining showed the expression and localization of BANF1 protein in MKN-45 and BGC-823 cells. All experiments were repeated at least three times. ns means no statistical difference, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 8

Knockdown of BANF1 inhibits the proliferation, migration and invasion ability of GC cells in vitro. (A) Knockdown efficiency of MKN-45 cell line was detected using RT-qPCR and WB. (B) CCK8 assay was used to detect the effect of BANF1 knockdown on the viability of MKN-45 cell line. (C) Knockdown efficiency of BGC-823 cell line was detected using RT-qPCR and WB. (D) CCK8 assay was used to detect the effect of BANF1 knockdown on the viability of BGC-823 cell line. (E, F) The effect of BANF1 knockdown on the proliferative capacity of MKN-45 and BGC-823 cells was detected by plate clone formation assay. (G, H) Effects of BANF1 knockdown on the migration ability of MKN-45 and BGC-823 cells were detected by wound healing assay. (I, J) Effect of BANF1 knockdown on the migration and invasion of MKN-45 and BGC-823 cells by Transwell assay. All experiments were repeated at least three times. ^***p < 0.001.

Figure 9

Knockdown of BANF1 inhibited tumor growth in in vivo experiments. (A) Photographs of nude mice injected with BANF1 knockdown MKN-45 (n = 6) and blank control MKN-45 cells (n = 6). (B) Photographs of subcutaneous tumors of nude mice in the knockdown group (n = 6) and control group (n = 6). (C, D) Comparison of subcutaneous tumor volume and weight of nude mice in knockdown and control groups. (E) Immunohistochemical staining of Ki67 in subcutaneous tumors of nude mice. (F) TUNEL staining of subcutaneous tumors in nude mice. All experiments were repeated at least three times. ^**p < 0.01, ^***p < 0.001.