Elevated SLC2A1 Expression Correlates with Poor Prognosis in Patients with Surgically Resected Lung Adenocarcinoma: A Study Based on Immunohistochemical Analysis and Bioinformatics

Abstract

Lung adenocarcinoma (LUAD) accounts for an increasing proportion of non–small-cell lung cancer and an increasing number of cancer-related deaths worldwide. However, few biomarkers are available for prognosis and patient stratification. In all eight datasets from the Oncomine and The Cancer Genome Atlas (TCGA) LUAD cohorts, solute carrier family 2 member 1 (SLC2A1) was significantly more highly expressed in LUAD tissue than in normal lung tissue. High SLC2A1 expression was also significantly (p < 0.05) associated with a poor prognosis in stage I, II, and III subgroups using the Kaplan–Meier plotter. In the National Cancer Center of China (NCC) cohort, SLC2A1 expression correlated significantly (p < 0.05) with several parameters, including sex, smoking history, tumor size, tumor differentiation, T stage, N stage, and pathologic TNM stage. Moreover, multivariate Cox regression indicated SLC2A1 to be an independent prognostic factor (p < 0.05) in both TCGA and NCC cohorts. Eleven hallmark pathways were significantly enriched (p < 0.01, false discovery rate <0.25) in the high-SLC2A1 expression group. SLC2A1 is a promising biomarker that can be used to predict the prognosis of LUAD.

Introduction

Lung cancer is now the leading cause of cancer-related death in both males and females, and approximately four-fifths of lung cancer patients have non–small-cell lung cancer (NSCLC) (Torre et al., 2015; Chen et al., 2016). In recent years, lung adenocarcinoma (LUAD) has accounted for an increasing proportion of NSCLC, posing a great challenge to global health (Mao et al., 2016). With the development of molecular genetics, several key events in tumorigenesis and the development of lung cancer have been identified, including both the activation of oncogenes and inactivation of tumor suppressor genes (Barr Kumarakulasinghe et al., 2015; Bol et al., 2015; Yuan et al., 2015; Kalemkerian et al., 2018; Lindeman et al., 2018; Rizvi et al., 2018). Through this deep insight, the emergence of targeted therapy and immunotherapy has brought new hope for treatments, including EGFR tyrosine kinase inhibitors (EGFR TKIs) (Mok et al., 2017; Soria et al., 2018), receptor tyrosine kinase (RTK) inhibitors (Liao et al., 2015; Steuer et al., 2015) and immune checkpoint blockade (Postow et al., 2015; Lee et al., 2017). However, due to great heterogeneity, only a portion of patients benefit from these treatments, and overall survival remains disappointing. Therefore, additional molecular biomarkers are urgently needed.

It has long been known that tumor cells can promote proliferation and survival by altering metabolic patterns (Griffin and Shockcor, 2004; DeBerardinis et al., 2008; Boroughs and DeBerardinis, 2015). One of the most famous features of tumor cells is the Warburg effect (Liberti and Locasale, 2016), which is characterized by increased glucose uptake and the fermentation of glucose to lactate. Although the benefits of this phenomenon to tumor cells are not completely understood, studies have shown that it is involved in several functions, including supporting the biosynthetic requirements of rapid proliferation, providing an advantage for cell growth in the tumor microenvironment, and conferring cellular signals to tumor cells (Liberti and Locasale, 2016). Solute carrier family 2 member 1 (SLC2A1) encodes a major glucose transporter in the mammalian blood–brain barrier and is a crucial mediator of the Warburg effect (Lopez-Serra et al., 2014; Ooi and Gomperts, 2015). Recent studies have revealed that SLC2A1 is overexpressed and promotes biological processes involved in several cancer types, including glioma, gastric cancer, colorectal cancer, and lung cancer (Yan et al., 2015; Wang et al., 2017; Yang et al., 2017; Shi et al., 2019). Nonetheless, its potential as a biomarker in LUAD has not been carefully evaluated.

In general, the development of high-throughput sequencing has profoundly altered our understanding of tumorigenesis, metastasis, and chemoresistance (Stark et al., 2019). Moreover, a large number of public databases are available for exploration. Indeed, several biomarkers have been identified through a data mining approach and validated in independent cohorts. The Cancer Genome Atlas (TCGA, www.cancer.gov/tcga) is one of the most complete public cancer databases and contains genomic, epigenomic, transcriptomic, and clinical information, and utilization of these data is beneficial for gaining new insight into the molecular characteristics of LUAD patients.

The current study aimed to explore the prognostic value of the major glucose transporter SLC2A1 in LUAD using both public databases and an independent cohort from our hospital. The expression level of SLC2A1 between tumor and paired normal tissues was analyzed first, after which its association between SLC2A1 expression and clinicopathological parameters was assessed. Kaplan–Meier (KM) curve and Cox multivariate regression analyses were also performed in the LUAD cohort of TCGA and our own cohort to evaluate the potential for SLC2A1 to predict prognosis. Finally, gene set enrichment analysis (GSEA) was conducted to identify SLC2A1-related pathways and determine potential targets for further study.

Materials and Methods

Patient samples

All clinical patient samples in this study were collected between June 2006 and June 2014 at the National Cancer Center/Cancer Hospital, CAMS. A total of 415 patients with LUAD who underwent R0 resection were included in the analysis. The entire enrollment process is shown in Figure 1. All patients provided informed consent before surgery. The collected clinical information included sex, age, tumor location, tumor differentiation status, T stage, lymph node metastasis, and TNM stage. All of the specimens were pathologically confirmed by two experienced pathologists and stored properly in our biobank. The 8th edition of the TNM staging system was used to guide pathological classification of the primary tumor and the degree of lymph node metastasis.

FIG. 1.

Flowchart of the enrollment process.

The current study was conducted in accordance with the Declaration of Helsinki. The Clinical Research Ethics Committee of the National Cancer Center/Cancer Hospital, CAMS approved this study (19/305-2089). Patients were followed up in the outpatient department every 3–6 months for the first 2 years after surgery and then annually until March 4, 2019. The follow-up material consisted of medical history, physical examinations, and chest computed tomography.

mRNA expression and clinical information in the dataset from TCGA

RNA-seq expression profiles and clinical information of the LUAD cohort of TCGA (535 patients) were downloaded from the Genomic Data Commons (GDC) portal (https://portal.gdc.cancer.gov). The fragments per kilobase of transcript per million mapped reads value was first transformed into transcripts per million (TPM) for between-sample comparisons. Clinical data, including age, sex, race, smoking history, T stage, N stage, M stage, TNM stage, overall survival time, and survival status, were extracted.

SLC2A1 expression analysis using Oncomine and gene expression profiling interactive analysis

Oncomine is a web-based application containing more than 700 independent datasets with curated data and consistent oncological terms and interpretations (Rhodes et al., 2004). Eight LUAD datasets, including both tumor and normal tissues were extracted for SLC2A1 expression analysis. Gene expression profiling interactive analysis (GEPIA) is another newly developed website that uses RNA-seq data from TCGA and the Genotype-Tissue Expression (GTEx) project (Tang et al., 2017); it is commonly used to compare gene expression profiles between different cancer types and to perform survival analysis. In this study, we used both Oncomine and GEPIA to preliminarily explore SLC2A1 expression. In addition, survival analysis of the LUAD cohort of TCGA was conducted using GEPIA. The median TPM value was chosen as the cutoff point for grouping.

Survival analysis of SLC2A1 in the lung cancer database with the KM plotter

The KM plotter is an online survival analysis tool that integrates datasets from the Gene Expression Omnibus (GEO), European Genome-phenome Archive (EGA) and TCGA (Győrffy et al., 2013). It combines gene expression and clinical data and calculates both the hazard ratio (HR) and log-rank p value. In the present study, the KM plotter was used to explore the prognostic value of SLC2A1 expression in different cohorts and in subgroups stratified by TNM stage. The median expression value was used as the cutoff.

GSEA of SLC2A1 and function of the SLC2A1 network

To identify pathways closely related to SLC2A1, GSEA was performed on data for 515 LUAD patients from TCGA. GSEA is a powerful computational method that identifies a significantly aggregated biological function or molecular pathway between two different biological states (Subramanian et al., 2005). Patients were divided into groups of high and low expression based on the median expression level of SLC2A1. Hallmark gene sets containing 50 gene sets were employed for GSEA. According to the recommended parameters, the number of permutations was set to 1000, and the enrichment statistic was set to “weighted.” Pathways with p < 0.01 and FDR <0.25 were considered significant, as listed in the Results section.

The Search Tool for the Retrieval of Interacting Genes/Proteins (STRING, version 11.0) online database (Franceschini et al., 2012) was employed to construct a protein–protein interaction (PPI) network centered on SLC2A1 and for functional annotation using Gene Ontology (GO) (Ashburner et al., 2000) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (Kanehisa, 2002). An interaction score >0.4 was considered significant.

Tissue microarrays

Formalin-fixed paraffin-embedded (FFPE) samples from LUAD patients stained with Hematoxylin and Eosin (H&E) were reviewed independently by two experienced pathologists. A representative area, ∼4 μm thick, harboring a high tumor/stroma ratio was selected for each sample. In total, the tissue microarray (TMA) contained 415 LUAD cases from the National Cancer Hospital, CAMS.

Immunohistochemistry of SLC2A1

TMAs were first deparaffinized, rehydrated, and treated with 2 N HCl for 15 min, followed by 100 mM Tris-HCl (pH 8.5) for 10 min. The TMAs were then blocked with 3% H₂O₂ for 30 min and goat serum at room temperature for 30 min. After blocking, a rabbit anti-SLC2A1 polyclonal antibody (1:2000, HPA004117; Sigma-Aldrich, St. Louis, MO) was added and incubated at 4°C overnight, followed by a polyclonal peroxidase-conjugated anti-rabbit IgG (Zhongshanjinqiao, Beijing, China) at room temperature for 20 min according to the manufacturer's instructions.

Immunohistochemistry (IHC) staining was evaluated and scored based on the density and percentage of SLC2A1-positive tumor cells by two experienced pathologists who were blinded to the clinical information of the LUAD patients. An immunoreactive score for each case was calculated based on the product of the following multipliers: (1) scores representing staining intensity (no staining = 0, weak = 1, moderate = 2, strong = 3) and (2) scores representing the positive staining area (0–25% = 0, 25–50% = 1, >50% = 2). In the current study, tissues with scores ≤1 were grouped together and considered negative for SLC2A1; tissues with scores ≥2 were considered positive for SLC2A1.

Statistical analyses

Statistical analyses were performed by R version 3.5.3 (The R Foundation for Statistical Computing). Supplementary Table S1 of clinical data and SLC2A1 expression was analyzed using the χ² test or Fisher's exact test. Survival analysis was performed using the KM method, and p values were calculated using the log-rank test. Associations between SLC2A1 expression values and clinicopathological characteristics were determined by the Kruskal–Wallis test and Wilcoxon test. Multivariate Cox regression was applied to identify independent prognostic variables, and the median TPM value was set as the cutoff. A p value <0.05 was considered statistically significant.

Results

Expression analysis of SLC2A1 in LUAD

SLC2A1 expression analysis was first conducted in various cancer types using the Oncomine database. The results showed SLC2A1 to be significantly highly expressed in several tumor types, including bladder cancer, breast cancer, colorectal cancer, esophageal cancer, gastric cancer, head and neck cancer, kidney cancer, leukemia, lung cancer, lymphoma, ovarian cancer, pancreatic cancer, and many other cancers (Fig. 2A). TCGA pan-cancer SLC2A1 expression analysis using GEPIA revealed similar results (i.e., SLC2A1 expression was elevated in multiple cancer types) (Fig. 2B). SLC2A1 expression was significantly higher in tumor tissues than in normal tissues in all datasets (Table 1). Significantly upregulated SLC2A1 expression was also found in cancerous tissue in the LUAD cohort from TCGA (Fig. 2C). In addition, the KM curve showed that the prognosis of LUAD patients with high SLC2A1 expression was significantly worse than that of LUAD patients with low SLC2A1 expression (Fig. 2D).

FIG. 2.

Aberrant SLC2A1 expression and prognostic value in cancer. (A) Expression profile of SLC2A1 in different cancer types by Oncomine. The numbers indicated quantity of datasets with significant (p < 0.01) SLC2A1 overexpression (red) or downexpression (blue). (B) Expression of SLC2A1 in different cancer types by GEPIA. (C) Abnormally high expression of SLC2A1 in LUAD compared with normal lung tissues. Normal lung tissues were derived from the TCGA and the GTEx projects. *p < 0.05. (D) Overall survival analysis using Kaplan–Meier curve in TCGA LUAD cohort. GEPIA, gene expression profiling interactive analysis; GTEx, Genotype-Tissue Expression; LUAD, lung adenocarcinoma; SLC2A1, solute carrier family 2 member 1; TCGA, The Cancer Genome Atlas.

Table 1.

The Significant Changes of SLC2A1 Expression in Transcription Level Between Lung Adenocarcinoma and Normal Tissues

References	p	Fold change	Rank (%)	Tumor	Normal
Su Lung	1.12E-10	3.160	1	27	30
Selamate Lung	2.39E-23	5.153	1	58	58
Hou lung	2.68E-14	4.801	1	45	65
Okayama Lung	5.86E-21	2.843	1	226	20
Beer Lung	5.99E-8	3.190	2	86	10
Landi Lung	2.15E-16	2.086	2	58	49
Yamagata Lung	0.008	3.792	6	9	2
Garber Lung	0.002	5.648	11	38	6

Source: Oncomine database.

SLC2A1, solute carrier family 2 member 1.

Association between SLC2A1 expression and clinicopathological factors in the cohort from TCGA

Comparison of SLC2A1 expression grouped by clinical indicators revealed consistent results based on TCGA LUAD data. The clinicopathological factors that were extracted for analysis included age, sex, race, smoking history, N stage, and pathologic stage. In multiple comparisons, all six tests showed great significance (p < 2.2e-16) (Fig. 3A–F). Furthermore, after stratification by different clinical factors, significant differences were observed between each tumor subgroup and the reference/normal group (p < 0.05) (Fig. 3A–F). Therefore, to further validate the prognostic value of SLC2A1, multivariate Cox regression analysis (Table 2) was performed, and the results illustrated that T stage (HR: 1.310, 95% confidence interval [CI]: 1.079–1.591), N stage (HR: 2.319, 95% CI: 1.711–3.142), and SLC2A1 expression (HR: 1.152, 95% CI: 1.016–1.306) were independent prognostic factors (p < 0.05) for the LUAD cohort from TCGA. In contrast, M stage was not an independent prognostic factor in the multivariate Cox model.

FIG. 3.

The relationship between SLC2A1 expression and clinical parameters (TCGA). Association between SLC2A1 relative expression and (A) age, (B) gender, (C) race, (D) smoke history, (E) N stage, (F) pathologic stage. *p < 0.05, ***p < 0.001, ****p < 0.001. Normal tissues were set as reference group for comparisons in each graphic.

Table 2.

Multivariate Cox Proportional Hazards Regression on Lung Adenocarcinoma Overall Survival

Variables	HR	95% CI	p
T
T₁ vs. T₂ vs. T₃ vs. T₄	1.310	1.079–1.591	0.006
N
N₀ vs. N_1–3	2.319	1.711–3.142	0.000
M
M₀ vs. M₁	1.577	0.878–2.834	0.128
SLC2A1 expression	1.152	1.016–1.306	0.028

Source: The Cancer Genome Atlas.

CI, confidence interval; HR, hazard ratio.

Prognostic analysis of SLC2A1 using the KM plotter

A total of 720 LUAD patients with SLC2A1 probe information were included in analysis using KM plotter. The KM curve containing all patients demonstrated that high SLC2A1 expression was significantly (p < 1.2e-10) (Fig. 4A) associated with poor overall survival, and the role of SLC2A1 in prognosis was consistent after the patients were stratified into different subgroups based on TNM stage and N stage. High SLC2A1 expression was significantly (p < 0.05) associated with a poor prognosis in the stage I, stage II, stage III, N₀, and N₁ subgroups (Fig. 4B–F).

FIG. 4.

SLC2A1 expression analysis and survival analysis of LUAD patients in KM-Plot lung cancer database. (A) All 720 LUAD patients; (B) stage I; (C) stage II; (D) stage III; (E) N₀; (F) N₁. KM, Kaplan–Meier.

Correlations between SLC2A1 expression and clinicopathological parameters in the National Cancer Center of China cohort

As shown in Table 3, χ² tests were employed to assess correlation between SLC2A1 expression (high and low) and clinicopathological parameters, and the results revealed SLC2A1 expression to be significantly associated with sex, smoking history, tumor size, tumor differentiation, T stage, N stage, pathologic TNM stage, and therapeutic strategy. It can also be seen from the table that patients with high levels of SLC2A1 expression tended to present at advanced stages, which suggests that SLC2A1 is linked to patient prognosis.

Table 3.

Correlations Between SLC2A Expression and Clinicopathological Parameters of 415 Patients with Lung Adenocarcinoma

Category	Cases, number (%)	SLC2A1 expression		p
Category	415 (100)	Low (n = 357)	High (n = 221)	p
Age (years)				1.000
≤60	216 (52.0)	137	79
>60	199 (48.0)	127	72
Gender				0.024
Male	233 (56.1)	137	96
Female	182 (43.9)	127	55
Smoking				0.001
Ever	232 (55.9)	100	83
Never	183 (44.1)	164	68
Tumor length (cm)				0.001
≤4	280 (67.5)	199	81
>4	135 (32.5)	65	70
Differentiation				0.000
Well	57 (13.7)	49	8
Moderate	159 (38.3)	105	54
Poor	199 (48.0)	110	89
T stage				0.000
T₁	174 (41.9)	118	39
T₂	102 (24.6)	115	77
T₃	137 (33.0)	15	26
T₄	2 (0.5)	16	9
N stage				0.014
N₀	174 (41.9)	123	51
N₁	102 (24.6)	65	37
N₂	139 (33.5)	76	63
TNM stage				0.002
I	154 (37.1)	111	43
II	102 (24.6)	68	34
III	159 (38.3)	85	74
Treatment				0.038
Surgery only	170 (41.0)	117	53
Surgery with chemotherapy	159 (38.3)	89	70
Surgery with chemoradiotherapy	86 (20.7)	58	28

Boldface indicates p value < 0.05.

Validation of the prognostic value of SLC2A1 in the NCC cohort

A total of 415 LUAD patients from our NCC cohort were divided into high (n = 151) and low SLC2A1 expression groups according to IHC staining results. Representative photomicrographs of H&E and SLC2A1 immunohistochemical staining are shown in Figure 5. A KM curve was generated using all patients and subgroups. The results first revealed that high SLC2A1 expression was significantly (p < 0.001) (Fig. 6A) associated with poor overall survival. In subgroup analysis, SLC2A1 remained a significant indicator (p < 0.05) of a poor prognosis in stage I, stage II, stage III, N₀, and N₊ (patients with lymphatic metastasis) subgroups (Fig. 6B–F). To further verify its independent prognostic value, univariate and multivariate Cox regression analyses were performed (Table 4). In the univariate analysis, overall survival was significantly associated with age, sex, smoking history, tumor size, tumor differentiation, T stage, lymph node metastasis, pathologic TNM stage, therapeutic strategy, and SLC2A1 expression (p < 0.05). In the subsequent multivariate analysis, factors, including patient age, tumor size, lymph node metastasis, therapeutic strategy, and SLC2A1 expression remained significantly associated with overall survival.

FIG. 5.

Representative photomicrographs of LUAD TMA sections. (A, B) H&E staining of LUAD tissues with SLC2A1 expression. (C, D) Immunohistochemical staining for SLC2A1 in LUAD with negative result. (E, F) Immunohistochemical staining for SLC2A1 in LUAD with positive result. Left: low-power field; right: high-power field. H&E, Hematoxylin and Eosin; TMA, tissue microarray.

FIG. 6.

Kaplan–Meier curves showing survival of the 415 patients with LUAD according to SUSD2 expression. (A) All 415 patients; (B) stage I; (C) stage II; (D) stage III; (E) N₀; (F) N₊.

Table 4.

Univariate Analysis and Multivariate Analysis of Risk Factors for Prognosis of 415 Lung Adenocarcinoma Patients

	Univariate analysis			Multivariate analysis
	p	HR	95% CI	p	HR	95% CI
Age (≤60, >60 years)	0.002	1.537	1.172–2.016	0.001	1.796	1.361–2.372
Gender (female, male)	0.046	0.755	0.574–0.995	0.690	0.926	0.635–1.350
Smoking (ever, never)	0.009	1.431	1.093–1.873	0.487	1.140	0.788–1.650
Tumor length (cm)
≤4
>4	0.002	2.471	1.884–3.240	0.001	1.920	1.403–2.627
Differentiation (well/moderate, poor)	0.001	1.887	1.435–2.483	0.411	1.134	0.841–1.529
T stage (T₁–T₂, T₃–T₄)	0.002	1.919	1.390–2.648	0.634	0.910	0.616–1.344
Lymph node metastasis (negative, positive)	0.001	2.775	2.025–3.749	0.001	3.357	1.994–5.653
TNM stage (I–II, III)	0.001	2.456	1.873–3.220	0.293	1.222	0.841–1.777
Treatment (surgery only, surgery with adjuvant therapy)	0.001	1.990	1.478–2.680	0.004	0.515	0.328–0.809
SLC2A1 expression (negative, positive)	0.001	2.762	2.106–3.623	0.001	2.324	1.749–3.090

Boldface indicates p value < 0.05.

GSEA and functional annotation of the SLC2A1 PPI network

A PPI network was constructed using 10 genes identified from the STRING database significantly associated with SLC2A1, namely, STOM, GIPC1, LDHA, HIF1A, AKT1, TP53, SMUG1, HK2, VEGFA, and INS (Fig. 7A). All of these genes retained an interactive score >0.4. To investigate the potential role of this network, functional annotation analyses (GO and KEGG) were carried out, and the significant enrichment terms are shown in Figure 7B. To investigate the biological processes in which SLC2A1 may be involved, GSEA was performed, and the results showed glycolysis, the G2M checkpoint, MTORC1 signaling, and hypoxia to be the most significantly associated processes (Fig. 8A–D and Table 5). Detailed information on each biological process is listed in Tables 6 –9. The top 100 genes that were significantly positive or negatively correlated with SLC2A1 are displayed in Figure 8E.

FIG. 7.

PPI network and functional annotation. (A) PPI network construction using SLC2A1 and significantly correlated genes (STOM, GIPC1, LDHA, HIF1A, AKT1, TP53, SMUG1, HK2, VEGFA, INS). (B) GO terms and KEGG analysis using DAVID revealed significant biologic processes or pathways. DAVID, Database for Annotation, Visualization, and Integrated Discovery; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PPI, protein–protein interaction.

FIG. 8.

Gene Set Enrichment Analysis centered on SLC2A1. (A–D) The most significantly associated pathways with SLC2A1. (A) Glycolysis. (B) G2M checkpoint. (C) MTORC1 signaling. (D) Hypoxia. (E) Heat Map of the top 100 genes that were most significantly positive or negative correlated with SLC2A1.

Table 5.

Top Biological Processes Enriched in Lung Adenocarcinoma Centered on SLC2A1

Rank	Name of pathway	ES	NES	NOM p-value	FDR q-value	FWER p-value
1	GLYCOLYSIS	0.62	2.42	0.000	0.000	0.000
2	G2M_CHECKPOINT	0.75	2.22	0.000	0.001	0.001
3	MTORC1_SIGNALING	0.66	2.22	0.000	0.000	0.001
4	HYPOXIA	0.55	2.17	0.000	0.002	0.006
5	E2F_TARGETS	0.78	2.16	0.000	0.002	0.006
6	MYC_TARGETS_V1	0.72	2.04	0.000	0.007	0.024
7	UNFOLDED_PROTEIN_RESPONSE	0.55	1.98	0.000	0.013	0.043
8	MITOTIC_SPINDLE	0.55	1.93	0.000	0.016	0.063
9	MYC_TARGETS_V2	0.71	1.89	0.002	0.019	0.087
10	EPITHELIAL_MESENCHYMAL_TRANSITION	0.57	1.80	0.022	0.035	0.147
11	PI3K_AKT_MTOR_SIGNALING	0.41	1.73	0.010	0.054	0.222

ES, enrichment score; NES, normalized enrichment score; NOM, nominal p-value; FDR, false discovery rate; FWER, familywise-error rate.

Table 6.

Gene Set Enrichment Analysis Detail of Hallmark Glycolysis Pathway

Probe	Rank metric score	Running ES	Core enrichment
PKM	0.941	0.0179	Yes
PGAM1	0.899	0.0350	Yes
SLC16A3	0.873	0.0517	Yes
PGK1	0.838	0.0675	Yes
TPI1	0.818	0.0830	Yes

ES, enrichment score; Running ES is the enrichment score for this set at this point in the ranked list of genes.

Table 7.

Gene Set Enrichment Analysis Detail of Hallmark G2M Checkpoint

Probe	Rank metric score	Running ES	Core enrichment
CCNA2	0.887	0.0117	Yes
PRC1	0.884	0.0237	Yes
CCNB2	0.864	0.0353	Yes
PLK1	0.863	0.0470	Yes
KIF4A	0.830	0.0581	Yes
CDC20	0.823	0.0692	Yes
MAK2L1	0.813	0.0800	Yes
KIF2C	0.800	0.0907	Yes

Table 8.

Gene Set Enrichment Analysis Detail of Hallmark MTORC1 Signaling

Probe	Rank metric score	Running ES	Core enrichment
SLC2A1	1.593	0.0264	Yes
GAPDH	0.973	0.0425	Yes
RRM2	0.915	0.0576	Yes
PLK1	0.863	0.0715	Yes
PGK1	0.838	0.0852	Yes
TPI1	0.818	0.0986	Yes

Table 9.

Gene Set Enrichment Analysis Detail of Hallmark Hypoxia Pathway

Probe	Rank metric score	Running ES	Core enrichment
SLC2A1	1.593	0.0318	Yes
GAPDH	0.973	0.0512	Yes
PGK1	0.838	0.0672	Yes
TPI1	0.818	0.0834	Yes

Discussion

LUAD is a highly malignant and heterogeneous disease with various prognoses. Patients at the same stage may have quite different outcomes, even at an early stage. Thus, it is urgent to identify novel biomarkers to fine-tune the prediction of patient prognosis and to adjust treatment strategies. In the present study, we integrated mRNA expression with immunohistochemical staining analyses to investigate the significance, prognostic value, and hypothetical mechanism of SLC2A1 in LUAD.

SLC2A1 has been widely studied as the primary glucose transporter, and it has been identified as a probable prognostic factor in several cancer types. For example, Huang and colleagues performed integrated bioinformatics analysis and revealed the potential research value of the SLC family in small cell lung cancer (Liao et al., 2019). Concerning NSCLC, a meta-analysis, including 26 studies, was conducted in 2018, and the results suggested that positive SLC2A1 expression predicted a poor prognosis (Zhang et al., 2018). Nonetheless, there are several problems associated with previous studies, including the ambiguous cutoff value of SLC2A1 expression between different studies and incomplete prognostic information in some studies. In addition, conflicting results have been obtained in different studies when multivariate analysis was used. Another study explored the association of single nucleotide polymorphisms of SLC2A1 and patient prognosis and demonstrated that SLC2A1 variants may be useful in predicting the survival of those with early stage NSCLC (Do et al., 2018). Gao and Wang (2018) also conducted an in silico study and revealed SLC2A1 as a prognosis-related differentially expressed gene in LUAD; however, only two datasets were included, and the survival analysis was presented using the KM curve and did not exclude the influence of other factors. Thus, the prognostic value of SLC2A1 in LUAD needs to be verified.

In our study, the cohort from TCGA and eight independent datasets consistently showed that SLC2A1 was significantly highly expressed in LUAD tissues compared with paired normal tissues, indicating that SLC2A1 plays a role in tumorigenesis and progression. Furthermore, we demonstrated that SLC2A1 has great potential in predicting patient prognosis. High SLC2A1 expression correlated significantly with poor overall survival in multiple datasets, including those from TCGA, the KM plotter, and our own cohort. Additionally, subgroup analysis indicated that SLC2A1 expression can be used to stratify the prognosis of patients with early stage LUAD. It may also be useful when identifying early stage patients with undesirable prognoses and subsequently guiding their therapeutic regimen. SLC2A1 was further shown to be an independent factor in both the dataset from TCGA and our dataset, indicating its potential application as a complementary marker for the TNM staging system. It should be noted that investigating whether SLC2A1 may be used as a diagnostic marker is beyond the scope of this study.

With respect to the molecular mechanism involved, an in vitro study conducted by Wang et al. (2017) showed that knockdown of SLC2A1 in LUAD cells inhibits cellular proliferation, metastasis, and glucose utilization. High SLC2A1 expression is significantly associated with the cell cycle pathway. Another study confirmed these results and demonstrated that suppressing SLC2A1 may lead to diminished tumor growth in vivo (Ooi and Gomperts, 2015). Furthermore, an SLC2A1 inhibitor exhibited antitumor activity, even in chemoresistant cancer cells (Ooi and Gomperts, 2015). Our results are highly similar to those from previous pathway analyses, and we utilized more samples for exploration and identified 14 enriched hallmark pathways, including the G2M checkpoint, hypoxia, and MTORC1 signaling. These specific pathways need to be further verified both in vitro and in vivo.

There are certain limitations to the present study. First, the immunohistochemical cohort from our hospital was obtained from a single center, and the data were collected retrospectively. Thus, bias is inevitable. Second, all of our samples were obtained by surgical resection. Therefore, the molecular biology characteristics may not be extrapolated to the entire LUAD cohort. Prospective studies using multicenter cohorts are needed to verify these results. In summary, this is the first time in which in silico analysis was combined with a large cohort to immunohistochemically analyze and verify the relationship between SLC2A1 and LUAD. Our results strongly support the prognostic value of SLC2A1 in LUAD.

Footnotes

Acknowledgment

The authors thank all staff in the Department of Thoracic Surgery for their support during the study.

Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by the National Key R&D Program of China (2017YFC1311000, 2018YFC1312100), CAMS Initiative for Innovative Medicine (2017-I2M-1-005, 2017-I2M-2-003, 2019-I2M-2-002), Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (2018PT32033,2017PT32017), Innovation team development project of Ministry of Education (IRT_17R10).

Supplementary Material

Supplementary Table S1

References

Ashburner

, Ball

C.A.

, Blake

J.A.

, Botstein

, Butler

, Cherry

J.M.

, et al. (2000). Gene ontology: tool for the unification of biology. Nat Genet, 25, 25.

Barr Kumarakulasinghe

, Zanwijk

N.V.

, and Soo

R.A.

(2015). Molecular targeted therapy in the treatment of advanced stage non-small cell lung cancer (NSCLC). Respirology, 20, 370–378.

Bol

G.M.

, Vesuna

, Xie

, Zeng

, Aziz

, Gandhi

, et al. (2015). Targeting DDX3 with a small molecule inhibitor for lung cancer therapy. EMBO Mol Med, 7, 648–669.

Boroughs

L.K.

, and DeBerardinis

R.J.

(2015). Metabolic pathways promoting cancer cell survival and growth. Nat Cell Biol, 17, 351–359.

Chen

, Zheng

, Baade

P.D.

, Zhang

, Zeng

, Bray

, et al. (2016). Cancer statistics in China, 2015. CA Cancer J Clin, 66, 115–132.

DeBerardinis

R.J.

, Lum

J.J.

, Hatzivassiliou

, and Thompson

C.B.

(2008). The biology of cancer: metabolic reprogramming fuels cell growth and proliferation. Cell Metab, 7, 11–20.

S.K.

, Jeong

J.Y.

, Lee

S.Y.

, Choi

J.E.

, Hong

M.J.

, Kang

H.-G.

, et al. (2018). Glucose transporter 1 gene variants predict the prognosis of patients with early-stage non-small cell lung cancer. Ann Surg Oncol, 25, 3396–3403.

Franceschini

, Szklarczyk

, Frankild

, Kuhn

, Simonovic

, Roth

, et al. (2012). STRING v9. 1: protein-protein interaction networks, with increased coverage and integration. Nucleic Acids Res, 41, D808–D815.

Gao

L.-W.

, and Wang

G.-L.

(2018). Comprehensive bioinformatics analysis identifies several potential diagnostic markers and potential roles of cyclin family members in lung adenocarcinoma. Onco Targets Ther, 11, 7407.

10.

Griffin

J.L.

, and Shockcor

J.P.

(2004). Metabolic profiles of cancer cells. Nat Rev Cancer, 4, 551.

11.

Győrffy

, Surowiak

, Budczies

, and Lánczky

(2013). Online survival analysis software to assess the prognostic value of biomarkers using transcriptomic data in non-small-cell lung cancer. PLoS One, 8, e82241.

12.

Kalemkerian

G.P.

, Narula

, Kennedy

E.B.

, Biermann

W.A.

, Donington

, Leighl

N.B.

, et al. (2018). Molecular testing guideline for the selection of patients with lung cancer for treatment with targeted tyrosine kinase inhibitors: American Society of Clinical Oncology Endorsement of the College of American Pathologists/International Association for the Study of Lung Cancer/Association for Molecular Pathology Clinical Practice Guideline Update. J Clin Oncol, 36, 911–919.

13.

Kanehisa

. (2002). The KEGG database. In Novartis Foundation Symposium. Chichester: John Wiley, pp. 91–100.

14.

Lee

C.K.

, Man

, Lord

, Links

, Gebski

, Mok

, et al. (2017). Checkpoint inhibitors in metastatic EGFR-mutated non–small cell lung cancer—a meta-analysis. J Thorac Oncol, 12, 403–407.

15.

Liao

B.-C.

, Lin

C.-C.

, and Yang

J.C.-H.

(2015). Second and third-generation epidermal growth factor receptor tyrosine kinase inhibitors in advanced nonsmall cell lung cancer. Curr Opin Oncol, 27, 94–101.

16.

Liao

, Yin

, Wang

, Zhong

, Fan

, and Huang

(2019). Identification of candidate genes associated with the pathogenesis of small cell lung cancer via integrated bioinformatics analysis. Oncol Lett, 18, 3723–3733.

17.

Liberti

M.V.

, and Locasale

J.W.

(2016). The Warburg effect: how does it benefit cancer cells?. Trends Biochem Sci, 41, 211–218.

18.

Lindeman

N.I.

, Cagle

P.T.

, Aisner

D.L.

, Arcila

M.E.

, Beasley

M.B.

, Bernicker

E.H.

, et al. (2018). Updated molecular testing guideline for the selection of lung cancer patients for treatment with targeted tyrosine kinase inhibitors: guideline from the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology. J Thorac Oncol, 13, 323–358.

19.

Lopez-Serra

, Marcilla

, Villanueva

, Ramos-Fernandez

, Palau

, Leal

, et al. (2014). A DERL3-associated defect in the degradation of SLC2A1 mediates the Warburg effect. Nat Commun, 5, 3608.

20.

Mao

, Yang

, He

, and Krasna

M.J.

(2016). Epidemiology of lung cancer. Surg Oncol Clin, 25, 439–445.

21.

Mok

T.S.

, Wu

Y.-L.

, Ahn

M.-J.

, Garassino

M.C.

, Kim

H.R.

, Ramalingam

S.S.

, et al. (2017). Osimertinib or platinum–pemetrexed in EGFR T790M-positive lung cancer. N Engl J Med, 376, 629–640.

22.

Ooi

A.T.

, and Gomperts

B.N.

(2015). Molecular pathways: targeting cellular energy metabolism in cancer via inhibition of SLC2A1 and LDHA. Clin Cancer Res, 21, 2440–2444.

23.

Postow

M.A.

, Callahan

M.K.

, and Wolchok

J.D.

(2015). Immune checkpoint blockade in cancer therapy. J Clin Oncol, 33, 1974.

24.

Rhodes

D.R.

, Yu

, Shanker

, Deshpande

, Varambally

, Ghosh

, et al. (2004). ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia, 6, 1–6.

25.

Rizvi

, Sanchez-Vega

, La

, Chatila

, Jonsson

, Halpenny

, et al. (2018). Molecular determinants of response to anti–programmed cell death (PD)-1 and anti–programmed death-ligand 1 (PD-L1) blockade in patients with non–small-cell lung cancer profiled with targeted next-generation sequencing. J Clin Oncol, 36, 633.

26.

Shi

, Zhang

, Qin

, Wang

, and Zhu

(2019). Long non-coding RNA LINC00174 promotes glycolysis and tumor progression by regulating miR-152-3p/SLC2A1 axis in glioma. J Exp Clin Cancer Res, 38, 1–12.

27.

Soria

J.-C.

, Ohe

, Vansteenkiste

, Reungwetwattana

, Chewaskulyong

, Lee

K.H.

, et al. (2018). Osimertinib in untreated EGFR-mutated advanced non–small-cell lung cancer. N Engl J Med, 378, 113–125.

28.

Stark

, Grzelak

, and Hadfield

(2019). RNA sequencing: the teenage years. Nat Rev Genet, 20, 631–656.

29.

Steuer

C.E.

, Khuri

F.R.

, and Ramalingam

S.S.

(2015). The next generation of epidermal growth factor receptor tyrosine kinase inhibitors in the treatment of lung cancer. Cancer, 121, E1–E6.

30.

Subramanian

, Tamayo

, Mootha

V.K.

, Mukherjee

, Ebert

B.L.

, Gillette

M.A.

, et al. (2005). Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A, 102, 15545–15550.

31.

Tang

, Li

, Kang

, Gao

, Li

, and Zhang

(2017). GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res, 45, W98–W102.

32.

Torre

L.A.

, Bray

, Siegel

R.L.

, Ferlay

, Lortet-Tieulent

, and Jemal

(2015). Global cancer statistics, 2012. CA Cancer J Clin, 65, 87–108.

33.

Wang

, Shi

, Ding

, Wang

, Liu

, Yang

, et al. (2017). Metabolic reprogramming induced by inhibition of SLC2A1 suppresses tumor progression in lung adenocarcinoma. Int J Clin Exp Pathol, 10, 10759–10769.

34.

Yan

, Wang

, Chen

, Li

, and Fan

(2015). Deregulated SLC2A1 promotes tumor cell proliferation and metastasis in gastric cancer. Int J Mol Sci, 16, 16144–16157.

35.

Yang

, Wen

, Tian

, Lu

, Wang

, et al. (2017). GLUT-1 overexpression as an unfavorable prognostic biomarker in patients with colorectal cancer. Oncotarget, 8, 11788.

36.

Yuan

, Wu

, Xu

, Xiong

, Chu

, Yu

, et al. (2015). Notch signaling: an emerging therapeutic target for cancer treatment. Cancer Lett, 369, 20–27.

37.

Zhang

, Xie

, and Li

(2018). The clinicopathologic impacts and prognostic significance of GLUT1 expression in patients with lung cancer: a meta-analysis. Gene, 689, 76–83.

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

0.04 MB