Bioinformatics Analysis Reveals the Vital Role of AKR1B1 in Immune Infiltration and Clinical Outcomes of Gastric Cancer

Abstract

Infiltrated immune cells are an important constitute of tumor microenvironment, which exert complex effects on gastric cancer (GC) pathogenesis and progression. By using weighted gene co-expression network analysis, integrating the data from The Cancer Genome Atlas-stomach adenocarcinoma and GSE62254, we identify Aldo-Keto Reductase Family 1 Member B (AKR1B1) as a hub gene for immune regulation in GC. Notably, AKR1B1 is associated with higher immune infiltration and worse histologic grade of GC. In addition, AKR1B1 is an independent factor for predicting the survival rate of GC patients. In vitro experiments further demonstrated that AKR1B1-overexpressed THP-1-derived macrophages promoted the proliferation and migration of GC cells. Taken together, AKR1B1 plays an important role in GC progression by regulating immune microenvironment, which could be a biomarker for predicting GC prognosis as well as a potential therapeutic target for GC treatment.

Introduction

Gastric cancer (GC) is a kind of malignant tumor with high morbidity and mortality, which is fifth most common cancer globally and the fourth leading cause of cancer deaths, responsible for 70,000 deaths worldwide annually (Poorolajal et al., 2020). The 5-year survival rate of GC is estimated <30%, since most patients are diagnosed at an advanced stage (Lordick et al., 2022; Wang et al., 2022b). Therefore, it is important to identify efficient biomarkers and therapy targets to improve the prognosis and treatment for GC.

Immune cell infiltration is an important part of tumor microenvironment, which participates in the survival, invasiveness, metastasis, and responsiveness to therapeutics of cancers (Turley et al., 2015). Recent studies revealed that tumor-infiltrated immune cells, including dendritic cells (DC), macrophages, neutrophils, T cells, and B cells, have complex effects on the GC pathogenesis and progression (Zavros and Merchant, 2022). Under ideal conditions, the cancer-related immunity could totally eradicate malignant cells (Ferrone and Dranoff, 2010). However, in circumstances of chronic inflammation, accumulation of reactive oxygen species (ROS), and immunosuppression, immune infiltration promotes GC development (Ferrone and Dranoff, 2010; Ma et al., 2022). Therefore, systemic analysis on the relationship between immune infiltration and GC using bioinformatics tools is useful and necessary.

Estimation of Stromal and Immune cells in Malignant Tumors using Expression data (ESTIMATE) algorithm is a method utilizing gene expression signatures to infer the fraction of stromal and immune cells in tumors (Yoshihara et al., 2013). Moreover, weighted gene co-expression network analysis (WGCNA) is a systematic strategy integrating highly correlated genes into several modules and investigating the association between gene groups and clinical/molecular phenotypes (Langfelder and Horvath, 2008). In this study, we analyzed genes that are highly associated with immune infiltration by ESTIMATE score and WGCNA using the data from The Cancer Genome Atlas (TCGA)-stomach adenocarcinoma (STAD) and GSE62254 and identify Aldo-Keto Reductase Family 1 Member B (AKR1B1) as a hub gene.

Aldo-keto reductase family 1 (AKR1) is known as oxidoreductases, which reduce electrons from nicotinamide adenine dinucleotide phosphate and play an important role in cellular energy metabolism (Penning, 2015; Srivastava et al., 2005). AKR1B1 participates in the polyol pathway catalyzing the conversion of excess glucose to sorbitol (Saraswathy et al., 2014). Recently, pioneer researches revealed high expression of AKR1B1 was associated with numerous cancer types (Khayami et al., 2020). Overexpression of AKR1B1 was proven to promote the proliferation and migration of cancer cells of colorectal cancer, breast cancer, and pancreatic cancer by in vitro experiments (Khayami et al., 2020; Reddy et al., 2017; Taskoparan et al., 2017; Xiao et al., 2018). However, the role of AKR1B1 in GC remains largely unknown. In this article, we investigated AKR1B1's functional annotation, association with immune infiltration, relationship to clinicopathological characteristics, and predictive value for GC prognosis.

Materials and Methods

Data collection

The RNA sequencing data of TCGA-STAD cohort containing 443 samples in fragments per kilobase per million form were downloaded from TCGA database. In addition, the GSE62254 dataset, including 300 samples, was obtained from the Gene Expression Omnibus (GEO) database. This study contains no data from original clinical trials. All human samples analyzed in this article were retrieved from public databases (TCGA and GEO).

Estimation of stromal and immune cells in malignant tumor tissues using expression data

Estimation of stromal and immune cells in malignant tumor tissues using expression data (ESTIMATE) algorithm is a tool to evaluate the immune score, stromal score, and tumor purity of each sample, which was calculated in R with packages of “estimate” and “tidyverse” (Yoshihara et al., 2013). The ESTIMATE scores were then defined as phenotypic data for subsequent WGCNA analysis.

Weighted gene co-expression network analysis

The “WGCNA” package was used to identify key gene modules remarkably related to immune infiltration. The expression profiling of TCGA-STAD and GSE62254 cohorts were input for WGCNA. TOM matrix was generated in the condition of the soft-thresholding power = 4 and the scale-free R ² = 0.9. Then genes were clustered and the hub modules of TCGA-STAD and GSE62254 cohorts were identified, respectively.

Enrichment analyses and clinical significance evaluation of overlapped hub genes

The genes in hub modules of TCGA-STAD and GSE62254 cohorts were overlapped using jvenn (Bardou et al., 2014). Then “clusterProfiler” package was used to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses in R on the cutoff value of adj.p < 0.05. In addition, the Least Absolute Shrinkage and Selection Operator (LASSO) and Univariate Cox regression analyses were performed on overlapped genes using the “glmnet” package to determine possible hub genes (Tibshirani, 1996).

Gene set enrichment analyses for AKR1B1

According to AKR1B1 median expression, samples in TCGA-STAD cohort were divided into high and low expression groups. The differentially expressed genes (DEGs) were identified using “DESeq2” package in R. Then GO and KEGG were conducted and visualized with “clusterProfiler,” “GOplot,” and “ggplot2” packages. Moreover, GSEA analysis was adopted to find vital biological pathways between groups with different AKR1B1 expression.

Immunomodulation analysis for AKR1B1

Infiltration enrichment of 24 common immune cells in TCGA-STAD tissues were analyzed with single-sample Gene Set Enrichment Analysis (ssGSEA) and CIBERSORT algorithm (Newman et al., 2015). The association between AKR1B1 expression and immune cell infiltrates, as well as PDCD1 and CTLA4 expression was investigated using Spearman's correlation.

Clinical relevance investigation and a nomogram construction

Information of patients' clinical outcomes in TCGA-STAD cohort, including overall survival (OS), disease-specific survival (DSS), and progression-free interval (PFI), were collected. Prognosis value of AKR1B1 expression on clinical relevance factors was examined using Kaplan–Meier analysis, logistic regression analysis, and univariate/multivariate Cox regression analysis. Next, nomograms, which integrated AKR1B1 expression with other factors, were constructed as a prognostic model to evaluate the predictive probability of 1-, 2-, and 3-year OS, DSS, and PFI.

Cell culture and transfection of plasmid

MGC803 cell line (Human GC cell line) and THP-1 cell line (Human monocyte-macrophage cell line) were purchased from Yuanchuang Cooperation (China), which were cultured using normal medium (RPMI-1640 with 10% fetal bovine serum and 1% Penicillin/Streptomycin).

The plasmid overexpresses AKR1B1 was constructed by Genomeditech Cooperation (Shanghai, China). The transfection of AKR1B1-overexpressed and control plasmids to THP-1 cells was performed using Hieff Trans Suspension Cell-Free Liposomal Reagent (Yeasen Cooperation, China) according to the manufacturer's instructions.

In vitro co-culture experimental design

After transfection with AKR1B1 overexpressed and control plasmids, THP-1 cells were induced to differentiate into macrophages by co-culturing with 10 ng/mL PMA for 48 h (Sigma-Aldrich). Then the medium was collected for further use. THP-1 cells were then lysed for isolating total RNA using an RNA extraction kit (Yishan, China). Then complementary DNA was synthesized by HiScript III RT SuperMix (Vazyme, China). And mRNA levels of TNFα, IL1β, IL4, IL10, TGFβ, and GAPDH were assessed by real-time PCR using Taq Pro Universal SYBR qPCR Master Mix (Vazyme, China). Primers used in this research are listed in Supplementary Table S1.

The collected THP-1 cell medium was mixed with normal medium at a ratio of 1:2 as conditioned medium (CM) to culture MGC803 cell line as a co-culture system. Then cell viability of MGC803 cells was measured by CCK-8 (Dojindo, Japan) following the manufacturer's instructions. In addition, the cell migration ability of MGC803 cells was evaluated by scratch wound assay and transwell migration assay, as previously described (Chen et al., 2016; Wang et al., 2018).

Statistical analysis

Data are expressed as mean ± standard deviation. Two groups were compared by Student's t test and multiple groups were compared by one-way analysis of variance (ANOVA) using GraphPad prism software. Differences were considered significant at p < 0.05.

Results

Identification of modules by WGCNA

Clustered gene modules were constructed using the data from TCGA-STAD and GSE62254 cohort (Fig. 1A, B). Then the association between modules and ESTIMATE scores measured by Spearman correlation analysis was shown by heat maps (Fig. 1C, E). For GSE62254 cohort, 24 co-expression modules were constructed (Fig. 1C). Notably, greenyellow module had the highest positive association with StromalScore (Cor = 0.77, p = 2e-59), ImmuneScore (Cor = 0.92, p = 2e-126), and ESTIMATE Score (Cor = 0.98, p < 2e-200), as well as negative relationship with Tumor Purity (Cor = −0.91, p = 2e-118) (Fig. 1C, D).

FIG. 1.

Identification of modules correlated with the immune infiltration of GC. (A, B) Gene dendrogram and clustered modules in the TCGA-STAD cohort (A) and GSE62254 cohort (B). (C, E) Heat map showing Pearson correlation analysis of merged modules and immune infiltration in TCGA-STAD cohort (C) and GSE62254 cohort (E). (D, F) Scatter plot showing the relationship between module membership in the yellowgreen module in the TCGA-STAD cohort (D) and grey module in the GSE62254 cohort (F) with immune infiltration indicators (stromal score, immune score, and ESTIMATE Score and Tumor Purity). ESTIMATE, Estimation of Stromal and Immune cells in Malignant Tumors using Expression data; GC, gastric cancer; STAD, stomach adenocarcinoma; TCGA, The Cancer Genome Atlas.

For GSE62254 cohort, 28 co-expression modules were clustered, with the blue module having the strongest correlation with StromalScore (Cor = 0.77, p = 2e-59), ImmuneScore (Cor = 0.92, p = 2e-126), ESTIMATE Score (Cor = 0.92, p = 2e-119), and Tumor Purity (Cor = −0.91 p = 2e-118) (Fig. 1C, D). Therefore, 662 genes in the greenyellow module and 1745 genes in the blue module were screened as potential immune infiltration-related hub genes for further investigation in this study.

Enrichment analysis of the overlapped genes and identification of hub genes

Genes in the above greenyellow and blue modules were overlapped and screened to 131 potential key genes, which were input for enrichment analysis (Fig. 2A). GO analysis results showed that these genes were enriched into the process of organelle fission, nuclear division, spindle, and condense chromosomal region in terms of Biological Process (BP) and Cellular Component (CC) (Fig. 2B). Tubulin binding, Microtubule binding, Extracellular matrix structural constituent, and Chemokine receptor binding were mainly enriched terms in Molecular Function (MF) (Fig. 2B). In addition, KEGG analysis demonstrated that cell cycle, cytokine-cytokine receptor interaction, cellular senescence, oocyte meiosis, and chemokine signaling pathway were main processes related to overlapped genes (Fig. 2C).

FIG. 2.

Functional analyses of the ESTIMATE-based signature. (A) The hub immune infiltration-related genes were overlapped in the yellowgreen module and the dark grey module. (B) KEGG enrichment analysis. (C) GO enrichment analysis. (D, E) LASSO regression analysis. (F) Univariate Cox regression analysis of common hub genes (p < 0.05). GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; LASSO, Least Absolute Shrinkage and Selection Operator.

These 131 genes were further subjected to LASSO regression model using the data from TCGA-STAD cohort to determine the minimum l value (Fig. 2D, E). Then these genes were analyzed by Univariate Cox regression analysis model with OS from TCGA-STAD cohort as dependent variables. The results are shown in Supplementary Figure S1 and genes with p < 0.05 were listed in Figure 2F. Notably, AKR1B1 is the only gene with p < 0.001 in Univariate Cox regression analysis; thus, it was identified as a hub gene for immune infiltration of GC.

Functional annotation of AKR1B1-associated DEGs in GC

To determine AKR1B1-associated signaling pathways, TCGA-STAD datasets were divided into high and low AKR1B1 expression groups based on AKR1B1 median expression. DEGs were identified with the cutoff value of log2FC >1, adj.p < 0.05 (Fig. 3A). As shown in Figure 3B, the most significant terms of BP, CC, MF, and KEGG were humoral immune response, intermediate filament, receptor ligand activity, and cytokine-cytokine receptor interaction, respectively. To find GO or KEGG terms having related DEGs, a network was drawn, in which significant terms were Treg proliferation, leukocyte cell-cell adhesion, lysosomal lumen, vacuolar lumen, coreceptor activity, and immune network for lgA production (Fig. 3C).

FIG. 3.

Functional annotation of DEGs in GC patients with distinct AKR1B1 levels. (A) Volcano plot showing DEGs in GC patients with distinct AKR1B1 levels. (B) Bubble chart showing significant terms in GO and KEGG enrichment analysis (C) Visualization network showing significant terms and related genes in GO and KEGG enrichment analysis (D) Representative GSEA analysis. AKR1B1, Aldo-Keto Reductase Family 1 Member B; DEGs, differentially expressed genes; GSEA, Gene Set Enrichment Analysis.

We then performed GSEA analyses and representative significant terms are listed in Figure 4D. Signaling pathways, including Onder CDH1 Targets 2, Jaeger metastasis, Rickman tumor differentiation well versus poorly, and Lin silenced by tumor microenvironment, were remarkably downregulated in high AKR1B1 expression group (Fig. 3D).

FIG. 4.

Correlation of immune cell infiltration and AKR1B1 expression in patients with GC. (A) Associations between infiltration levels of 24 immune cell types and AKR1B1 expression by Spearman's analysis. (B) The most correlated immune cells, including iDC, macrophages, DC, Th1, NK CD56bright, and Th17 cell between high and low AKR1B1 expression groups, are shown. aDCs, activated dendritic cells; DCs, dendritic cells; iDCs, immature dendritic cells; NK, natural killer; pDCs, plasmacytoid dendritic cells; Tcm, T central memory; Tem, T effector memory; Tfh, T follicular helper; Tgd, T gamma delta; Th, T helper cells; Th1, T helper 1; Treg, regulatory T cells.

Association of AKR1B1 and immune cell infiltration in GC

To investigate the association between AKR1B1 and immune cell infiltration, ssGSEA method was adopted to reveal the proportion of immune cells in TCGA-STAD datasets according to high and low AKR1B1 expression groups. High AKR1B1 expression was positively associated with the accumulation of immature dendritic cells (iDC), macrophages, DC, T helper 1 (Th1) cells, mast cells, T follicular helper cells, T cells, eosinophils, activated dendritic cells, T effector memory; B cells, regulatory T cells (Treg), CD8 T cells, cytotoxic cells, natural killer (NK) cells, neutrophils, T gamma delta, plasmacytoid dendritic cells, and NK CD56 dim cells, but negatively with NK CD56 bright cells (Fig. 4A, B).

No significant correlationship was found between AKR1B1 expression and Th17 cells, Th2 cells, and T central memory cells (Fig. 4A). Then we also used CIBERSORT algorithm to assess immune cell infiltration and similar results were obtained (Fig. 5A, B). Finally, we investigated the relationship between PDCD-1/CTLA4 and AKR1B1 expression by Pearson correlation analysis and found that PDCD-1 (r = 0.267, p < 0.001) and CTLA4 (r = 0.226, p < 0.001) were positively associated with AKR1B1 (Fig. 5C, D). Taken above, these results demonstrated that AKR1B1 was related to increased immune infiltration and immunosuppressive effects in GC.

FIG. 5.

Immune cell landscape in TCGA-ESCC cohort with high and low AKR1B1 expression. (A) The bar plot showing the proportion of infiltrated immune cells calculated by the CIBERSORT algorithm. (B) The violin plot showing the difference between 24 infiltrated immune cells in the tumor microenvironment. (C, D) Relationship between AKR1B1 expression and PDCD1/CTLA4 expression by Spearman's correlation. ESCC, esophageal squamous cell carcinoma.

Association of AKR1B1 expression and clinicopathological characteristics in patients with GC

We explored the clinicopathological characteristics of GC patients with distinct AKR1B1 expression, as shown in Table 1. Compared to low AKR1B1 expression patients, individuals with high AKR1B1 expression had more severe T stages and worse histological grade. However, no significant difference was observed in gender, age, N stage, M stage, pathologic stage, and primary therapy outcome (Table 1). We also performed logistics analysis to investigate the correlation between AKR1B1 expression and clinicopathological characteristics (Table 2). Similarly, T stage and histological grade were positively associated with AKR1B1 levels (Table 2).

Table 1.

Clinicopathological Characteristics of Stomach Adenocarcinoma Patients with Different Aldo-Keto Reductase Family 1 Member B Expression

Characteristic	Level	Low AKR1B1 group (n = 187)	High AKR1B1 group (n = 188)	p
Gender	Female	63 (16.8%)	71 (18.9%)	0.474
Age	≤65	77 (20.8%)	87 (23.5%)	0.371
Age	>65	108 (29.1%)	99 (26.7%)
T stage	T1	15 (4.1%)	4 (1.1%)	0.050 ^*
	T2	35 (9.5%)	45 (12.3%)
	T3	87 (23.7%)	81 (22.1%)
	T4	49 (13.4%)	51 (13.9%)
N stage	N0	63 (17.6%)	48 (13.4%)	0.434
	N1	44 (12.3%)	53 (14.8%)
	N2	38 (10.6%)	37 (10.4%)
	N3	37 (10.4%)	37 (10.4%)
M stage	M0	168 (47.3%)	162 (45.6%)	0.646
M stage	M1	11 (3.1%)	14 (3.9%)
Pathologic stage	Stage I	34 (9.7%)	19 (5.4%)	0.085
	Stage II	52 (14.8%)	59 (16.8%)
	Stage III	79 (22.4%)	71 (20.2%)
	Stage IV	15 (4.3%)	23 (6.5%)
Primary therapy outcome	PD	34 (10.7%)	31 (9.8%)	0.731
	SD	11 (3.5%)	6 (1.9%)
	PR	3 (0.9%)	1 (0.3%)
	CR	125 (39.4%)	106 (33.4%)
Histologic grade	G1	5 (1.4%)	5 (1.4%)	<0.001 ^***
	G2	85 (23.2%)	52 (14.2%)
	G3	92 (25.1%)	127 (34.7%)

All significant factors with their p values are shown in bold. ^* p < 0.05, ^*** p < 0.001.

AKR1B1, Aldo-Keto Reductase Family 1 Member B; CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease.

Table 2.

Logistic Regression Analysis of Association Between Clinicopathological Characteristics and Aldo-Keto Reductase Family 1 Member B Expression in Stomach Adenocarcinoma Patients

Characteristics	OR	p
Gender (female vs. male)	1.194 (0.783–1.826)	0.410
Age (>65 vs. < = 65)	0.848 (0.562–1.277)	0.430
T stage (T2&T3&T4 vs. T1)	5.584 (1.820–24.315)	0.007 ^**
N stage (N1&N2&N3 vs. N0)	1.424 (0.908–2.242)	0.125
M stage (M1 vs. M0)	1.304 (0.577–3.020)	0.525
Primary therapy outcome (PR&CR vs. PD&SD)	1.019 (0.616–1.691)	0.942
Histologic grade (G3 vs. G1&G2)	2.286 (1.495–3.520)	<0.001 ^***

All significant factors with their p values are shown in bold. ^** p < 0.01, ^*** p < 0.001.

OR, odds ratio.

Predictive value of AKR1B1 for GC prognosis

We utilized Kaplan-Meier and Cox regression analyses to assess the predictive value of AKR1B1 on clinical outcomes. High AKR1B1 expression group was manifested with worse OS [hazard ratio (HR): 1.83, p < 0.001), DSS (HR: 1.61, p = 0.009), and PFI (HR: 1.69, p = 0.014) (Fig. 6A–C). In addition, AKR1B1 expression was an independent risk factor for OS (HR: 2.026, p < 0.001), DSS (HR: 2.170, p < 0.001), and PFI (HR: 1.906, p = 0.002) in multivariate Cox regression (Table 3). Notably, primary therapy outcome was also a risk factor for OS (HR: 4.103, p < 0.001), DSS (HR: 8.446, p < 0.001), and PFI (HR: 7.808, p < 0.001) (Table 3). Besides, age was associated with worse OS, and histologic grade was related to worse OS and PFI (Table 3).

FIG. 6.

Predictive value of AKR1B1 expression for clinical outcomes in patients with GC. The K-M analyses comparing OS (A), DSS (B), and PFI (C) between high and low AKR1B1 expression groups. DSS, disease-specific survival; K-M, Kaplan–Meier; OS, overall survival; PFI, progression-free interval.

Table 3.

Cox Regression Analysis for Clinical Outcomes in Stomach Adenocarcinoma Patients

Characteristics	HR for OS		HR for DSS		HR for PFI
Characteristics	Univariate	Multivariate	Univariate	Multivariate	Univariate	Multivariate
Gender (male vs. female)	1.267		1.573	1.48	1.638 ^*	1.52
Age (>65 vs. ≤65)	1.620 ^**	1.953 ^**	1.211		0.858
T stage (T4 vs. T1&T2&T3)	1.204		1.152		0.957
N stage (N1&N2&N3 vs. N0)	1.925 ^**	1.307	1.807 ^*	0.859	1.640 ^*	0.983
M stage (M1 vs. M0)	2.254 ^**	1.515	2.438 ^*	1.463	2.224 ^*	1.535
Pathologic stage (Stage III&IV vs. Stage I&II)	1.947 ^***	1.277	2.146 ^***	1.517	1.676 ^**	1.098
Treatment choices (chemotherapy vs. radiotherapy)	0.702		0.654		0.591
Primary therapy outcome (PD&SD vs. PR&CR)	4.103 ^***	4.704 ^***	7.990 ^***	8.446 ^***	7.808 ^***	8.525 ^***
Histologic grade (G3 vs. G1&G2)	1.353	1.530 ^*	1.338		1.540 ^*	1.527 ^*
AKR1B1 (high vs. low)	1.816 ^***	2.026 ^***	1.621 ^*	2.170 ^***	1.578 ^*	1.906 ^**

All significant factors with their p values are shown in bold. ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

DSS, disease-specific survival; HR, hazard ratio; OS, overall survival; PFI, progression-free interval.

Then prognostic nomograms were established by inputting all significant prognostic factors identified in multivariate Cox analysis, of which efficiency was examined by calibration curve. Age, primary therapy outcome, historical grade, and AKR1B1 expression were included to construct a predictive nomogram for OS with a C-index of 0.710 (Fig. 7A). Primary therapy outcomes and AKR1B1 expression were integrated in a nomogram for DSS (C-index: 0.781) (Fig. 7C). Moreover, primary therapy outcome, historical grade, and AKR1B1 expression were used to establish a predictive nomogram for OS, which had a C-index of 0.767 (Fig. 7E). In accord, the calibration curve of the nomogram of 1-, 2-, and 3-year clinical outcomes manifested satisfied fitting between the predicted and observed values (Fig. 7B, D, F).

FIG. 7.

Construction and validation of nomograms based on AKR1B1 expression. Nomograms of AKR1B1 expression-based risk scoring models for 1-, 2-, and 3-year OS (A), DSS (C), and PFI (E). Calibration plots validating the efficiency of nomograms for OS (B), DSS (D), and PFI (F).

AKR1B1-overexpressed macrophages promote the proliferation and migration of GC cells

To further demonstrate whether AKR1B1 could regulate the interaction between macrophages and GC cells, in vitro co-culture experiments were performed using THP-1 cells (human monocyte-macrophage cell line) and MGC803 cells (human gastric carcinoma cell line). We constructed AKR1B1-overexpressed THP-1-derived macrophages by plasmid transfection (Fig. 8A). Then it was shown that several cytokines, including IL1β, IL4, IL10, and TGFβ, increased after AKR1B1 was overexpressed (Fig. 8B, C).

FIG. 8.

The construction of AKR1B1-overexpressed THP-1-derived macrophages. The mRNA expression levels of AKR1B1 (A), inflammatory cytokines TNFα and IL1β (B), as well as anti-inflammatory cytokines IL-4, IL-10, and TGFβ (C) in negative control-overexpressed (OE-NC) and AKR1B1-overexpressed (OE- AKR1B1) THP-1-derived macrophages. n = 4 per group. Data are expressed as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001. OE-NC, negative control-overexpressed; SD, standard deviation.

To build a co-culture system, the medium of THP-1-derived macrophages was collected and mixed with normal medium at a ratio of 1:2 as CM to culture MGC803 cells for further experiments (Fig. 9A). CCK-8 assay showed that CM from macrophages with overexpressed AKR1B1 significantly enhances the proliferation of MGC803 cells (Fig. 9B). In addition, as evidenced by the transwell assay and scratch wound assay, AKR1B1-overexpressed macrophages notably promote the migration ability of EPCs (Fig. 9C–F). These experiments indicated that AKR1B1 could regulate macrophage phenotype to impact GC progression.

FIG. 9.

Coculture of THP-1 derived macrophages and MGC803 GC cells. (A) Schematic illustration. (B) Cell viability measured by CCK-8 kit; n = 6 per group. (C, D) GC cells' motility capacity in different groups evaluated by the scratch wound assay; n = 3 per group. (E, F) GC cell migration capacity in different groups measured by the transwell migration assay; n = 3 per group. Data are expressed as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001.

Discussion

Immune cells are an important part of tumor microenvironment and play a vital role in tumor biology (Barkley et al., 2022). The improper immune infiltration could cause lasted low-grade inflammation and regional immunosuppression (Greten and Grivennikov, 2019). Chronic inflammation caused by H. pylori, gastritis, aging, and unhealthy diet could trigger tissue repair reactions, which produce various cytokines, angiogenic cytokines, and immune regulatory molecules, and then create a microenvironment suitable for GC pathogenesis, progression, and metastasis (Lei et al., 2022; Zavros and Merchant, 2022).

In addition, GC cells express some immune checkpoint molecules, including PDCD1 and CTLA4, which could inhibit cytotoxic CD8+ T cells and promote the differentiation of immunosuppressive Treg and Breg cells (Hamilton et al., 2022). In this study, by analyzing genes enriched in immune infiltration-related WGCNA modules, we reported that these genes were highly associated with cell cycle, chemokine recognition, and cellular senescence. Then LASSO regression model and univariate Cox regression analysis were performed to further identify hub genes with clinical value. It was noticed that AKR1B1 expression was significantly related to survival rate in GC patients.

AKR1B1 is widely expressed in most tissues and regulates energy metabolism in physiological state (Bailly, 2022). However, recent studies have revealed that AKR1B1 is overexpressed in several cancers, including colorectal carcinoma, hepatocellular carcinoma, and lung cancer (Penning, 2015). The elevation of AKR1B1 upregulates the expression of ROS and prostaglandin by activating nuclear factor kappa-light-chain enhancer of activated B cells (NFκB), mammalian target of rapamycin, and protein kinase B, which lead to chronic inflammation and cancer progression (Khayami et al., 2020). In this research, we demonstrated that high AKR1B1 expression group has more immune cell infiltration, especially macrophages and iDCs. GO and KEGG analysis revealed AKR1B1 regulates Treg activity, chemokine receptor physiology, and tumor differentiation. Moreover, AKR1B1 expression is positively associated with PDCD1 and CTLA4 expression. These results support the idea that AKR1B1 is associated with creating an immunosuppressive tumor microenvironment.

To improve the prognosis and treatment of GC, many genetic and epigenetic signatures has been identified using bioinformatics. The team of TCGA Research Network has published a GC molecular classification system consisting of genomically stable (e.g., CDH1, RHOA), chromosomal instability (e.g., TP53), microsatellite instable (e.g., KRAS, PIK3A, ARID1A), and Epstein–Barr virus positive (e.g., CDKN2A) (Network, 2014; Panarese et al., 2017). Other epigenetic modifications such as DNA methylation (DNA methyl methyltransferase family), histone modifications (histone acetyl transferase and deacetylating enzyme family), and noncoding RNA (lncRNA, miRNA, and circRNA) were also used to characterize GCs (Sogutlu et al., 2022).

In addition, series of genes, including cystathionine beta-synthase (CBS), GLIS Family Zinc Finger 3 (GLIS3), and COMM Domain Containing 10 (COMM10), were reported as independent factors for the prognosis of GC (Ding et al., 2023; Zhao et al., 2023; Zhao et al., 2021). However, our understanding of GC pathogenesis, especially the role of immune environment, is still inadequate. More studies are still needed in this field to explore valuable biomarkers for GC.

In previous studies, AKR1B1 has been included in prognosis models consisting of multiple genes for GC (Chen et al., 2022; Wang et al., 2022a; Zhang et al., 2022). In addition, human protein atlas database has reported that higher AKR1B1 expression was associated with worse survival rate. However, the relationship between AKR1B1 and GC has not been analyzed by systemic bioinformatics methods. To further investigate the clinical value of AKR1B1 in GC, we then tested the relationship between AKR1B1 expression and clinicopathological characteristics. AKR1B1 is positively associated with T stage and histologic grade, further confirming its participation in tumor proliferation and immune infiltration.

However, N stage, M stage, and primary therapy outcome are not statistically correlated with AKR1B1, suggesting AKR1B1 may have limited effects in tumor metastasis and chemotherapy responses. We then demonstrated that AKR1B1 level is an independent factor for predicting the prognosis of GC patients. GC patients with high AKR1B1 expression manifested with significantly decreased OS, DSS, and PFI. Combined with primary therapy outcome, historic grade, and age, prognostic nomogram was established with good fitting between the predicted and observed survival rate of GC patients. Taken above, AKR1B1 could serve as a biomarker for GC.

Finally, to confirm the role of AKR1B1 in tumor immune components, we performed in vitro experiments by co-culturing THP-1-derived macrophages and MC803 GC cells. TNFα, IL-1β, IL-4, IL-10, and TGFβ were important cytokines secreted from macrophages and related to worse prognosis of GC (Zavros and Merchant, 2022). TNF-α, IL-1β, IL-10, and TGFβ directly promote the carcinogenesis, invasion, and metastasis of GC (Chen et al., 2019; Cui et al., 2020; Huang et al., 2014; Park et al., 2002). IL-4 levels were significantly higher in GC patients than healthy controls, indicating its pernicious role (Gabitass et al., 2011).

The AKR1B1-overexpressed macrophages showed higher levels of inflammatory cytokine IL1β, but not TNFα. In addition, the expression of immunosuppressive cytokines, including IL4, IL10, and TGFβ, increased in macrophages with upregulated AKR1B1. We further demonstrated that CM from macrophages with AKR1B1 overexpression promoted the proliferation and migration of GC cells. These results suggest that AKR1B1 influences immune environment, especially the cytokines from macrophages to assist GC progression.

Conclusion

In summary, this study identifies AKR1B1 as an important regulator in the immune infiltration and clinical outcomes of GC. Since AKR1B1 overexpression is found related to immunosuppression and worse survival, it could be a biomarker for predicting GC prognosis as well as a potential therapeutic target for GC treatment.

Data Availability Statement

The datasets in this study were downloaded from the TCGA database, GEO database, and IMvigor210CoreBiologies R package. The data of in vitro experiments presented in this study are available on request from the corresponding author.

Footnotes

Disclosure Statement

No competing financial interests exist.

Funding Information

This research was supported by Western Medicine Guidance Project of Shanghai Science and Technology Commission (19411970700) to X.Z.

Supplementary Material

Supplementary Figure S1

Supplementary Table S1

References

Bailly

. Moving toward a new horizon for the aldose reductase inhibitor epalrestat to treat drug-resistant cancer. Eur J Pharmacol, 2022; 931:175191; doi: 10.1016/j.ejphar.2022.175191

Bardou

, Mariette

, Escudié

, et al. jvenn: An interactive Venn diagram viewer. BMC Bioinform, 2014; 15(1):293; doi: 10.1186/1471-2105-15-293

Barkley

, Moncada

, Pour

, et al. Cancer cell states recur across tumor types and form specific interactions with the tumor microenvironment. Nat Genet, 2022; 54(8):1192–1201; doi: 10.1038/s41588-022-01141-9

Chen

, Shi

, Zhu

, et al. IL-10 secreted by cancer-associated macrophages regulates proliferation and invasion in gastric cancer cells via c-Met/STAT3 signaling. Oncol Rep, 2019; 42(2):595–604; doi: 10.3892/or.2019.7206

Chen

, Yuan

, Yu

, et al. Bioinformatics analysis and structure of gastric cancer prognosis model based on lipid metabolism and immune microenvironment. Genes (Basel), 2022; 13(9):1581; doi: 10.3390/genes13091581

Chen

, Zhu

, Lin

, et al. The novel EZH2 inhibitor, GSK126, suppresses cell migration and angiogenesis via down-regulating VEGF-A. Cancer Chemother Pharmacol, 2016; 77(4):757–765; doi: 10.1007/s00280-016-2990-1

Cui

, Zhang

, Cao

, et al. Cytokine TNF-α promotes invasion and metastasis of gastric cancer by down-regulating Pentraxin3. J Cancer, 2020; 11(7):1800–1807; doi: 10.7150/jca.39562

Ding

, Wang

, Chen

, et al. The gene regulatory molecule GLIS3 in gastric cancer as a prognostic marker and be involved in the immune infiltration mechanism. Front Oncol, 2023; 13:1091733; doi: 10.3389/fonc.2023.1091733

Ferrone

, Dranoff

. Dual roles for immunity in gastrointestinal cancers. J Clin Oncol, 2010; 28(26):4045–4051; doi: 10.1200/jco.2010.27.9992

10.

Gabitass

, Annels

, Stocken

, et al. Elevated myeloid-derived suppressor cells in pancreatic, esophageal and gastric cancer are an independent prognostic factor and are associated with significant elevation of the Th2 cytokine interleukin-13. Cancer Immunol Immunother, 2011; 60(10):1419–1430; doi: 10.1007/s00262-011-1028-0

11.

Greten

, Grivennikov

. Inflammation and cancer: Triggers, mechanisms, and consequences. Immunity, 2019; 51(1):27–41; doi: 10.1016/j.immuni.2019.06.025

12.

Hamilton

, Anholt

, Nelson

. Tumour immunotherapy: Lessons from predator-prey theory. Nat Rev Immunol, 2022; 22(12):765–775; doi: 10.1038/s41577-022-00719-y

13.

Huang

, Lan

, Wang

, et al. IL-1β-induced activation of p38 promotes metastasis in gastric adenocarcinoma via upregulation of AP-1/c-fos, MMP2 and MMP9. Mol Cancer, 2014; 13:18; doi: 10.1186/1476-4598-13-18

14.

Khayami

, Hashemi

, Kerachian

. Role of aldo-keto reductase family 1 member B1 (AKR1B1) in the cancer process and its therapeutic potential. J Cell Mol Med, 2020; 24(16):8890–8902; doi: 10.1111/jcmm.15581

15.

Langfelder

, Horvath

. WGCNA: An R package for weighted correlation network analysis. BMC Bioinform, 2008; 9:559; doi: 10.1186/1471-2105-9-559

16.

Lei

, Teng

, Tian

, et al. Signaling pathways and therapeutic interventions in gastric cancer. Signal Transduct Target Ther, 2022; 7(1):358; doi: 10.1038/s41392-022-01190-w

17.

Lordick

, Carneiro

, Cascinu

, et al. Gastric cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol, 2022; 33(10):1005–1020; doi: 10.1016/j.annonc.2022.07.004

18.

, Sun

, Liu

, et al. The immune microenvironment in gastric cancer: Prognostic prediction. Front Oncol, 2022; 12:836389; doi: 10.3389/fonc.2022.836389

19.

Network

CGAR

. Comprehensive molecular characterization of gastric adenocarcinoma. Nature, 2014; 513(7517):202–209; doi: 10.1038/nature13480

20.

Newman

, Liu

, Green

, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods, 2015; 12(5):453–457; doi: 10.1038/nmeth.3337

21.

Panarese

, De Vita

, Ronchi

, et al. Predictive biomarkers along gastric cancer pathogenetic pathways. Expert Rev Anticancer Ther, 2017; 17(5):417–425; doi: 10.1080/14737140.2017.1301207

22.

Park

, 2nd, Son

, Song

, et al. Role of TGF-beta 1 and TGF-beta type II receptor in gastric cancer. Korean J Intern Med, 2002; 17(3):160–166; doi: 10.3904/kjim.2002.17.3.160

23.

Penning

. The aldo-keto reductases (AKRs): Overview. Chem Biol Interact, 2015; 234:236–246; doi: 10.1016/j.cbi.2014.09.024

24.

Poorolajal

, Moradi

, Mohammadi

, et al. Risk factors for stomach cancer: A systematic review and meta-analysis. Epidemiol Health, 2020; 42:e2020004; doi: 10.4178/epih.e2020004

25.

Reddy

, Kumar

, Srinivasulu

, et al. Overexpression and enhanced specific activity of aldoketo reductases (AKR1B1 & AKR1B10) in human breast cancers. Breast, 2017; 31:137–143; doi: 10.1016/j.breast.2016.11.003

26.

Saraswathy

, Anand

, Kunnumpurath

, et al. Chromosomal aberrations and Exon 1 mutation in the AKR1B1 gene in patients with diabetic neuropathy. Ochsner J, 2014; 14(3):339–342.

27.

Sogutlu

, Pekerbas

, Biray Avci

. Epigenetic signatures in gastric cancer: Current knowledge and future perspectives. Expert Rev Mol Diagn, 2022; 22(12):1063–1075; doi: 10.1080/14737159.2022.2159381

28.

Srivastava

, Ramana

, Bhatnagar

. Role of aldose reductase and oxidative damage in diabetes and the consequent potential for therapeutic options. Endocr Rev, 2005; 26(3):380–392; doi: 10.1210/er.2004-0028

29.

Taskoparan

, Seza

, Demirkol

, et al. Opposing roles of the aldo-keto reductases AKR1B1 and AKR1B10 in colorectal cancer. Cell Oncol (Dordr), 2017; 40(6):563–578; doi: 10.1007/s13402-017-0351-7

30.

Tibshirani

. Regression shrinkage and selection via the lasso. J Royal Stat Soc Series B (Methodol), 1996; 58:267–288.

31.

Turley

, Cremasco

, Astarita

. Immunological hallmarks of stromal cells in the tumour microenvironment. Nat Rev Immunol, 2015; 15(11):669–682; doi: 10.1038/nri3902

32.

Wang

, Qi

, Sun

, et al. Integrative analysis identifies a TNFα-derived gene signature for predicting prognosis, tumor immunity, and treatment sensitivity in gastric cancer. Front Genet, 2022a;13:882519; doi: 10.3389/fgene.2022.882519

33.

Wang

, Tang

, Wang

, et al. EPCR promotes MGC803 human gastric cancer cell tumor angiogenesis in vitro through activating ERK1/2 and AKT in a PAR1-dependent manner. Oncol Lett, 2018; 16(2):1565–1570; doi: 10.3892/ol.2018.8869

34.

Wang

, Han

, Xue

, et al. Nationwide gastric cancer prevention in China, 2021–2035: A decision analysis on effect, affordability and cost-effectiveness optimisation. Gut, 2022b;71(12):2391–2400; doi: 10.1136/gutjnl-2021-325948

35.

Xiao

, Jin

, Jiao

, et al. β2-AR regulates the expression of AKR1B1 in human pancreatic cancer cells and promotes their proliferation via the ERK1/2 pathway. Mol Biol Rep, 2018; 45(6):1863–1871; doi: 10.1007/s11033-018-4332-3

36.

Yoshihara

, Shahmoradgoli

, Martínez

, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun, 2013; 4:2612; doi: 10.1038/ncomms3612

37.

Zavros

, Merchant

. The immune microenvironment in gastric adenocarcinoma. Nat Rev Gastroenterol Hepatol, 2022; 19(7):451–467; doi: 10.1038/s41575-022-00591-0

38.

Zhang

, Dong

, Liu

, et al. Prognosis and tumor immune microenvironment of patients with gastric cancer by a novel senescence-related signature. Medicine (Baltimore), 2022; 101(40):e30927; doi: 10.1097/md.0000000000030927

39.

Zhao

, Zhao

, Ding

, et al. Low CBS expression can identify patients who benefit from adjuvant chemotherapy in gastric cancer. Expert Rev Anticancer Ther, 2021; 21(11):1287–1298; doi: 10.1080/14737140.2021.1962298

40.

Zhao

, Lin

, Cheng

, et al. Comprehensive analysis of COMMD10 as a novel prognostic biomarker for gastric cancer. PeerJ, 2023; 11:e14645; doi: 10.7717/peerj.14645

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.28 MB

0.02 MB