Molecular and Immune Characteristics for Lung Adenocarcinoma Patients With ERLIN2 Overexpression

Background Endoplasmic reticulum lipid raft-associated protein 2 (ERLIN2) is protein contained in the membrane of the endoplasmic reticulum. In lung adenocarcinoma (LUAD), the molecular function of ERLIN2 and the correlation between ERLIN2 and tumor-infiltrating immune cells have been unclear. The aim of our study was to determine the role of ERLIN2 in LUAD development to provide a better understanding of the molecular pathogenesis of this disease and identify new therapeutic targets for its treatment. Methods Immunohistochemistry, Western blotting, and real-time quantitative polymerase chain reaction were used to detect protein and mRNA levels of ERLIN2 in LUAD and adjacent normal tissues. Using the A549, H1299 cell line, ERLIN2-short hairpin RNA was applied to silence ERLIN2 to determine its role in LUAD cell proliferation and invasion. Based on mRNA expression of ERLIN2 from the Cancer Genome Atlas (TCGA) database, we identified ERLIN2-related protein-coding genes and analyzed the Kyoto Encyclopedia of Genes and Genomes pathway to explore its potential biological functions and determined the correlation between ERLIN2 and tumor-infiltrating immune cells. Results ERLIN2 was abnormally expressed in a variety of tumor tissues and is highly expressed in LUAD. This overexpression was associated with histological grade (P = 0.044), TNM stage (P = 0.01), and lymph node metastasis (P = 0.038). Patient overall survival was poorer with ERLIN2 overexpression. Downregulation of ERLIN2 inhibited LUAD cell proliferation and invasion in vitro. Based on mRNA expression of ERLIN2 from the TCGA database, 13 ERLIN2-related genes and 10 pathways were identified and showed a correlation between ERLIN2 and naive B cells and neutrophils. Conclusion ERLIN2 could serve as a potential diagnostic and prognostic biomarker for LUAD and has demonstrated to be correlated with immune infiltrates, which suggests that it may represent a new therapeutic target for LUAD.


INTRODUCTION
Lung cancer is one of the main cancer-related deaths worldwide, accounting for~20% of all cancer deaths (1). Lung adenocarcinoma (LUAD) is the most important subtype of lung cancer and usually metastasizes, which leads to a poor prognosis for the patient (2). Although the existing treatment methods have made some progress, the 5-year survival rate remains only 10-20% (3), which is important when identifying the current therapeutic restrictions associated with the disease. The mechanisms of LUAD tumorigenesis remain unclear; therefore, it is urgent that we identify any biomarkers for the diagnosis of this disease.
Endoplasmic reticulum lipid raft-associated protein 2 (ERLIN2), also known as stomatin/prohibitin/flotillin/HflK/C (SPFH2) or C8ORF2, is a protein within the membrane of the endoplasmic reticulum (ER) that contains an evolutionarily conserved SPFH domain (4). In recent years, ERLIN2 has been considered to be a new medium related to ER degradation by binding to activated inositol triphosphate receptors (IP3Rs) and other ER-related degradation substrates, leading to polyubiquitination and their subsequent degradation (5). ERLIN2 can also interact with ER-resident protein from insulin-induced gene 1 to regulate the activation of sterol regulatory element binding protein 1c without acting as an ER degradation medium (6). Through this regulating mechanism, ERLIN2 helps the cells maintain high levels of cytoplasmic lipids and gain a growth advantage during tumorigenic stress. ERLIN2 gene mutation has been found to be related to motor neuron diseases in children (7). In the research, ERLIN2 has been reported only in breast cancer to indicate its effects on the cellcycle processes of breast cancer cells (8).
As far as we know, there has been no study on the role of ERLIN2 in LUAD carcinogenesis; therefore, the aim of our study was to explore the expression of ERLIN2 in LUAD samples and analyze the correlation between ERLIN2 expression and certain clinical parameters, as well as the prognosis for LUAD patients.

Lung Adenocarcinoma (LUAD) Clinical Samples and Immunohistochemistry Assay
At the Affiliated Hospital of Nantong University in China, 284 pairs of LUAD and adjacent normal tissues were treated. All patients were treated by surgical resection between 2007 and 2011. All clinical data on the patients were carefully recorded after the diagnosis of LUAD by two pathologists. The pathological stage was determined according to the 8th Edition of the TNM Classification for Lung Cancer (9). The follow-up was completed by June 30, 2014, and the median follow-up duration was 52 months. All experiments involving patient specimens were approved by the Ethics Committee of the Affiliated Hospital of Nantong University, China.
An immunohistochemistry (IHC) assay was conducted as previously described (10). Briefly, the LUAD samples were deparaffinized and rehydrated. The primary antibodies were those against ERLIN2 (1:100 dilution; ab129207; abcam). The scoring criteria for IHC staining were based on the intensity of the stain and the percentage of immunoreactive cells, as previously described (10).

Analyses of Western Blotting and Real-Time Quantitative Polymerase Chain Reaction
Western blotting analyses were conducted as previously described (11) using 50 mg protein samples from fresh tissues. The primary antibodies were ERLIN2 (1:2,000 dilution; ab129207; abcam) and b-actin (1:10,000 dilution; 66009-1; proteintech). RNA from tumor tissues was extracted using TRIzol Reagent (Invitrogen), and the cDNA was obtained through reverse transcription using a PrimeScript ™ RT Reagent Kit (TaKaRa, Shiga, Japan). Real-time quantitative polymerase chain reaction (qPCR) was conducted in triplicate for each cell sample using a SYBR Premix Ex Taq II Reagent Kit (TaKaRa).

Lung Adenocarcinoma (LUAD) Cell Lines and Cell Culture
The A549, H1299 human LUAD cell line was purchased from the American Type Culture Collection (Manassas, VA, USA). The A549, H1299 cells were cultured in swell Park Memorial Institute-1640 medium (Thermo Fisher Scientific, Inc., Waltham, MA, USA) containing 10% fetal bovine serum (Haoyang Biological Manufacture Co. Ltd., Tianjin, China) and 100 units penicillinstreptomycin in a humidified atmosphere with 5% CO 2 at 37°C.

Construction of Plasmids and Transfection
A549, H1299 cells were transfected with plasmids encoding ERLIN2, or short-hairpin (sh)RNA against ERLIN2, along with vector control. The shRNA targeting sequence for ERLIN2 was GGGTAACAAAGCCCAACATAC. The cDNA encoding fulllength human ERLIN2 was cloned into a PCDH vector. The expression constructs were confirmed by DNA sequencing. The transfection process was as described before (10).
Cell Viability Assays, 5-Ethynyl-2′-Deoxyuridine Staining, Cell Wound Healing Assay, and Transwell Assay Cell viability was measured using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sigma-Aldrich) according to the manufacturer's instructions. Cell growth was also evaluated using the 5-ethynyl-2′-deoxyuridine kit (RiboBio, Science City, China) according to the manufacturer's instructions. Cells were cultured to full confluence in a six-cell plate, after which a micropipette tip was used to scratch the surface. The cells were washed with phosphate-buffered saline and cultured in serumfree medium. The scratches were then photographed at 0, 24, and 48 h, and cell migration was compared by measuring the gap size in each field. A549, H1299 cell invasion assay was using a Transwell system (Corning, Tewksbury, MA) based on previously described methods (12).

Flow Cytometry Analysis
The BD Fluorescence-activated cell sorting (FACS) Calibur flow cytometry system (Becton Dickinson, Franklin Lakes, NJ, USA) was used to detect cell cycle distribution. A549, H1299 cells were harvested and fixed with 70% ice-cold ethanol. After treating with RNaseA, the cells were stained with propidium iodide (PI) for 30 min.

Gene Set Enrichment Analysis
In the Cancer Genome Atlas (TCGA)-Persons at Risk (PARD) database, 535 LUAD cases were divided into two expression-level groups according to the median expression value of ERLIN2. A gene set enrichment analysis was then conducted to detect the gene sets that were enriched in the gene rank in the two groups for identifying a potential hallmark of LUAD. For each analysis, 1,000 repetitions of gene set permutations were completed. The phenotype label put forth was the expression level of ERLIN2. In addition, we used the nominal p-value and normalized enrichment score to sort the enriched pathways in each phenotype (13). Gene sets with a discovery rate (FDR) <0.05 were considered to be significantly enriched.

Immune Infiltrate Analysis
To assess the relative variations of gene expression among sets in the samples, we used the deconvolution algorithm CIBERSORT based on gene expression (http://cibersort.stanford.edu/) (14) to measure the immune response of 21 tumor infiltrating immune cells (TIICs), evaluate their association with ERLIN2 expression in LUAD, and determine any correlation among TIICs. We used standard annotation files to establish gene expression datasets and the default signature matrix at 1,000 permutations. Using the Monte Carlo method, CIBERSORT approximated a p-value for deconvolution to determine the levels of confidence in the results.
To analyze the influence of ERLIN2 on the microenvironment of the immune system, 338 tumor samples were used and classified into two groups. To determine the types of lymphocytes affected by ERLIN2, p-value <0.05 was set as significant.
The Tumor and Immune System Interactions Database (TISIDB) provides a user-friendly web portal (http://cis.hku. hk/TISIDB) that allows users to explore the function of a gene of interest and its role in tumor-immune interactions through high-throughput data analysis and literature mining (15).

Statistical Analyses
The data acquired from TCGA were merged and analyzed using R 3.5.3. The correlations between the clinical information and ERLIN2 expression were analyzed using logistic regression. The multivariate Cox analysis was used to evaluate the influence of ERLIN2 expression on survival. All of the experiments were repeated at least three times. Differences between groups were determined using Student's t-test. P-value <0.05 indicated statistical significance.

The Levels of Endoplasmic Reticulum Lipid Raft-Associated Protein 2 (ERLIN2) mRNA in Lung Adenocarcinoma (LUAD) and Other Cancers
The analysis of TCGA RNA-seq data using the TIMER database (http://cistrome.org/TIMER/) showed that ERLIN2 mRNA expression was significantly higher in esophageal carcinoma, head and neck cancer, lung adenocarcinoma, lung squamous cell carcinoma, and stomach adenocarcinoma and lower in kidney chromophobe, kidney renal clear cell carcinoma, kidney renal papillary carcinoma, thyroid carcinoma, and uterine corpus endometrial carcinoma tissues compared with that in adjacent normal tissues ( Figure 1) (Supplementary Figure 1). FIGURE 1 | The level of ERLIN2 expression in different tumor types from the TCGA database in TIMER. Note: *P < 0.05, **P < 0.01, ***P < 0.001.

Endoplasmic Reticulum Lipid Raft-Associated Protein 2 (ERLIN2) Predicts Poor Prognosis for Lung Adenocarcinoma (LUAD) Patients
To explore the role of ERLIN2 in the development of LUAD, we first analyzed the expression levels of ERLIN2 in 10 pairs of fresh tumor tissues and matched normal tissues using Western blotting and qPCR. Our results showed that ERLIN2 was markedly overexpressed in the LUAD samples compared with that in the matched adjacent normal tissues ( Figure 2). The results of IHC also revealed that ERLIN2 was overexpressed in LUAD compared with that in adjacent normal tissues ( Figure 3). ERLIN2 was also highly expressed in 129 of 284 (45.4%) lungtumor tissues. Moreover, the expression levels of ERLIN2 were correlated with histological grade (P = 0.044), TNM stage (P = 0.01), and lymph node metastasis (P = 0.038; Table 1). Kaplan-Meier analysis revealed that the overall survival (OS) rate of patients with a higher ERLIN2 expression was lower than those with a low ERLIN2 expression (P = 0.015; Figure 4A). In addition, Cox regression analyses demonstrated that ERLIN2 was an independent predictor for LUAD (P = 0.019; Figure 4B). Therefore, these results indicated that ERLIN2 overexpression might predict a poor prognosis in LUAD patients.

Effect of Endoplasmic Reticulum Lipid
Raft-Associated Protein 2 (ERLIN2) on Lung Adenocarcinoma (LUAD) Cell Growth In Vitro High levels of ERLIN2 can predict a poor prognosis in LUAD patients, which prompted us to study whether ERLIN2 might be involved in oncogene function. After silencing ERLIN2, expression of ERLIN2 protein decreased in A549 ( Figure 5A). EdU evaluated how ERLIN2 regulates DNA replication. After silencing ERLIN2, the proliferation of A549 cells was significantly inhibited ( Figure 5B). Cell healing and invasion experiments showed that silencing the expression of ERLIN2 decreased the migration and invasion abilities of the cells (Figures 5C, D). To further study the effect of ERLIN2 on cell growth, flow cytometry was conducted to analyze the cell cycle. Our results indicated that ERLIN2 depletion blocked the cell cycle at the G0/G1 stage in the treated LUAD cells ( Figure 5E). The same results were shown in H1299 ( Figure 6).

Identification of Endoplasmic Reticulum Lipid Raft-Associated Protein 2 (ERLIN2)-Related Protein-Coding Genes in Lung Adenocarcinoma (LUAD)
To explore the potential molecular regulatory mechanisms of ERLIN2 in LUAD tumorigenesis, we found protein-coding genes closely related to the expression level of ERLIN2. A two-sided Pearson correlation coefficient analysis and z-test were performed using R based on gene expression data extracted from TCGA. |Pearson correlations| >0.50 and z-test p <0.001 were used as cutoff criteria. Ultimately, 13 protein-coding genes were identified as ERLIN2-related genes (Table 2, Figure 7A).

Gene-Set Enrichment Analysis of Interrelated Pathways
KEGG pathway analyses were conducted to explore the potential biological functions of ERLIN2. GSEA revealed significant differences (FDR < 0.050, p-value < 0.050) in enrichment of the KEGG pathways in samples with high levels of ERLIN2. As shown in Table 3 and Figures 7B-K, KEGG pathway analysis showed the following 10 pathways that had the strongest correlation with ERLIN2 expression: glycosylphosphatidylinositol (GPI) anchor biosynthesis, hematopoietic cell lineage, intestinal immune network for immunoglobulin A (IgA) production, allograft rejection, graft versus host disease, asthma, cytokine receptor interaction, type I diabetes mellitus, ribosome, and natural killer cell (NKC)-mediated cytotoxicity.

Association Between Endoplasmic Reticulum Lipid Raft-Associated Protein 2 (ERLIN2) Expression and Tumor Infiltrating Immune Cells (TIIC) Composition
As shown in Figure 8A, the immune cells that have a significant correlation with ERLIN2 expression were active B cells, memory B cells, immature B cells, and neutrophils. According to the median expression value of ERLIN2, 338 LUAD tumor tissues were downloaded from the TCGA database and divided into high-and low-expression groups comprising169 high-and 169 low-expression groups that met the screening criteria.
CIBERSORT was used to explore gene expression profiles of the downloaded samples to determine the levels of 21 types of immune cells. The CIBERSORT algorithm applied to the 21 immune cell subtypes helped to assess the differences in their expression levels in the high-and low-expression groups. Naive B cells and neutrophil cells have difference between the high and low expressions of ERLIN2 ( Figure 8B). As shown in Figure 8C, the correlation heatmap reflects a higher correlation within the proportions of different TIIC subgroups. We analyzed the relationship between ERLIN2 expression of TIICs and cell surface markers through the "correlation" module of the Gene Expression Profiling Interactive Analysis tool. This study shows that the immune cells affected by ERLIN2 gene expression were CD8A of CD8+T cells, CD2 and CD3E of T cells (general), CD19 and CD79A of B cells, CCR7 of neutrophils, T-bet, STAT4 of Th1, STAT6 of Th2, BCL-6 of Tfh, STAT3 and IL17A of Th17, CTLA4 of T cell exhaustion, and TPSB2 of mast cells ( Table 4). We used the Spearman correlation coefficient to evaluate the correlation. The results of B cell and neutrophil markers were similar to those obtained from TISIDB and CIBERSORT.

DISCUSSION
Although significant progress has been made in the diagnosis and treatment of LUAD, the 5-year survival of LUAD patients is still poor. Therefore, it is essential to elucidate the molecular mechanisms of LUAD development and identify new prognostic markers and therapeutic targets for LUAD. Previous studies have reported that some proteins including HMGA1, IDH1, CEA and CYFRA play a role in the development of lung adenocarcinoma (16). In the present study, we revealed for the first time that   overexpression of ERLIN2 may help to predict a poor prognosis in LUAD Patients. First, the analysis of TCGA RNA-seq data using the TIMER database showed that ERLIN2 mRNA expression was significantly higher in LUAD and some other cancers. We then compared ERLIN2 expression in LUAD tissues with that in adjacent normal tissues using Western blotting, qPCR, and IHC. Our results showed that ERLIN2 was markedly upregulated in LUAD tissues compared with that in normal tissues. Our results also indicated that the expression levels of ERLIN2 were correlated with histological grade, TNM stage, and lymph node metastasis in LUAD patients. In addition, patients with high ERLIN2 expression levels demonstrated a shorter OS time than those with low ERLIN2 expression levels; therefore, these results suggest that ERLIN2 is an oncogene. ERLIN2 has been reported to be overexpressed in human breast cancer and promotes cancer cell proliferation (8,(17)(18)(19). Recent studies have found that the down-regulation of ERLIN2 gene increases cell apoptosis and inhibits the proliferation, invasion and migration of breast cancer cells, which may be related to PI3K/ AKT signaling pathway (20).To evaluate the carcinogenic effect of ERLIN2 in LUAD, we knocked down ERLIN2 in A549 and H1299 cells. ERLIN2 knockdown markedly inhibited the growth rate of the tumor cells and reduced other malignant tumor cell behaviors, such as clonogenic survival, colonies counting, and capability for DNA replication. These results suggest that ERLIN2 can promote cell proliferation in LUAD. The unlimited proliferation of cancer cells results in a disorder of the cell cycle (21). Compared with that in the control group, ERLIN2 knockdown arrested the G0/G1 phase in LUAD cells, which indicated that it can also promote LUAD.   Accordingly, we identified the ERLIN2-related protein-coding genes to investigate the potential biological process in LUAD. Based on the TCGA database, we found the following 13 ERLIN2related protein-coding genes in LUAD: PROSC, BRF2, LSM1, ASH2L, WHSC1L1, BAG4, TM2D2, PPAPDC1B, EIF4EBP1, DDHD2, ZNF703, FGFR1, and ADAM9. To further investigate the functions of ERLIN2 in LUAD, we conducted GSEA using TCGA data, which showed that GPI anchor biosynthesis, hematopoietic cell lineage, intestinal immune network for IgA production, allograft rejection, graft versus host disease, asthma, cytokine receptor interaction, type I diabetes mellitus, ribosome, and NKC-mediated cytotoxicity in KEGG are differentially enriched in the ERLIN2 high-expression phenotype.
Tumor immunotherapy has developed rapidly in recent years, and there is growing recognition of the role of the immune system in the development of cancer (22,23). Some studies have shown that immune cell infiltration has an influence on survival  in lung cancer (24,25). In our study, CIBERSORT analysis suggested that the expression levels of ERLIN2 had a significant effect on the infiltration levels of naive B cells and neutrophils in the LUAD tumor microenvironment. Similarly, the correlation between B cells, neutrophils, and ERLIN2 expression using TISIDB was the same as that from the CIBERSORT analysis. More and more attention has been paid to the role of tumor-infiltrating B cells in the tumor microenvironment. Most studies have shown that the infiltration of B cells, especially naive B cells, memory B cells, and plasma cells, in non-small cell lung cancer is associated with a good prognosis (26). Meanwhile, neutrophils are potential targets for immunotherapy (27,28). Previous studies have shown that neutrophil autophagy is closely related to neutrophil immune activity, cytokine secretion, and extracellular trap formation (29,30); therefore, the signaling pathway of the cytokine-receptor interaction found in our study may be related to neutrophil infiltration. In addition, we used the correlation module in GEPIA to analyze the gene markers in immune cells and the expression of ERLIN2. The correlation between ERLIN2 and surface marker expression in immune cells was basically the same. These results suggest that ERLIN2 may be associated with the immune microenvironment in LUAD. ERLIN2 is an independent prognostic factor of LUAD, and its high expression suggests a poor prognosis. In addition, ERLIN2 expression significantly correlates with several tumor-infiltrating immune cells, particularly naive B cells and neutrophils. These findings suggest that ERLIN2 can be used as a prognostic biomarker for determining a prognosis and immune infiltration in LUAD patients.

DATA AVAILABILITY STATEMENT
The datasets presented in this article are not readily available.
Requests to access the datasets should be directed to ntdxbiantingting@sina.com.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Ethics Committee of the Affiliated Hospital of Nantong University. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
YL and PX conceived the project and wrote the manuscript. DJ conceived the project and participated in the data analysis. JL and JZ collected the data. TB participated in the discussion and language editing. JS reviewed the manuscript. All authors contributed to the article and approved the submitted version.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2020. 568440/full#supplementary-material SUPPLEMENTARY FIGURE 1 | High or low expression of ERLIN2 in different human cancer tissues compared with normal tissues using the TCGA database.