Multidisciplinary analysis of the prognosis and biological function of NUBPL in gastric cancer

Abstract

Background: Researchers are currently concentrating on molecular markers and potential therapeutic targets associated with gastric cancer in light of recent developments in precision medicine and molecular biology. Disulfidptosis was first proposed in 2023 as a novel programmed cell death mode associated with the cytoskeleton. Disulfidptosis-related proteins are essential for the preservation of protein stability and abnormal expression of disulfidptosis-related genes may be linked to cancer development and drug resistance.

Materials and method: The gastric cancer transcriptomic data were retrieved from TCGA database, and disulfidptosis-related genes were identified through literature search. Utilizing machine learning methods such as LASSO, Random Forest (RF), Boruta, SVM-RFE, and XGBoost, the disulfidptosis-related gene NUBPL was determined as a potential predictor for gastric cancer. PPI network was constructed, and the GO database as well as the KEGG database were employed to analyze the protein interactions and pathway enrichment of NUBPL in gastric cancer. Meanwhile, the ESTIMATE algorithm was used for immune infiltration analysis and prediction of immunotherapy response, and the Genomics of Drug Sensitivity in Cancer (GDSC) database was utilized for the drug sensitivity analysis of NUBPL. The role of NUBPL in gastric cancer and its inhibition of disulfidptosis were validated using molecular biological methods.

Results: The aberrant expression of NUBPL significantly impacts the prognosis of gastric cancer and modulates metabolic and immune-related pathways. In patients with elevated NUBPL expression levels, a reduced number of CD8-positive T cells is associated with adverse prognosis and gastric cancer progression. Elevated NUBPL expression levels can impair the function of chemokines. Moreover, patients with lower NUBPL expression levels exhibit better responses to immunotherapy. We have also identified drugs such as QS11, Imatinib, and AS601245 as potential inhibitors of NUBPL. In vitro experiments have shown that NUBPL affects the invasion and migration of gastric cancer cells, rather than proliferation and apoptosis, by regulating the PPP pathway and inhibiting disulfidptosis.

Conclusion: This study underscores the pivotal role of NUBPL in gastric cancer progression and highlights its significance as a potential target for targeted therapy and immunotherapy gastric cancer, NUBPL, disulfidptosis, biomarker, immunotherapy, machine learning.

Keywords: NUBPL; biomarker; disulfidptosis; gastric cancer; immunotherapy; machine learning.

Figure 1

Identification of diagnostic genes employing two machine learning algorithms. (A) Variable selection utilizing Lasso regression, illustrating the variations in variable coefficients. (B) Selection procedure for the optimal parameter λ in the Lasso regression model via cross-validation. (C, D) ROC curves for the test and validation sets in the Lasso model. (E) Analysis of gastric cancer risk utilizing Random Survival Forest methodology. (F, G) ROC curves for the test and validation sets in the Random Forest model. (H) Venn diagram illustrating candidate diagnostic genes identified by both algorithms.

Figure 2

Identification of diagnostic genes employing machine learning algorithms. (A, B) 23 important characteristic genes were identified by Boruta algorithm. (C) Support vector machine recursive feature elimination (SVM-RFE) algorithm was used to identify the top 10 genes in gene expression data. (D) xgboost algorithm was used to obtain 19 important genes. (E) The residual box plot shows the median, quartile range, outlier and other information of the residual. (F, G) The inverse cumulative distribution of sample residuals and ROC curve shows that almost 100% of sample residuals are concentrated near 0. (H) Each model outputs the top10 most important genes.

Figure 3

Clinical examination of hub genes (A, B) Kaplan-Meier survival curves illustrating the survival rates of high and low expression cohorts for NCKAP1 and NUBPL. (C, D) Expression levels of NCKAP1 and NUBPL at various stages of gastric cancer. (E, F) Variations in the expression of NCKAP1 and NUBPL among gastric cancer patients of varying ages. (G-H) Association between NCKAP1 and NUBPL Expression Levels and Gender.

Figure 4

Biological functions of NUBPL. (A, B) GO and KEGG enrichment analysis of differentially expressed genes in high versus low expression cohorts of NUBPL. (C) ClueGO analysis of the top 100 differentially expressed genes (DEGs) (D, E) GSEA analysis of Gene Ontology and KEGG gene sets. (F, G) Quantification of pathway correlation using the GSVA algorithm.

Figure 5

Molecular and immunological characteristics between high and low NUBPL expression cohorts. (A) Box plot that contrasts the NUBPL and ESTIMATE scores, as well as immune and stromal assessments. (B) Violin plot illustrates the disparities in immune cell infiltration between groups with high and low expression levels. (C) Dot plot is also presented to compare the expression of four key immune checkpoints across the same high and low groups. The statistical significance, denoted as p < 0.05, underscores the reliability of these findings. (D) Scatter plot reveals the association between NUBPL expressi on and the concentrations of four chemokines. *p < 0.05.

Figure 6

The influence of NUBPL expression on the response to immunotherapy. (A) SubMap analysis heatmap. (B) Comparison of Immunophenoscore (IPS) between high and low expression groups. (C) Box plot depicting the association of NUBPL levels with TIDE, T cell dysfunction, and T cell exclusion metrics; significance levels: ***p < 0.001.

Figure 7

The three-dimensional and two-dimensional interaction diagrams of the NUBPL protein with QS11 (A), Imatinib (B), and AS601245 (C).

Figure 8

The impact of NUBPL on gastric cancer cell phenotype. (A) Western blotting assessment of NUBPL protein levels in AGS and MKN cell lines. (B) CCK8 assay for cellular proliferation. (C) Colony formation assay. (D) Annexin-V/PI assay for cellular apoptosis. (E) Transwell assay for cellular invasion and migration. (F) Cell scratch assay. ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 9

The impact of NUBPL on disulfide-induced mortality in gastric cancer cells. (A-D) The influence of NUBPL on disulfidptosis in gastric cancer cells under conditions of glucose supplementation and deprivation. (E, F) Observation of cell morphology via microscopy. (G) Immunofluorescence assessment of F-actin expression. (H-K) Detection of disulfide bonds in non-reducing SDS-PAGE. ns, not significant; *p < 0.05; ***p < 0.001; ****p < 0.0001.

Figure 10

The regulatory mechanism of NUBPL in disulfide-induced apoptosis in gastric cancer cells. (A) Western blot analysis of SLC7A11 expression. (B) Measurement of the NADP⁺/NADPH ratio. (C) Correlation of NUBPL with pivotal enzymes in the PPP pathway (G6PD, PGD, RPE, TALDO1, TKT). (D) Quantitative PCR analysis of the impact of NUBPL on the expression of enzymes in the pentose phosphate pathway. ns, not significant; *p < 0.05; **p < 0.01.