Development and validation of a metabolic gene signature for predicting the overall survival of patients in lung adenocarcinoma with lymph node metastasis

doi:10.21203/rs.3.rs-3747618/v1

Download PDF

Research Article

Development and validation of a metabolic gene signature for predicting the overall survival of patients in lung adenocarcinoma with lymph node metastasis

https://doi.org/10.21203/rs.3.rs-3747618/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose

Metabolism reprogramming plays an essential role in cancers. The current study explored the prognostic potential of metabolic genes for lung adenocarcinoma (LUAD) with metastasis.

Methods

The RNA-seq data and clinical data for LUAD patients were retrieved from public databases. LASSO analysis was utilized to construct a multigene signature in the TCGA-LUAD cohort. LUAD patients from the GSE72094, GSE30219, GSE31210 dataset were used for external validation. The effect of altered GNPNAT1 expression on cell migration, invasion and EMT were explored in vitro.

Results

Our results showed that a total of 40 differentially expressed metabolic genes (DEMGs) were identified to be involved in the occurrence and lymph node metastasis of LUAD. Univariate Cox regression analysis demonstrated that 10 DEMGs were correlated with overall survival (OS) of LUAD patients. LASSO regression analysis indicated that the prognostic signature, including ALDOA, MTHFD1L, LDHA, GNPNAT1, POLR3G, GPD1L, PGS1, was developed in the TCGA-LUAD cohort. The prognostic value of this signature was successfully validated in the GSE72094, GSE31210, GSE30219 dataset. Receiver operating characteristic (ROC) curve analysis confirmed this signature's predictive capacity. Functional analysis revealed that several signaling pathways were enriched in the high-risk group. Furthermore, the calibration curve of the prognostic nomogram demonstrated good agreement between the predicted and observed survival rates for each of OS. Further analysis revealed that this signature could be an independent predictor for OS in patients with LUAD. Finally, we successfully detected expression level of 7 hub MGs at the transcription level and firstly found that GNPNAT1 might played an important role on LUAD cells migration, invasion, and EMT.

Conclusion

A novel metabolic gene signature can be used for prognostic prediction in LUAD metastasis. Targeting metabolism may be a therapeutic alternative for LUAD metastasis.

Lung adenocarcinoma

metabolic genes signature

prognosis

lymph node metastasis

Lung cancer is the most prevalent and the leading cause of cancer-related morbidity and mortalities worldwide (Bray et al. 2018). Approximately 40–60% patients with lung cancer present with metastasis during diagnosis (Kobrinsky et al. 2003; Satoh et al. 2009). Metastasis accounts for 70–90% of cancer-related mortalities (Wood et al. 2014). Lung adenocarcinoma (LUAD) is the most common pathological type of lung cancer, and accounts for 40% of all lung cancer cases. LUAD is characterized by early metastasis, and most cases are diagnosed in the late stages, which explains the high mortality in LUAD patients (Collins et al. 2007; Riihimaki et al. 2014). Therefore, identification of accurate therapeutic targets and prognostic biomarkers can improve overall survival (OS) of LUAD patients with metastasis.

Metabolic reprogramming, as one of the hallmarks of cancer, is closely related to cancer cell growth, proliferation, angiogenesis, invasion, metastasis, immunosuppression, and treatment resistance (Hanahan and Weinberg 2011; Hensley et al. 2016; Rosario et al. 2018; Crunkhorn 2019; Lane et al. 2020). Therefore, metabolic reprogramming provides a basis for discovery of new tumor biomarkers. Several prognostic metabolic gene signatures have been developed for different cancers, such as hepatocellular carcinoma (Cui et al. 2020; Wang et al. 2021b), clear cell renal carcinoma (Zhao et al. 2020), glioblastoma (He et al. 2020b), endometrial cancer (Jiang et al. 2020), rectal cancer (Zhang et al. 2020b), neuroblastoma (Meng et al. 2020). Previous studies reported that hundreds of metabolic genes were abnormally expressed in LUAD (Asavasupreechar et al. 2019; Vanhove et al. 2019). In addition, several metabolic gene signatures have been developed to predict prognosis of LUAD patients (Cao et al. 2020; He et al. 2020a; Zhang et al. 2020a; Wang et al. 2021a), but there are few studies on metabolic gene signature to predict the prognosis of LUAD patients with metastasis.

In the current study, metabolic genes associated lymph node metastasis were identified in TCGA-LUAD cohort. We developed a 7-MG signature on the basis of the findings from LASSO regression analysis in TCGA-LUAD cohort, which was further confirmed in the GSE72094, GSE31210, and GSE30219 dataset to search for an efficient metabolic biomarker for the more accurate stratification management of LUAD metastasis. In addition, based on the 7 hub MGs, a nomogram was established to predict the 1-, 3- and 5-year OS of LUAD. The calibration plot was applied to assess the accuracy of the nomogram prediction. Finally, the mRNA expression levels of 7 hub MGs were detected in the TCGA-LUAD cohort and GNPNAT1 was firstly found to promote LUAD cells migration, invasion, and EMT.

Data collection and processing

The RNA sequencing (RNA-seq) data and corresponding clinical information of LUAD cohort were downloaded from the cancer genome atlas (TCGA) database (https://portal.gdc.cancer.gov/). GSE72094, GSE31210, and GSE30219 as the validation dataset was downloaded from the Gene Expression Omnibus Database (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). The limma packages were utilized to conduct the normalization process, deleting the normal or repeated samples for subsequent analysis.

Metabolism-associated genes

Genes implicated in metabolism were downloaded from the molecular signatures database (MSigDB, V.7.1) (Liberzon et al. 2015). The key search words in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database were “metabolism” or “metabolic”.

Screening differentially expressed metabolic genes (DEMGs)

Based on the lymph node status, the LUAD tissues were divided into lymph node metastasis (LNM) and non-lymph node metastasis (NLNM) groups. Differentially expressed metabolic genes (DEMGs) between LUAD and adjacent non-tumorous samples (a) as well as differentially expressed metabolic genes (DEMGs) between LNM and NLNM (b) were screened using limma packages (Ritchie et al. 2015) by a criteria of a false discovery rate (FDR < 0.05). The overlapping genes by intersecting the results of a and b were used for further analysis.

Constructing the metabolism-associated gene signature in LUAD patients

In order to avoid the interference from irrelevant factors, patients with follow-up time from 0 to 2000 days in the TCGA-LUAD cohort were utilized to construct the metabolic gene signature. GSE72094, GSE30219 and GSE31210 were explored for external confirmation. Univariate Cox regression analysis was conducted to identify metabolic genes for predicting the overall survival (OS) of LUAD patients by a criteria of P value < 0.01. Least absolute shrinkage and selection operator (LASSO) regression analysis was utilized to identify coefficient of prognostic DEMGs. The metabolic gene signature was presented as risk score= Ʃ (βi x Expi), where βi represented the coefficient of gene i, and Expi represented the expression level of gene i. Survival analysis of the metabolic gene signature in LUAD patients was evaluated using Kaplan-Meier with log-rank test analysis. Predictive accuracy of the metabolic gene signature was evaluated using ROC analysis.

Validation of the metabolism-associated gene signature and construction of a nomogram based on the hub genes of the metabolism-associated gene signature

Predictive value of the metabolic gene signature for OS of LUAD patients was validated using GSE72094, GSE30219, and GSE31210 dataset retrieved from GEO database. Moreover, a predictive nomogram for OS of LUAD patients was also constructed in TCGA-LUAD cohort based on the hub genes of the metabolic gene signature. In addition, the correlation between the metabolic gene signature and clinicopathological characteristics, including age, gender, distant metastasis stage, lymph node metastasis stage, tumor stage, were assessed using wilcoxon test.

Functional enrichment analysis

GSEA was performed to identify pathways regulated by the 7-MG signature. Empirical gene-set associations were based on 1,000 permutations. FDR < 0.05 was considered statistically significant.

Cell culture and cell transfection

Human lung adenocarcinoma cell lines (A549 and H1299) were acquired from the Heilongjiang Cancer Institute (Harbin, China). A549 and H1299 cells were cultured in Roswell Park Memorial Institute 1640 medium, supplemented with 10% fetal bovine serum (Gibco, USA) at 37℃ and 5%CO2. A small interfering RNA for GNPNAT1 (siGNPNAT1 and a negative control (siNC) were used in knockdown function experiments. siGNPNAT1 (RiboBio, China) were transfected into A549 and H1299 cells using jetPRIME (Polyplus, France) reagent. For overexpression of GNPNAT1, pCMV3-GNPNAT1 from Shen Gong Company (Shanghai, China) was transfected into A549 and H1299 cells using jetPRIME reagent. For control experiment, A549 and H1299 cells were transfected with an equal amount of negative control siRNAs or pCMV3-vector. Forty-eight hours after transfection, cells were harvested for western blot analysis. siRNA sequences were as follows:

siGNPNAT1#1:5’-GCAAGAAACTGAACTGTTA-3’;

siGNPNAT1#2:5’-CGGCAACTCTGATTATAGA-3’.

Cell migration and invasion assays

Cell migration and invasion abilities were evaluated using Transwell assays (Corning, USA). 2×10⁴ cells were suspended in the upper chamber, and 600µl medium containing 10% FBS was placed into the lower chamber. After 24 h incubation, A549 and H1299 cells that had migrated through the membrane were fixed with methanol for 15 min and stained with 10% crystal violet. For invasion assays, Matrigel (BD, USA) was added into the upper chambers 12 h before the experiments. The number of cells were quantified in six random fields, and three independent experiments were performed.

Western blotting

Cell lysates for A549 and H1299 were produced by RIPA lysis buffer (Beyotime, China). BCA protein assay kit (Beyotime, China) was selected to detect protein concentration. Equal amount of proteins (30ug) were utilized to electrophoresis through 10% sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE) forand and then were transferred onto polyvinylidene difluoride (PVDF) membranes via electroblotting. The membranes were blocked with 5% fat-free milk in phosphate-buffered saline with 0.1% tween for 1h at room temperature. The membranes were probed at 4℃ overnight with the specific primary antibodies, followed by incubation with horseradish peroxides (HRP)-coupled secondary antibodies for 1h at room temperature. The enhanced chemiluminescene reagent was used to visualize the specific protein bands on the membrane. The antibodies against protein were listed as follow: anti-GNPNAT1(1/1000 dilution, Proteintech Group, Inc, China), anti-Vimentin (1/20000 dilution, Proteintech Group, Inc, China), anti-E-cadherin (1/1000 dilution, Proteintech Group, Inc, China), anti-N-cadherin (1/2000 dilution, Proteintech Group, Inc, China), anti-actin (1/10000 dilution, Proteintech Group, Inc, China).

Statistical analysis

R software and GraphPad prism 9 were utilized to perform statistical analysis. Prognostic DEMGs were explored through univariate Cox regression analysis. Risk score was constructed using LASSO regression analysis. Predictive potential of the metabolic gene signature for OS of LUAD patients was analyzed using Kaplan-Meier with log-rank test and ROC analysis. The general linear model for predicting the OS of LUAD patients was generated using the rms package. The difference comparison between two groups was evaluated using student’s t-test. P < 0.05 was considered statistically significant.

Identification of DEMGs and prognostic DEMGs in the TCGA-LUAD cohort

The flowchart for this study was shown in Fig. 1. In order to obtain DEMGs, we firstly identified 40 metabolic pathways from the molecular signatures database. Subsequently, we matched metabolic genes from 40 metabolic pathways with the RNA-seq data in the TCGA-LUAD cohort. 674 DEMGs between LUAD and adjacent non-tumorous samples of TCGA-LUAD cohort were identified via the criteria of FDR < 0.05(Fig S1A, B). Likewise, 55 DEMGs between LNM and NLNM samples of TCGA-LUAD cohort were identified via the criteria of FDR < 0.05 (Fig S1C, D). As shown in Fig. 2A, a total of 40 overlapping genes in the intersection of 674 DEMGs and 55 DEMGs were identified to be involved in the occurrence and lymph node metastasis of LUAD. Subsequently, univariate Cox regression analysis was performed on 40 overlapping genes. It was shown in Fig. 2B that 10 DEMGs were significantly correlated with OS of LUAD patients. Among them, ALDOA, CDA, MTHFD1L, LDHA, GNPNAT1, RRM2, POLR3G, and TK1 were classified as a risk type, whose HR was > 1 with lymph node metastasis, while GPD1L and PGS1 were classified as a protective type, whose HR was < 1 with non-lymph node metastasis. In order to further explore potential function of 10 prognostic DEMGs in LUAD, the interaction network from the STRING databases was shown in Fig. 2C, the correlation was presented in Fig. 2D.

Construction of a prognostic model for in the TCGA-LUAD cohort

LASSO regression analysis showed that fructose-bishosphate aldolase (ALDOA), methylenetetrahydrofolate dehydrogenase 1-like (MTHFD1L), lactate dehydrogenase A (LDHA), glucosamine-phosphate N-acetyltransferase 1 (GNPNAT1), RNA polymerase III subunit G (POLR3G), glycerol-3-phosphate dehydrogenase 1-like (GPD1L), and phosphatidylglycerophosphate synthase 1 (PGS1) were the most satisfactory 7 genes that were used to construct prognostic MG signature in LUAD. Based on the results of LASSO regression analysis, we used the coefficients to construct the prognostic model as following: Risk score = Exp (ALDOA) × 0.0017 + Exp (MTHFD1L) × 0.0095 + Exp (LDHA) × 0.0038 + Exp (GNPNAT1) × 0.0289 + Exp (POLR3G) × 0.0216 - Exp (GPD1L) × 0.0073 - Exp (PGS1) × 0.0551. The median risk score was used as the cut off value to classify the patients from TCGA-LUAD cohort into either high-risk or low-risk groups. The risk score, survival status and gene expression heatmap of these prognostic MGs were shown in (Fig. 2E). It was shown in Fig. 2F that compared with the low-risk group, the high-risk group had significantly shorter OS. The area under the ROC curve (AUC) of the 7-MG signature for 1-, 3-, 5-year OS was 0.705, 0.691, and 0.733, respectively (Fig. 2G). As shown in Fig. 2H, GSEA showed that the top 8 hallmarks were correlated with the high-risk group: cell cycle, DNA replication, mismatch repair, P53 signaling, pentose phosphate pathway, proteasome, pyrimidine metabolism and assembly of spliceosomes.

Validation of the 7-MG signature for LUAD

In order to confirm our findings in TCGA-LUAD cohort, three independent external datasets (GSE72094, GSE30219, GSE31210) from GEO database were applied to validate the ability of the 7-MG signature for predicting the OS of LUAD patients. According to the median risk score, the three independent external datasets were classified into low-risk and high-risk groups, survival status and gene expression heatmap of these prognostic MGs were shown in Fig. 3A, Fig S2A and Fig S2D. Survival analysis of the three independent external datasets was then executed to confirm the ability of the 7-MG signature for predicting the OS of LUAD patients. The LUAD patients in low-risk group had a better OS in all the three GEO dataset, including GSE31210 (p ༜ 0.001, 1-, 3-, 5-AUC:0.768, 0.683, 0.716), GSE30219 (p ༜ 0.001, 1-, 3-, 5-AUC: 0.608, 0.585, 0.937), GSE72094 (p ༜ 0.001, 1-, 3-, 5-AUC: 0.736, 0.620, 0.631) (Fig. 3B, C, Fig S2B, C, Fig S2E, F). To better predict 1-, 3- and 5-year survival in patients with LUAD, A nomogram was constructed with the seven hub MGs that had most significant values in LASSO regression analysis (Fig. 3D). In addition, the calibration curve of the nomogram indicated good agreement between the predicted and observed survival rates for each of OS (Fig. 3E-G).

Clinical significance of the 7-MG signature

To evaluate whether the 7-MG signature was an independent factor for predicting the OS of LUAD patients, univariate and multivariate Cox regression analysis was performed. The results showed that the 7-MG signature could be used as an independent prognostic factor (Fig. 4A, B). To further verify the clinical value of the 7-MG signature, the correlation between the risk score and clinical characteristics were subsequently evaluated. The result revealed that high risk scores were positively correlated with tumor stage (Fig. 4E, p ༜ 4.052e-04), lymph node stage (Fig. 4F, p ༜ 1.662e-05), and clinical stage (Fig. 4H, p ༜ 6.761e-07) in patients with LUAD, while no association was found between GNPNAT1 expression and age (Fig. 4C), gender (Fig. 4D), distant metastasis (Fig. 4G).

Expression level and prognostic prediction analysis of 7 hub MGs

To explore the seven hub genes at the transcription level, the TCGA-LUAD cohort was utilized to detected their mRNA expression levels. The expression of ALDOA, POLR3G, MTHFD1L, LDHA, and GNPNAT1 in LUAD tissues were all increased than those of adjacent normal lung tissues (Fig. 5A, B, Fig. 5I-K), while the expression of GPD1L and PGS1 was decreased than those of adjacent tissues (Fig. 5C, D). Subsequently, we also investigated the effect of these hub genes on the OS of LUAD patients, the results showed that the high expression of ALDOA, POLR3G, MTHFD1L, LDHA and GNPNAT1 was positively associated with poor OS in patients with LUAD (Fig. 5E, F, Fig. 5L-M), whereas the high expression of GPD1L and PGS1 had an opposite result (Fig. 5G, H).

GNPNAT1 promotes LUAD cell migration, invasion and EMT

For the seven hub MGs (ALDOA, MTHFD1L, GNPNAT1, LDHA, POLR3G, GPD1L, PGS1), we randomly selected GNPNAT1 to perform cell experiment. It was shown in Fig. 6A-D that the efficacy of GNPNAT1 knockdown or overexpressed in A549 and H1299 was confirmed using western blot analysis. Transwell assay validated that GNPNAT1 knockdown suppressed the migration and invasion of A549 and H1299 cells (Fig. 6E, F). In contrast, cells with GNPNAT1 overexpression exhibited more aggressive migratory and invasive potential (Fig. 6G, H). GNPNAT1 knockdown significantly increased the expression of E-cadherin and decreased the expression of vimentin, N-cadherin (Fig. 6I, J), while GNPNAT1 overexpression significantly increased the expression of vimentin, N-cadherin and decreased the expression of E-cadherin (Fig. 6K, L).

Lung adenocarcinoma is a highly heterogeneous disease with a poor prognosis. Several studies have explored gene signatures for prognostic prediction of LAUD patients owing to advances in high-throughput technology and bioinformatics (Li et al. 2020a; Li et al. 2020b; Wang et al. 2020; Xue et al. 2020). Moreover, identifying novel gene signatures for predicting prognosis of patients provides information on choosing effective treatment regimen and improving survival rate of LUAD patients (Wang et al. 2019; Ma et al. 2020).

Metabolic abnormalities play an important role in development and progression of several cancer types. Metabolic genes can be used as prognostic signature for LUAD patients by continuously exploring cancer metabolic heterogeneity. Although many studies on metabolism-associated prognostic signature in LUAD have been conducted in LUAD, metabolism-associated prognostic signatures for LUAD with metastasis were rare. Therefore, in the study, the relationship between dysregulated metabolism and LUAD with lymph node metastasis was explored. The analysis of TCGA-LUAD data identified 40 MGs were identified to be involved in the occurrence and lymph node metastasis of LUAD. Univariate Cox regression analysis demonstrated that 20 out of 40 MGs were correlated with OS of LUAD patients. Notably, 7 most dysregulated MGs were selected to construct a prognostic predictive model for LUAD. The 7-MG signature was successfully validated using GSE31210, GSE30219, GSE72094 dataset. The ROC analysis in TCGA-LUAD cohort, GSE31210, GSE30219, and GSE72094 dataset revealed that the accuracy of the 7-MG signature in predicting the OS of LUAD was satisfactory. The relevance analysis demonstrated that risk score was positively correlated with gender, tumor stage, lymph node metastasis, and pathological stage. Univariate and multivariate Cox regression analysis showed that the 7-MG signature was an independent prognostic predictor for LUAD patients. Moreover, a nomogram based on the 7 hub MGs accurately predicted the 1-, 3-, and 5-year OS of LUAD patients. GSEA analysis showed that the 7-MG signature was implicated in modulation of cell cycle, DNA replication, mismatch repair, P53 signaling, pentose phosphate pathway, proteasome, pyrimidine metabolism, and spliceosome assembly pathways in LUAD patients in the high-risk group. Notably, these signaling pathways were implicated in development and progression of tumor. These findings suggested that the 7-MG signature was should predict prognosis of LUAD patients with good feasibility and reliability. In addition, the 7-MG signature could provide novel insights for therapeutic targets and underlying molecular mechanism of LUAD with lymph node metastasis.

Most of the 7 hub MGs (ALDOA, MTHFD1L, GNPNAT1, LDHA, POLR3G, GPD1L, PGS1) have been reported to be involved in tumor development and progression across malignancies. ALDOA is among the most abundant glycolytic enzymes in all cancer cells(Gizak et al. 2019). It was reported that ALDOA promoted lung cancer metastasis via PHD-mediated stabilization of HIF-1α and the subsequent activation of MMP9(Chang et al. 2017). In addition, ALDOA was found to promote lung cancer metastasis through interaction with γ-actin and induce lung cancer stemness by decreasing miR-145 expression and activating Oct4 transcription(Chang et al. 2020). MTHFD1L is a mitochondrial enzyme involved in the folate cycle by catalyzing the reaction of formyltetrahydrofolate to formate and tetrahydrofolate (THF)(Walkup and Appling 2005; Pike et al. 2010; Tibbetts and Appling 2010). Previous studies reported that MTHFD1L played important role in neural tube defects(Parle-McDermott et al. 2009), coronary artery disease(Samani et al. 2007), depression and Alzheimer's disease(Eszlari et al. 2016). Presently, it was found that MTHFD1L was implicated in liver cancer(Lee et al. 2017), colorectal cancer(Agarwal et al. 2019; He et al. 2020c), bladder cancer(Eich et al. 2019), lung cancer(Do et al. 2020), and esophageal cancer(Yang et al. 2018). POLR3G as a subunit of RNA polymerase (Pol) III was reported to be involved in several cancers. It was reported that POLR3G inhibition led to cell-specific proliferation arrest and cell death in prostate cancer(Petrie et al. 2019). Deletion of POLR3G resulted in reduced tumor growth and suppressed metastasis in triple-negative breast cancer(Lautré et al. 2022). Knockdown of POLR3G resulted in reduced clonal formation of lung cancer cells and increased chemosensitivity of lung cancer cells to paclitaxel(Park et al. 2023). GPD1L, which is located on chromosome 3p22.3(Shrestha et al. 2018), catalyze the conversion of sn-glycerol 3-phosphate to glycerone phosphate(Yuan et al. 2021) and was found to be involved in more than one type of tumor. For example, Liu et al. indicated that GPD1L inhibited tumor growth and promoted mitophagy by activating the PINK1/Parkin pathway in renal cell carcinoma(Liu et al. 2023). Fan et al. discovered that GPD1L suppressed the proliferation of LUAD cells and negatively associated with Th2 cells, NK CD56dim cells, Tgd, Treg, and neutrophils(Fan et al. 2023).

LDHA promoted progression of different malignant tumors such as LUAD (Hou et al. 2019), breast cancer (Wang et al. 2012; Jin et al. 2017), thyroid cancer (Hou et al. 2021), gastric cancer (Zhu et al. 2018) and neuroblastoma (Dorneburg et al. 2018) as well as head and neck squamous cell carcinoma (Wang et al. 2012; Cai et al. 2019). Overexpression of LDHA promoted lymphatic metastasis, distant metastasis and was correlated with poor prognosis of LUAD. Furthermore, LDHA knockdown inhibited proliferation, migration, invasion and metastasis of LUAD cell lines by inhibiting epithelial-to-mesenchymal (EMT) transformation (Hou et al. 2019). Overexpression of LDHA promoted tumorigenesis, invasion and metastasis of breast cancer and was associated with poor prognosis of breast cancer patients (Wang et al. 2012; Jin et al. 2017). Moreover, LDHA overexpression resulted in poor prognosis and promoted metastasis of papillary thyroid cancer by inducing EMT transformation (Hou et al. 2021). A previous study reported that LDHA knockdown inhibited invasion and migration of gastric cancer cells, whereas LDHA silencing modulated growth of neuroblastoma (Dorneburg et al. 2018) and suppressed invasion and metastasis of head and neck squamous cell carcinoma (Wang et al. 2012; Cai et al. 2019). Studies reported that GNPNAT1 was an independent predictor of LUAD prognosis(Zheng et al. 2020; Liu et al. 2021; Zhang et al. 2021; Zhu et al. 2021). our studies further explored the role of GNPNAT1 in promoting migration, invasion and metastasis of LUAD.

The current study had some limitations. First, the prognostic model was developed and validated using retrospective data retrieved from TCGA and GEO databases, which may result in data loss and selection bias. Therefore, further prospective studies should be conducted to validate the clinical application of the model. Second, we only validated GNPNAT1 of 7 hub MGs, the effect of other hub MGs (ALDOA, MTHFD1L, LDHA, POLR3G, GPD1L, PGS1) will need to be further explored in our future studies

The current study constructed and validated a model for predicting the OS of LUAD patients using 7 hub MGs. The 7-MG signature could provide novel insights for therapeutic targets and underlying molecular mechanism of LUAD with lymph node metastasis.

LUAD			Lung adenocarcinoma
GNPNAT1			Glucosamine-phosphate N-acetyltransferase 1
LNM			Lymph node metastasis
NLNM			Non-lymph node metastasis
FBS			Fetal bovine serum
RPMI-1640			Roswell Park Memorial Institute-1640
MGs			Metabolic genes
FDR			False discovery rate
LASSO			Least absolute shrinkage and selection operator
DEMGs			Differentially expressed metabolic genes
OS			Overall survival
TCGA			The cancer genome atlas
GEO			Gene Expression Omnibus Database
ROC			Receiver operating characteristic curve
EMT			Epithelial-mesenchymal transition

Acknowledgements

We are very grateful to TCGA and GEO database for providing their platforms and contributors for uploading their meaningful datasets.

Author contributions

JH, FL: Conception, design, and study supervision. JH, FX, HT, YL, WZ: development of methodology, analysis, and interpretation of data. JH, FX, HT: performing the experiments. JH, YL, WZ: writing of the manuscript. FL, WZ: review of the manuscript. All authors approved the final version of the manuscript.

Funding No funding was received for conducting this study.

Data availability Publicly available datasets were analyzed in the present study. The data can be found here: TCGA database (https://portal.gdc.cancer.gov/) and GEO database (https://www.ncbi.nlm.nih.gov/geo/).

Conflict of interest The authors declare that there is no conflict of interest in this work.

Ethics statement The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The present study was approved by the ethic committee of Harbin Medical University cancer hospital.

Agarwal S, Behring M, Hale K et al (2019) MTHFD1L, A folate cycle enzyme, is involved in progression of colorectal cancer. Transl Oncol 12(11):1461–1467. https://doi.org/10.1016/j.tranon.2019.07.011.
Asavasupreechar T, Chan MSM, Saito R et al (2019) Sex steroid metabolism and actions in non-small cell lung carcinoma. J Steroid Biochem Mol Biol 193:105440. https://doi.org/10.1016/j.jsbmb.2019.105440.
Bray F, Ferlay J, Soerjomataram I et al (2018) Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 68(6):394–424. https://doi.org/10.3322/caac.21492.
Cai H, Li J, Zhang Y et al (2019) LDHA Promotes Oral Squamous Cell Carcinoma Progression Through Facilitating Glycolysis and Epithelial-Mesenchymal Transition. Front Oncol 9:1446. https://doi.org/10.3389/fonc.2019.01446.
Cao Y, Lu X, Li Y et al (2020) Identification of a six-gene metabolic signature predicting overall survival for patients with lung adenocarcinoma. PeerJ 8:e10320. https://doi.org/10.7717/peerj.10320.
Chang Y, Chan Y, Chang W et al (2017) Feedback regulation of ALDOA activates the HIF-1α/MMP9 axis to promote lung cancer progression. Cancer Lett 403:28–36. https://doi.org/10.1016/j.canlet.2017.06.001.
Chang Y, Yang Y, Chiou J et al (2020) Nonenzymatic function of Aldolase A downregulates miR-145 to promote the Oct4/DUSP4/TRAF4 axis and the acquisition of lung cancer stemness. Cell Death Dis 11(3):195. https://doi.org/10.1038/s41419-020-2387-2.
Collins L, Haines C, Perkel R et al (2007) Lung Cancer: Diagnosis and Management. Am Fam Physician 75:56–63.
Crunkhorn S (2019) Targeting cancer cell metabolism in glioblastoma. Nat Rev Cancer 19(5):250. https://doi.org/10.1038/s41568-019-0139-3.
Cui L, Xue H, Wen Z et al (2020) Prognostic roles of metabolic reprogramming-associated genes in patients with hepatocellular carcinoma. Aging (Albany NY) 12(21):22199–22219. https://doi.org/10.18632/aging.104122.
Do S, Choi S, Lee S et al (2020) Genetic variants in one-carbon metabolism pathway predict survival outcomes of early-stage non-small cell lung cancer. Oncology-Basel 98(12):897–904. https://doi.org/10.1159/000509658.
Dorneburg C, Fischer M, Barth TFE et al (2018) LDHA in neuroblastoma is associated with poor outcome and its depletion decreases neuroblastoma growth independent of aerobic glycolysis. Clin Cancer Res 24(22):5772–5783. https://doi.org/10.1158/1078-0432.CCR-17-2578.
Eich M, Rodriguez Pena M, Chandrashekar D et al (2019) Expression and Role of Methylenetetrahydrofolate Dehydrogenase 1 Like (MTHFD1L) in Bladder Cancer. Transl Oncol 12(11):1416–1424. https://doi.org/10.1016/j.tranon.2019.07.012.
Eszlari N, Kovacs D, Petschner P et al (2016) Distinct effects of folate pathway genes MTHFR and MTHFD1L on ruminative response style: a potential risk mechanism for depression. Transl Psychiatry 6(3):e745-e745. https://doi.org/10.1038/tp.2016.19.
Fan Z, Wu S, Sang H et al (2023) Identification of GPD1L as a potential prognosis biomarker and associated with immune infiltrates in lung adenocarcinoma. MEDIAT INFLAMM 2023:9162249. https://doi.org/10.1155/2023/9162249.
Gizak A, Wiśniewski J, Heron P et al (2019) Targeting a moonlighting function of aldolase induces apoptosis in cancer cells. Cell Death Dis. 10(10):712. https://doi.org/10.1038/s41419-019-1968-4.
Hanahan D, and Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell 144(5):646–674. https://doi.org/10.1016/j.cell.2011.02.013.
He L, Chen J, Xu F et al (2020a) Prognostic implication of a metabolism-associated gene signature in lung adenocarcinoma. Mol Ther Oncolytics 19:265–277. https://doi.org/10.1016/j.omto.2020.09.011.
He Z, Wang C, Xue H et al (2020b) Identification of a metabolism-related risk signature associated with clinical prognosis in glioblastoma using integrated bioinformatic analysis. Front Oncol 10:1631. https://doi.org/10.3389/fonc.2020.01631.
He Z, Wang X, Zhang H et al (2020c) High expression of folate cycle enzyme MTHFD1L correlates with poor prognosis and increased proliferation and migration in colorectal cancer. J Cancer 11(14):4213–4221. https://doi.org/10.7150/jca.35014.
Hensley CT, Faubert B, Yuan Q et al (2016) Metabolic Heterogeneity in Human Lung Tumors. Cell 164(4):681–694. https://doi.org/10.1016/j.cell.2015.12.034.
Hou X, Shi X, Zhang W et al (2021) LDHA induces EMT gene transcription and regulates autophagy to promote the metastasis and tumorigenesis of papillary thyroid carcinoma. Cell Death Dis 12(4):347. https://doi.org/10.1038/s41419-021-03641-8.
Hou XM, Yuan SQ, Zhao D et al (2019) LDH-A promotes malignant behavior via activation of epithelial-to-mesenchymal transition in lung adenocarcinoma. Biosci Rep 39(1). https://doi.org/10.1042/BSR20181476.
Jiang P, Sun W, Shen N et al (2020) Identification of a metabolism-related gene expression prognostic model in endometrial carcinoma patients. BMC Cancer 20(1):864. https://doi.org/10.1186/s12885-020-07345-8.
Jin L, Chun J, Pan C et al (2017) Phosphorylation-mediated activation of LDHA promotes cancer cell invasion and tumour metastasis. Oncogene 36(27):3797–3806. https://doi.org/10.1038/onc.2017.6.
Kobrinsky NL, Klug MG, Hokanson PJ et al (2003) Impact of smoking on cancer stage at diagnosis. J Clin Oncol 21(5):907–913. https://doi.org/10.1200/JCO.2003.05.110.
Lane AN, Higashi RM and Fan TW (2020) Metabolic reprogramming in tumors: Contributions of the tumor microenvironment. Genes Dis 7(2):185–198. https://doi.org/10.1016/j.gendis.2019.10.007.
Lautré W, Richard E, Feugeas J et al (2022) The POLR3G subunit of human rna polymerase iii regulates tumorigenesis and metastasis in triple-negative breast cancer. Cancers (Basel). 14(23):5372. https://doi.org/10.3390/cancers14235732.
Lee D, Xu IM-J, Chiu DK-C et al (2017) Folate cycle enzyme MTHFD1L confers metabolic advantages in hepatocellular carcinoma. J Clin Invest 127(5):1856–1872. https://doi.org/10.1172/JCI90253.
Li W, Gao L, Song P et al (2020a) Development and validation of a RNA binding protein-associated prognostic model for lung adenocarcinoma. Aging (Albany NY) 12:3558–3573. https://doi.org/10.18632/aging.102828.
Li W, Li N, Gao L et al (2020b) Integrated analysis of the roles and prognostic value of RNA binding proteins in lung adenocarcinoma. PeerJ 8:e8509. https://doi.org/10.7717/peerj.8509.
Liberzon A, Birger C, Thorvaldsdottir H et al (2015) The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 1(6):417–425. https://doi.org/10.1016/j.cels.2015.12.004.
Liu T, Zhu H, Ge M et al (2023) GPD1L inhibits renal cell carcinoma progression by regulating PINK1/Parkin-mediated mitophagy. J Cell Mol Med 27(16):2328–2339. https://doi.org/10.1111/jcmm.17813.
Liu W, Jiang K, Wang J et al (2021) Upregulation of GNPNAT1 predicts poor prognosis and correlates with immune infiltration in lung adenocarcinoma. Front Mol Biosci 8:605754. https://doi.org/10.3389/fmolb.2021.605754.
Ma X, Ren H, Peng R et al (2020) Identification of key genes associated with progression and prognosis for lung squamous cell carcinoma. PeerJ 8:e9086. https://doi.org/10.7717/peerj.9086.
Meng X, Feng C, Fang E et al (2020) Combined analysis of RNA-sequence and microarray data reveals effective metabolism-based prognostic signature for neuroblastoma. J Cell Mol Med 24(18):10367–10381. https://doi.org/10.1111/jcmm.15650.
Park C, Lee M, Lee S et al (2023) Regulating POLR3G by MicroRNA-26a-5p as a promising therapeutic target of lung cancer stemness and chemosensitivity. Noncoding RNA Res 8(3):273–281. https://doi.org/10.1016/j.ncrna.2023.03.001.
Parle-McDermott A, Pangilinan F, O'Brien KK et al (2009) A common variant in MTHFD1L is associated with neural tube defects and mRNA splicing efficiency. Hum Mutat 30(12):1650–1656. https://doi.org/https://doi.org/10.1002/humu.21109.
Petrie JL, Swan C, Ingram RM et al (2019) Effects on prostate cancer cells of targeting RNA polymerase III. Nucleic Acids Res 47(8):3937–3956. https://doi.org/10.1093/nar/gkz128.
Pike S, Rajendra R, Artzt K et al (2010) Mitochondrial C1-tetrahydrofolate synthase (MTHFD1L) supports the flow of mitochondrial one-carbon units into the methyl cycle in embryos. J BIOL CHEM 285(7):4612–4620. https://doi.org/10.1074/jbc.M109.079855.
Riihimaki M, Hemminki A, Fallah M et al (2014) Metastatic sites and survival in lung cancer. Lung Cancer 86(1):78–84. https://doi.org/10.1016/j.lungcan.2014.07.020.
Ritchie ME, Phipson B, Wu D et al (2015) limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43(7):e47. https://doi.org/10.1093/nar/gkv007.
Rosario SR, Long MD, Affronti HC et al (2018) Pan-cancer analysis of transcriptional metabolic dysregulation using The Cancer Genome Atlas. Nat Commun 9(1):5330. https://doi.org/10.1038/s41467-018-07232-8.
Samani NJ, Erdmann J, Hall AS et al (2007) Genomewide association analysis of coronary artery disease. New Engl J Med 357(5):443–453. https://doi.org/10.1056/NEJMoa072366.
Satoh H, Kurishima K, Nakamura R et al (2009) Lung cancer in patients aged 80 years and over. Lung Cancer 65(1):112–118. https://doi.org/10.1016/j.lungcan.2008.10.020.
Shrestha K, Onasanya AE, Eisenberg I et al (2018) miR-210 and GPD1L regulate EDN2 in primary and immortalized human granulosa-lutein cells. Reproduction 155(2):197–205. https://doi.org/10.1530/rep-17-0574.
Tibbetts A, and Appling D (2010) Compartmentalization of mammalian folate-mediated one-carbon metabolism. Annu Rev Nutr 30:57–81. https://doi.org/10.1146/annurev.nutr.012809.104810.
Vanhove K, Graulus GJ, Mesotten L et al (2019) The metabolic landscape of lung cancer: new insights in a disturbed glucose metabolism. Front Oncol 9:1215. https://doi.org/10.3389/fonc.2019.01215.
Walkup A, and Appling D (2005) Enzymatic characterization of human mitochondrial C1-tetrahydrofolate synthase. Arch Biochem Biophys 442(2):196–205. https://doi.org/10.1016/j.abb.2005.08.007.
Wang X, Yao S, Xiao Z et al (2020) Development and validation of a survival model for lung adenocarcinoma based on autophagy-associated genes. J Transl Med 18(1):149. https://doi.org/10.1186/s12967-020-02321-z.
Wang Z, Embaye KS, Yang Q et al (2021a) Establishment and validation of a prognostic signature for lung adenocarcinoma based on metabolism-related genes. Cancer Cell Int 21(1):219. https://doi.org/10.1186/s12935-021-01915-x.
Wang Z, Embaye KS, Yang Q et al (2021b) A novel metabolism-related signature as a candidate prognostic biomarker for hepatocellular carcinoma. j hepatocell carcinoma 8:119–132. https://doi.org/10.2147/JHC.S294108.
Wang Z, Wang Z, Niu X et al (2019) Identification of seven-gene signature for prediction of lung squamous cell carcinoma. Onco Targets Ther 12:5979–5988. https://doi.org/10.2147/OTT.S198998.
Wang ZY, Loo TY, Shen JG et al (2012) LDH-A silencing suppresses breast cancer tumorigenicity through induction of oxidative stress mediated mitochondrial pathway apoptosis. Breast Cancer Res Treat 131(3):791–800. https://doi.org/10.1007/s10549-011-1466-6.
Wood SL, Pernemalm M, Crosbie PA et al (2014) The role of the tumor-microenvironment in lung cancer-metastasis and its relationship to potential therapeutic targets. Cancer Treat Rev 40(4):558–566. https://doi.org/10.1016/j.ctrv.2013.10.001.
Xue L, Bi G, Zhan C et al (2020) Development and validation of a 12-gene immune relevant prognostic signature for lung adenocarcinoma through machine learning strategies. Front Oncol 10:835. https://doi.org/10.3389/fonc.2020.00835.
Yang YS, Yuan Y, Hu WP et al (2018) The role of mitochondrial folate enzyme MTHFD1L in esophageal squamous cell carcinoma. Scand J Gastroentero 53(5):533–540. https://doi.org/10.1080/00365521.2017.1407440.
Yuan M, Guo Y, Xia H et al (2021) Novel SCN5A and GPD1L variants identified in two unrelated han-chinese patients with clinically suspected brugada syndrome. Front Cardiovasc Med 8:758903. https://doi.org/10.3389/fcvm.2021.758903.
Zhang J, Zhang J, Yuan C et al (2020a) Establishment of the prognostic index reflecting tumor immune microenvironment of lung adenocarcinoma based on metabolism-related genes. J Cancer 11(24):7101–7115. https://doi.org/10.7150/jca.49266.
Zhang S, Zhang H, Li H et al (2021) Potential role of glucosamine-phosphate N-acetyltransferase 1 in the development of lung adenocarcinoma. Aging (Albany NY) 13(5):7430–7453. https://doi.org/10.18632/aging.202604.
Zhang ZY, Yao QZ, Liu HY et al (2020b) Metabolic reprogramming-associated genes predict overall survival for rectal cancer. J Cell Mol Med 24(10):5842–5849. https://doi.org/10.1111/jcmm.15254.
Zhao Y, Tao Z and Chen X (2020) A three-metabolic-genes risk score model predicts overall survival in clear cell renal cell carcinoma patients. Front Oncol 10:570281. https://doi.org/10.3389/fonc.2020.570281.
Zheng X, Li Y, Ma C et al (2020) Independent Prognostic Potential of GNPNAT1 in Lung Adenocarcinoma. Biomed Res Int 2020:8851437. https://doi.org/10.1155/2020/8851437.
Zhu P, Gu S, Huang H et al (2021) Upregulation of glucosamine-phosphate N-acetyltransferase 1 is a promising diagnostic and predictive indicator for poor survival in patients with lung adenocarcinoma. Oncol Lett 21(6):488. https://doi.org/10.3892/ol.2021.12750.
Zhu W, Ma L, Qian J et al (2018) The molecular mechanism and clinical significance of LDHA in HER2-mediated progression of gastric cancer. Am J Transl Res 10(7):2055–2067.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Development and validation of a metabolic gene signature for predicting the overall survival of patients in lung adenocarcinoma with lymph node metastasis

Status:

Version 1

Abstract

Purpose

Methods

Results

Conclusion

Figures

Introduction

Materials and Methods

Data collection and processing

Metabolism-associated genes

Screening differentially expressed metabolic genes (DEMGs)

Constructing the metabolism-associated gene signature in LUAD patients

Functional enrichment analysis

Cell culture and cell transfection

Cell migration and invasion assays

Western blotting

Statistical analysis

Results

Identification of DEMGs and prognostic DEMGs in the TCGA-LUAD cohort

Construction of a prognostic model for in the TCGA-LUAD cohort

Validation of the 7-MG signature for LUAD

Clinical significance of the 7-MG signature

Expression level and prognostic prediction analysis of 7 hub MGs

GNPNAT1 promotes LUAD cell migration, invasion and EMT

Discussion

Conclusion

Abbreviations

Declarations

Acknowledgements

Author contributions

References

Additional Declarations

Supplementary Files

Status:

Version 1