The protein-protein interaction network of intestinal gastric cancer patients reveals hub proteins with potential prognostic value

Abstract

BACKGROUND:

Gastric cancer (GC) is the third leading cause of cancer worldwide. According to the Lauren classification, gastric adenocarcinoma is divided into two subtypes: diffuse and intestinal. The development of intestinal gastric cancer (IGC) can take years and involves multiple factors.

OBJECTIVE:

To investigate the protein profile of tumor samples from patients with IGC in comparison with adjacent nontumor tissue samples.

METHODS:

We used label-free nano-LC-MS/MS to identify proteins from the tissues samples. The results were analyzed using MetaCore™ software to access functional enrichment information. Protein-protein interactions (PPI) were predicted using STRING analysis. Hub proteins were determined using the Cytoscape plugin, CytoHubba. Survival analysis was performed using KM plotter. We identified 429 differentially expressed proteins whose pathways and processes were related to protein folding, apoptosis, and immune response.

RESULTS:

The PPI network of these proteins showed enrichment modules related to the regulation of cell death, immune system, neutrophil degranulation, metabolism of RNA and chromatin DNA binding. From the PPI network, we identified 20 differentially expressed hub proteins, and assessed the prognostic value of the expression of genes that encode them. Among them, the expression of four hub genes was significantly associated with the overall survival of IGC patients.

CONCLUSIONS:

This study reveals important findings that affect IGC development based on specific biological alterations in IGC patients. Bioinformatics analysis showed that the pathogenesis of IGC patients is complex and involves different interconnected biological processes. These findings may be useful in research on new targets to develop novel therapies to improve the overall survival of patients with IGC.

Keywords

Protein-protein interaction intestinal gastric cancer proteomics hub proteins

1. Introduction

Gastric cancer (GC) is the third leading cause of cancer worldwide, with 1,033,701 new cases and 782,685 related deaths in 2018 [1]. According to the Lauren classification [2], gastric adenocarcinoma is divided into two subtypes: diffuse and intestinal. Intestinal gastric cancer (IGC) is the most frequent subtype in high incidence populations, although its incidence has been declining. IGC is a multifactorial disease that originates from the evolution of sequential lesions. Its development can take years and involves lifestyle, aging, genetics, socioeconomic factors, and infectious agents such as Helicobacter pylori (HP) [3, 4, 5]. HP infection can elicit continuous inflammation, immune response and abnormal epithelial cell proliferation, inducing a metaplastic process that can lead to atypia and occasional neoplasia characterizing the development of IGC [6, 7]. Despite the advances in diagnosis, GC is often detected in late stages due to the nonspecific symptoms. The five-year survival rate is only 20%, and chemotherapy and surgery have limited value in the treatment of advanced cases. Adding to this problem, there is a lack of targeted therapies that use molecular markers [3]. To understand the carcinogenesis of GC, molecular high-throughput studies, including genomics, transcriptomics and proteomics, have been performed. In cancer, these studies mainly address the quantitative and qualitative differences in the protein alterations in tumor and nontumor tissues since these alterations are involved in the carcinogenic process [8]. The use of these technologies for studying GC enables the identification of proteins in different biological specimens, such as blood, gastric fluids, tumor tissues and tumor cells [9]. However, the use of proteomic approaches to study IGC has not been extensively explored compared to their application to other solid tumors. Studies focusing on signaling pathways and interaction networks in IGC are indispensable since the regulatory control of protein-protein interactions (PPIs) plays a critical role in biological pathways in cancer, including apoptosis, cell proliferation, trafficking and angiogenesis. Thus, alterations in proteins and PPIs are related to malignancies and constitute critical targets for treating cancer [10]. A comprehensive analysis of proteomic alterations and their interactions using bioinformatics approaches would greatly benefit our understanding of the molecular background of IGC. Thus, with the aim of further understanding the proteomic changes in the occurrence and development of IGC, we used high-throughput proteomic analysis to investigate the protein profiles of tumor samples from IGC patients and compared them to adjacent nontumor tissue samples from the same patient using label-free nano LC-MS/MS. We identified 429 differentially expressed proteins whose pathways and processes were related mainly to protein folding, apoptosis, and immune response. The PPI network of these altered proteins showed enrichment modules related to the regulation of cell death, the immune system, neutrophil degranulation, the metabolism of RNA and chromatin DNA binding. We also identified 20 key hub proteins, and we assessed the prognostic value of the expression of the genes that encode them. Among them, the expression levels of four hub genes were significantly associated with the overall survival of IGC patients. Protein alterations and their relations to the specific signaling pathways found in this study will help us to better understand the tumor biology of IGC patients, highlighting new potential targets for further investigation.

2. Material and methods

2.1 Patients and study design

The study was conducted in accordance with the Declaration of Helsinki, and all participants signed an informed consent form approved by the National and Institutional Ethics Committee. Tumor samples (from biopsy) were obtained from patients ( $n=$ 18) diagnosed with GC at Instituto Nacional do Câncer José de Alencar Gomes da Silva – INCA (Rio de Janeiro, Brazil). The samples were collected at diagnosis, prior to surgery or chemotherapy. The inclusion criteria were patients older than 18 years old with a confirmed diagnosis of IGC who had not been treated with chemotherapy. The included patients were investigated according to the design shown in Fig. 1. The clinicopathological features of the patients are presented in Supplementary S1.

Figure 1.

Study design. The study schema is illustrated with the methods used and patient populations. Tumor tissues and nontumor adjacent tissues controls from IGC patients were submitted to protein extraction followed by label free proteomic approach. Chromatography coupled to mass spectrometry was used for protein identification and quantification. After statistical analysis and data mining, a list containing a profile of upregulated and downregulated proteins from IGC patients were generated. The list of proteins was submitted to functional enrichment analyses using important bioinformatics tools to identify significant biological pathways and processes. Subsequently, the interaction between the proteins was determined and the top interacting proteins were selected. The genes of the final selected proteins were then evaluated regarding its prognostic value in a cohort of 201 IGC patients.

2.2 The H. pylori status of the patients

During endoscopic evaluation, biopsies from patients were obtained. From the same stomach area, half of the biopsies were prepared for histopathologic evaluation, while the other half were frozen and sent to the National Tumor Biobank at INCA, where they were stored at $-$ 80 ${}^{\circ}$ C.

In the histopathologic examination, H. pylori status was investigated initially with Wayson stain and analyzed by two independent experienced pathologists. At that time, the first status of H. pylori was defined. To further validate these results, immunohistochemistry was performed using the rabbit monoclonal anti-Helicobacter pylori (SP48) antibody (Roche Ltd, Basel, Switzerland).

2.3 Protein extraction, concentration and digestion

Total protein was extracted from tumor and adjacent nontumor tissues, and the samples were concentrated as previously reported [11] and quantified using the Braford method [12]. Two hundred micrograms of protein was used for tryptic digestion, as previously reported [11]. After digestion, 5 mM ammonium formate with 5% acetonitrile was added to the sample for a final volume of 160 $\mu$ L. Then, 40 $\mu$ L of yeast alcohol dehydrogenase was added to a final concentration of 0.2 pmol/ $\mu$ L, which was the concentration used for standardized quantification.

2.4 Liquid chromatography and mass spectrometry

The samples were subjected to nanoscale chromatographic separation (2DnanoLC) using the nanoACQUITY UPLC system from Waters ${}^{\text{\textregistered}}$ according to the method of Pizzatti and colleagues (2012) [13] with modifications. Mass spectrometry analysis was performed as described previously [14]. The MS profile was adjusted to ensure that the LC-MS data generated at low energy were effectively acquired at a 400 to 2000 m/z interval, ensuring that all m/z values lower than 400 in the LC/MSE were the result of dissociation in the collision cell.

2.5 Identification, quantification and in silico analysis

Database searching and protein quantification were performed as previously reported [13]. Briefly, ProteinLynx Global Server v.3.0 (PLGS) with ExpressionE software was used to process the spectra and databank search results. The UniProtKB databank (release 2017_07) was used with manually reviewed annotations. Only the proteins found in all replicates under each condition were subjected to analysis with the PLGS Expression ${}^{\text{E}}$ tool algorithm. The list containing the identified differentially expressed proteins was submitted to functional enrichment and interaction analysis using MetaCore™ (GeneGo, Thomson Reuters). Canonical pathways were obtained by KEGG Mapper. Protein-protein interaction (PPI) analyses were predicted using Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) version 11.0 [15] against the human database. Associations were visualized with a high confidence cutoff (0.900). IGC hub proteins were determined using the Cytoscape plugin “cytoHubba” by using the Maximal Clique Centrality (MCC) method [16]. We also used the open access tool cBioPortal for Cancer Genomics (https://www.cbioportal.org/; version: 3.4.16) [17], which provides visualization, analysis, and downloads regarding cancer genomics alterations of many types of tumors, to analyze genomic alterations in the hub genes in IGC.

2.6 Survival analysis

The survival analysis of genes was performed using KM plotter (https://kmplot.com/analysis/), a web server capable of assessing the effect of 54 k genes (mRNA, miRNA, and protein) on survival in 21 cancer types, including gastric cancer, from different databases (the GEO, EGA, and TCGA databases) ( $n=$ 1,440) [18]. Survival analysis was performed using the Kaplan-Meier method, and hazard ratios with 95% confidence intervals and log-rank $P$ values were calculated. Cohorts of GC patients were split by the trichotomization of gene expression values, where the lower quartile and the upper quartile were selected as the cutoff values. The following parameters were selected: survival $=$ overall survival; follow-up threshold $=$ 120 months; probe set options $=$ all probe sets per gene; Lauren classification $=$ intestinal; treatment $=$ surgery alone; and array quality control $=$ excluded biased array. For the selected parameters, the analysis was performed in 201 patients.

3. Results

3.1 Proteomic profile evaluation of the IGC tumor samples

To identify proteomic alterations, we used label-free protein quantification by mass spectrometry to compare the proteomic profiles of tumor and adjacent nontumor tissue samples (from biopsy) from Brazilian patients ( $n=$ 18) with IGC. First, the proteomic profile of the tumor samples was compared with that of the respective nontumor adjacent tissue samples. For this comparison, the patient samples were allocated into pools: one pool for the tumor tissue (T) samples and another pool for the adjacent nontumor tissue (NT) samples. The samples in the pools were analyzed using label-free nano-liquid chromatography-mass spectrometry (nano-LC-MS). The PLGS Expression ${}^{\text{E}}$ tool algorithm was used to organize the identified proteins into a statistically significant list corresponding to the increased and decreased levels of proteins in the IGC T samples compared to those in the NT samples. After additional applied filters, a final list of the total number of differentially expressed proteins in tumor and nontumor tissues was generated. The differentially expressed proteins were separated into unique (exclusive to T or NT) and common protein categories. A total of 429 proteins were found to be differentially expressed ( $p=$ 0.05, fold change $\geqslant$ 2) between T and NT samples. In total, 268 were found to be downregulated proteins, and 161 were upregulated proteins in tumor samples. The complete list of differentially expressed proteins identified and further information is shown in Supplementary S2.

Table 1
Functional enrichment pathways related to the 429 differentially expressed proteins in IGC patients

Functional enrichment analysis ${}^{\rm a}$	FDR	Protein ${}^{\rm b}$
		Up	Down
Pathways
Possible regulation of HSF-1/chaperone pathway in Huntington’s disease	2.503E-03	HSP70, E2I, HSP90, HSP90 alpha, HSP90 beta	p23 co-chaperone
CFTR folding and maturation (normal and CF)	2.954E-03	HSP70, HSP90 alpha, HSP90 beta	ERp29, Sti1, p23 co-chaperone
Development_Gastrin in differentiation of the gastric mucosa	2.443E-02	ERK1/2	TFF1, Chromogranin A, ICRF, TFF2, Gastrin 17
HSP70 and HSP40-dependent folding in Huntington’s disease	2.443E-02	HSP70, HSP90, HSP90 alpha, HSP90 beta	Sti1
Inhibition of apoptosis in gastric cancer	2.443E-02	Bid, tBid	Gastrin 17, Gastrin 17-Gly, Progastrin, c-FLIP (Short)
NF-kB-, AP-1- and MAPKs-mediated proinflammatory cytokine production by eosinophils in asthma	2.443E-02	ERK1/2, ERK2 (MAPK1), Myeloblastin, Cathepsin G, Leukocyte elastase	Tryptase
Apoptosis and survival_TNFR1 signaling pathway	2.443E-02	Bid, jBid, tBid	MKK7 (MAP2K7), c-FLIP (Short), RAIDD
Apoptosis and survival_TNF-alpha-induced Caspase-8 signaling	2.443E-02	HSP90, HSP90 alpha, Bid, tBid	c-FLIP (Short), Cathepsin B
Apoptosis and survival_Granzyme A signaling	3.228E-02	Histone H3, Histone H1, Ku70	Granzyme A, PHAP1 (pp32)
HIF-1 in gastric cancer	5.060E-02	ERK1/2, HSP90, LDHA	CDC42, p21, LAMR1

${}^{\rm a}$ Enrichment analysis was performed using Metacore TM. ${}^{\rm b}$ Proteins which were identified to be significantly up- or down-regulated by proteomic experiment were shown.

Table 2

Functional enrichment process networks related to the 429 differentially expressed proteins in IGC patients

Functional enrichment analysis ${}^{\rm a}$	FDR	Protein ${}^{\rm b}$
		Up	Down
Process networks
Cytoskeleton_Regulation of cytoskeleton rearrangement	4.528E-03	ERK1/2, ERK2 (MAPK1), Profilin I, Profilin, CAPZA1, CAPZA, Alpha-actinin, Alpha-actinin 2	ARF3,14-3-3 gamma, 14-3-3, CDC42, 14-3-3 eta,Vimentin, Cofilin, ARF1
Inflammation_IL-6 signaling	4.528E-03	ERK1/2, ERK2 (MAPK1), Ceruloplasmin, Fibrinogen beta	Alpha 1-antitrypsin, 14-3-3 gamma, 14-3-3,14-3-3 sigma, 14-3-3 eta, p21, PAP,TATI
Protein folding_Folding in normal condition	4.528E-03	HSP70, PPID, FKBP12, HSP90, HSP90 alpha, HSP90 beta, HSP10 (mitochondrial)	HSPA7, HSPA6, Sti1, p23 co-chaperone, PDI
Protein folding_Response to unfolded proteins	4.538E-03	HSP70, Glutaredoxin, HSP90, HSP90 alpha, HSP90 beta, HSP10 (mitochondrial)	HSPA7, HSPA6, PDI
Apoptosis_Apoptotic mitochondria	8.309E-03	HSP70, ERK2 (MAPK1), Glutaredoxin 1, Glutaredoxin, Bid, tBid	14-3-3 gamma, 14-3-3, 14-3-3 eta
Transcription_mRNA processing	1.026E-02	PSF, NSAP1, SNRPD1 (SMD1), hnRNP H1, PTBP1	Mago nashi, ALY, LSM2, BAT1 (UAP56), DDX39, PABP2, FLJ10292, hnRNP L
Proteolysis_Proteolysis in cell cycle and apoptosis	1.252E-02	Myeloblastin, Bid, tBid	Granzyme A, UBE2C, Cathepsin C, E2N (UBC13), c-FLIP (Long), c-FLIP, c-FLIP (Short), Cathepsin B
Translation_Regulation of initiation	1.252E-02	ERK1/2, ERK2 (MAPK1)	eIF4A, eIF4A1, CDC42, PP1-cat, PP1-cat gamma, eIF2S1, Gastrin 17, PP1-cat alpha,, MLCP (cat)
Translation_Translation initiation	4.043E-02	eIF6 (ITGB4BP), eIF1, RPS4	RPS5, eIF4A, eIF4A1, RPL30, RPS3, LAMR1, eIF4H, eIF2S1, RPS12
Inflammation_Complement system	6.272E-02	Factor B, Factor Bb, C4a, C4b, C4, FHR-3	Clusterin

${}^{\rm a}$ Enrichment analysis was performed using Metacore TM. ${}^{\rm b}$ Proteins which were identified to be significantly up- or down-regulated by proteomic experiment were shown.

Figure 2.

PPI analysis of 429 differentially expressed proteins exclusively found in IGC patients. Network nodes represent interacting proteins. Nodes also indicate the three dimensional protein structures that are known or predicted. Red nodes represent the proteins related to the regulation of cell death. Green nodes represent the proteins related to the immune system. Blue nodes represent the proteins related to neutrophil degranulation. Pink nodes represent the proteins related to the metabolism of RNA. Yellow nodes represent proteins related to chromatin DNA binding. Different line colors indicate the types of evidence supporting these proposed associations. Proteins not connected with the network were removed for better visualization.

Table 3

Top 20 hub proteins in the PPI network, ranked by the maximal clique centrality method

Rank	Gene name	Protein names	Score	Expression
1	LYZ	Lysozyme C	2.09E $+$ 13	Down
2	ELANE	Neutrophil elastase	2.09E $+$ 13	Up
3	CTSC	Dipeptidyl peptidase 1 (Cathepsin C)	2.09E $+$ 13	Down
4	MAPK1	Mitogen-activated protein kinase 1	2.09E $+$ 13	Up
5	TXNDC5	Thioredoxin domain-containing protein 5	2.09E $+$ 13	Down
6	CTSG	Cathepsin G	2.09E $+$ 13	Up
6	PRTN3	Myeloblastin (Leukocyte proteinase 3)	2.09E $+$ 13	Up
8	PA2G4	Proliferation-associated protein 2G4	2.09E $+$ 13	Down
9	NAPRT	Nicotinate phosphoribosyltransferase	2.09E $+$ 13	Up
10	FABP5	Fatty acid-binding protein 5	2.09E $+$ 13	Down
11	HEXB	Beta-hexosaminidase subunit beta (Cervical cancer proto-oncogene 7 protein)	2.09E $+$ 13	Down
12	SERPINB3	Serpin B3 (Protein T4-A) (Squamous cell carcinoma antigen 1)	2.09E $+$ 13	Up
12	NPC2	NPC intracellular cholesterol transporter 2	2.09E $+$ 13	Down
12	PYCARD	Apoptosis-associated speck-like protein containing a CARD (hASC)	2.09E $+$ 13	Up
12	VAT1	Synaptic vesicle membrane protein VAT-1 homolog	2.09E $+$ 13	Up
12	IST1	IST1 homolog (hIST1) (Putative MAPK-activating protein PM28)	2.09E $+$ 13	Up
12	HEBP2	Heme-binding protein 2 (Placental protein 23)	2.09E $+$ 13	Down
18	EIF4A3	Eukaryotic initiation factor 4 A-III	4.76E $+$ 07	Down
19	ALYREF	THO complex subunit 4 (Tho4)	4.72E $+$ 07	Down
20	PTBP1	Polypyrimidine tract-binding protein 1 (hnRNP I)	4.72E $+$ 07	Up

3.2 Alterations related to apoptosis, protein metabolism and immune response are common characteristics in IGC

Based on proteomic results, we used MetaCore™ software (GeneGo Inc., Encinitas, USA) to investigate the pathways and biological processes related to the 429 differentially expressed proteins. The results showed that enriched pathways and processes were related mainly to protein folding, apoptosis, and immune response (Tables 1 and 2). The top 10 signaling pathways and process network results showed that a group of molecular chaperones (upregulated HSP70, HSP90 alpha, HSP90, HSP90 beta, and HSP10), an important family of proteins commonly expressed in response to stress conditions and also known as heat shock proteins (HSPs), was overrepresented in these pathways (Supplementary S3). HSPs were found in chaperone-specific pathways and in networks related to gastric cancer and apoptosis (Table 1). We also found proteins related to apoptosis pathways in these patients. A protein structurally similar to caspase-8, CASP8 and FADD-like apoptosis regulator (c-FLIP, which was downregulated) that functions as a crucial link between cellular survival and death was altered, and BH3-interacting domain death agonist (BID, which was upregulated), a member of the BCL-2 protein family, was also altered (Supplementary S4). We also identified the downregulation of important proteins related to the differentiation of gastric mucosa and its protection, such as trefoil factor 1 and 2 (TFF1 and TFF2), chromogranin A, regenerating family member 1 alpha (ICRF, also known as REG1A) and the mitogen factor gastrin 17 (Supplementary S5). Moreover, we observed that histones, important components in gene regulation, had deregulated expression (histones from H1 and H3 families were upregulated, and the H2 histone was downregulated) and an N-lysine methyltransferase was upregulated. Furthermore, enrichment analysis of the top 10 process networks showed downregulated regulatory proteins, such as the 14-3-3 family proteins (14-3-3 sigma, 14-3-3 gamma, 14-3-3 theta, and 14-3-3 eta) and mitogen-activated protein kinase 1 (MAPK1), in networks related to inflammation, cytoskeleton regulation and apoptosis.

3.3 Protein-protein interaction (PPI) in IGC patients

To investigate the interactions among the 429 differentially expressed proteins in IGC patients, we built a PPI network. A complex protein-protein association was observed in which most proteins had direct or indirect interactions through the number of observed nodes (Fig. 2). Corroborating our previous results, the following biological processes were overrepresented: the regulation of cell death (62 proteins), the immune system (63 proteins), and neutrophil degranulation (45 proteins). Furthermore, the metabolism of RNA (30 proteins) and chromatin DNA binding (7 proteins). Some proteins in the PPI network had a high number of interactions, such as MAPK1, which showed 27 interactions and was connected to proteins involved in the regulation of cell death, the immune system and neutrophil degranulation processes. Elastase neutrophil expressed (ELANE) is a serine protease that hydrolyzes proteins within specialized neutrophil lysosomes and has 22 interactions. Proliferation-associated 2G4 (PA2G4), a protein related to growth regulation and differentiation in cancer, showed 20 interactions. Moreover, proteins from a family of heterogeneous nuclear ribonucleoproteins, polypyrimidine tract-binding protein (PTBP1), displayed 18 interactions and heterogeneous nuclear ribonucleoprotein K (HNRNPK), with 18 interactions, are all involved in processing mRNAs and act as trans-factors in regulating gene expression. DExD-box helicase 39B (DDX39B), with 16 interactions, is involved in various cellular processes, such as mRNA export, splicing and translation.

Figure 3.

Analysis of hub differentially expressed genes in IGC patients. (a) PPI network of hub protein nodes representing highly interacting proteins. Nodes also indicate the three dimensional protein structures that are known or predicted. Blue nodes represent the proteins related to neutrophil degranulation. (b) Enrichment analyses of hub proteins. Top ten significantly ranked enriched biological processes, reactome pathways, molecular functions and cellular components of hub proteins. (c) The oncoPrint plot showing the different genomic alterations in a set of 109 IGC patient samples. (d) An overview summarizing the alterations in the 20 hub genes in the genomics datasets of IGC in the TCGA database.

Figure 4.

The cumulative survival curves of the hub genes in IGC patients plotted using KM plotter. (a) IST1 (also known as KIAA0174). (b) VAT1. (c) ALYREF. (d) HEXB. $P<$ 0.05 was considered statistically significant.

3.4 Determination of hub proteins in IGC patients

Hub proteins are characterized by a large number of interactions and are frequently essential in multiple pathways; therefore, they are important for disease development. Thus, to identify highly connected protein nodes, we determined the hub proteins from the PPI network based on the STRING interaction data. We used the plug-in “cytoHubba” to rank the nodes by their network features through topological analysis methods. The hub proteins were chosen based on the following standards: the top 20 nodes from the network identified by the maximal clique centrality (MCC) analysis method, which has better performance regarding the precision of predicting essential proteins [16]. The results indicated 20 proteins that were identified as hub proteins in the PPI network. Ten of these hub proteins were upregulated, and 10 were downregulated (Table 3). Furthermore, the results of the enrichment analysis of hub proteins in the STRING interaction data showed a significant representation of neutrophil degranulation and immune system biological processes and pathways, as well as intracellular organelles for cellular components (Fig. 3a and b). Figure 3c and d shows the genomic alterations of the 20 hub genes according to cBioportal (data from the TCGA, Firehose Legacy). The 20 hub genes were altered in 43 (39%) of the 109 sequenced patients, mutations 8.26% (9 patients), deep deletions 9.17% (10 patients) and amplification 22.02% (24 patients).

3.5 Survival analysis of the hub genes in KM plotter

Due to the small size of our cohort and the lack of follow-up of the patients, it was not possible to evaluate the prognostic value of hub proteins. To address this question, we decided to analyze the prognostic value of the genes that encode the 20 hub proteins in public software (KM plotter) with data sourced from 1,440 gastric cancer patients [18, 19]. After filtration of these gastric cancer samples, data from 201 IGC patients were used in survival analysis. Of the 20 hub genes, four were significantly associated with overall survival (OS). High expression of IST1 (also known as KIAA0174) (log-rank $P=$ 0.0045, HR $=$ 2.31 (1.28–4.19)) and VAT1 (log-rank $P=$ 2.1e-05, HR $=$ 4.08 (2.03–8.21)) was associated with a short OS, and a high expression of ALYREF (log-rank $P=$ 0.016, HR $=$ 0.46 (0.24–0.88)) and HEXB (log-rank $P=$ 0.012, HR $=$ 0.47 (0.25–0.86)) had a protective effect (Fig. 4).

4. Discussion

The purpose of this study was to identify global proteomic alterations in a Brazilian cohort of patients with ICG. We used label-free nano-LC-MS/MS and in silico approaches to assess the relevant biological alterations in tumors from IGC patients, focusing on alterations in biological pathways and PPI networks that are not available in the current literature.

Overall, when we compared IGC tissue samples with their respective adjacent nontumor samples, we observed that biological pathways and processes inherent to cellular and systemic homeostasis, such as protein folding, cytoskeleton rearrangement, apoptosis and the immune response, were commonly altered. To that extent, we observed that some proteins, especially those related to protein folding, such as HSP, were commonly upregulated and related to chaperone activity, apoptosis and the immune system.

HSPs have a cytoprotective function that is important to cells under normal conditions and constitute the first defense responses to stress conditions [20, 21]. The main activity of HSPs is their chaperone function, through which they are important to the proper folding of proteins, maintaining their structures and functions in cells even under adverse conditions [22, 23]. In cancer cells with high metabolic requirements, it is common to observe the deregulation of these proteins, since the majority of relevant oncoproteins require high levels of HSP chaperones to ensure proper protein folding, stabilization, aggregation, activation, operation and degradation. Among the HSPs, upregulated members from the HSP90 family were found; these proteins are the most studied HSPs, with altered expression in several solid tumors [24]. In gastric cancer, high expression of HSP90 is associated with the staging, proliferation and aggressiveness of the tumor. We also observed dysregulation of HSP70 family proteins (upregulation of HSPA1L and downregulation of HSPA6), which are related not only to specific antitumor immunity but also to regulatory T cell responses [25]. HSPA1L is known to increase cell survival and protection against apoptotic and necrotic stimuli [26]. In addition, it is related to the progression of several solid tumors and to poor prognosis for patients [27, 28]. High expression of HSP70 was observed in GC cell lines [29]. HSP70 overexpression also correlated with poor prognosis in patients with esophageal squamous cell carcinoma and potentially with the progression of other tumors, such as cervical cancer [30], breast cancer [31] and nasopharyngeal carcinoma [32]. The 10 kDa mitochondrial heat shock protein HSPE1, an HSP10 family protein, was also upregulated. This protein is related to immunomodulation, differentiation and proliferation, and it can inhibit apoptosis [33, 34]. High levels of this protein were detected in several tumors [34, 35, 36, 37, 38]. From all these observations, it is possible that pathways related to apoptosis and survival are highly impacted in patients with ICG through the dysregulation of HSPs. Among the identified proteins, the upregulation of BID was found. BID is a pro-apoptotic protein member of the BCL-2 family and is an important regulator that links components of the extrinsic and mitochondrial (intrinsic) pathways of apoptosis [39].

We noted that differentiation of the gastric mucosa pathway was enriched through alterations in important proteins related to its normal function, such as the downregulation of CHGA, TFF1, TFF2, and ICRF (also known as REG1A). In a previous work by our group [40], we verified the downregulation of mRNA levels of CHGA, TFF2 and ICRF (also known as REG1A) in a common molecular signature for intestinal gastric cancer worldwide, corroborating these findings. The genes of the TFF family play important roles in the digestive system of mammals, are mainly related to epithelial cell reconstruction, protection of gastric mucosa, signal transduction and regulation of apoptosis and proliferation, and may act in tumor suppression or promotion. TFF1 and TFF2 are downregulated in gastric cancer tissue compared with normal tissue, and TFF2 downregulation may be associated with hypermethylation [41, 42].

Proteins need to interact with each other to fulfill their functions, and this characteristic is fundamental in most biological processes. Thus, to better comprehend the biological function is essential to understand protein interactions. Highly connected proteins (hubs) in PPIs are often essential to network architecture and functioning [43]. Since high-throughput proteomic analysis has acquired prominence in cancer research, drug targeting investigations using hub proteins and their interactions could be especially interesting [44, 45, 46].

PPI analyses showed the enrichment of proteins with direct and indirect interactions in processes related to apoptosis and neutrophil degranulation, and according to STRING enrichment analyses, the hub proteins are highly involved in neutrophil degranulation. Mature neutrophils represent approximately 50–70% of all leucocytes in adult peripheral blood and are important components of the tumor microenvironment [47]. Neutrophil functionality depends on its cytoplasmic granule composition, which includes important proteins that can promote carcinogenesis [48]. One of these proteins, neutrophil elastase (ELANE), which was also identified as a hub protein, was upregulated. The protumorigenic role of ELANE has been observed in lung, breast, and colon cancers [49, 50, 51]. Notably, in the PPI analysis results, ELANE interacted with other important hub proteins, including the upregulated MAPK1. MAPK1 was the top interacting protein in the PPI analysis, and fourth in the hub analysis; therefore, MAPK1 may be a master regulator related to both apoptosis and neutrophil degranulation processes in IGC patients.

It is interesting to note that in our cohort, all patients were positive for Helicobacter pylori (HP) infection. In addition, an important protein related to tight junctions, occludin, was downregulated in these patients. Long-term infection with HP leads to the disruption of gastric epithelial junctional structures through the downregulation of proteins related to membrane integrity, such as E-cadherin and occludin, and contributes to tissue alterations related to carcinogenic processes [52, 53]. PPI analyses also revealed clusters of interacting proteins related to chromatin DNA binding, where histone variants were dysregulated. The accessibility of DNA for transcription is regulated by complex posttranslational modifications of those histones, such as acetylation, methylation, phosphorylation, ubiquitination, and ADP-ribosylation [54, 55]. MCF-7 cells from breast cancer treated with estrogenic compounds showed an increase in the proliferative rates accompanied by changes in the levels of histones [56]. Additionally, increased histone variant expression is a strong prognostic biomarker in cervical cancer patients [57]. Based on this information, it is reasonable to affirm that alterations in histone variant expression can regulate gene expression and are related to the malignancy of tumors.

In CG, HP infection was observed to induce histone modifications, such as the dephosphorylation of histone H3, thereby regulating gene expression, the cell cycle and cell proliferation [58, 59], and acetylation of H4, which is associated with p21 gene expression in gastric epithelial cells [60]. However, there is little information about the impact of HP infection on the levels of histone expression or the consequences of any such impact in CG patients. We also observed a cluster of proteins related to mRNA metabolic processes, especially proteins from the heterogeneous nuclear ribonucleoprotein (hnRNP) family, such as upregulated SYNCRIP and downregulated HNRNPL. These proteins are RNA-binding proteins that control posttranscriptional regulation, mediating alternative splicing and polyadenylation, thereby directly influencing protein abundance and diversity. SYNCRIP overexpression is related to cell proliferation, growth and survival and maintains the undifferentiated status of leukemic cells [61]. However, there is no information on this protein in GC patients. In contrast to our results, HNRNPL was found to be upregulated in GC patients from a Chinese population; however, the HP status of these patients was not available [62]. These results indicate that HP infection may have an impact on gene expression through histone variants and hnRNP protein alterations.

Furthermore, among the 20 hub proteins identified, IST1 (also known as KIAA0174), VAT1, ALYREF, and HEXB could predict the overall survival. IST1 regulates the disassembly of endosomal sorting complexes in yeast and is related to poor prognosis in GC [63]. Vesicle amine transport 1 (VAT1) is upregulated in glioblastomas and promote migration [64]; however, there is no information about this protein in GC. ALYREF is a nuclear protein that acts as a chaperone, thereby regulating DNA binding, dimerization, transcriptional activity, and mRNA export [65, 66], and is significantly related to the survival outcome in hepatocellular carcinoma [67]. Beta-hexosaminidase subunit beta (HEXB) is a lysosomal enzyme involved in the hydrolysis of ganglioside GM2 to GM3, and its protein expression is related to poor survival in malignant melanoma [68].

This study reveals important findings to comprehend IGC development; however, it has some limitations. The work had initially a small set of the patients with IGC ( $n=$ 18) and mRNA data was used to evaluate our proteomics results. Once mRNA, prior to its translation, commonly undergoes significant post-transcriptional modifications, mRNA levels are not perfect predictors of the function of the protein. Also, in this study the tumor tissue of HP non infected patients were not used as a comparative, since, at the time of the biopsy collection and lab experiments investigation, the majority of samples were HP positive, which is a common infection in Brazilian patients. There weren’t enough HP non infected patient samples that could be included in the study as a good comparative. Notwithstanding, those limitations do not compromise the results since the study design was chose aiming a broader approach investigation to find global alterations in IGC.

In conclusion, based on the differences in the expression of the proteins observed in the proteomic profile of IGC, we were able to identify important proteins and biological processes by integrating comprehensive proteomic approaches and bioinformatics analysis, which showed that the pathogenesis of IGC patients is complex and involves different interconnected biological processes. The present study revealed a considerable number of differentially expressed proteins, which are mainly involved in protein metabolism, cytoskeleton rearrangement, apoptosis, immune response RNA metabolism and chromatin DNA binding. We also identified 20 hub proteins that may have important roles in the molecular background of IGC development. Moreover, the overall survival prediction of IST1 (also known as KIAA0174), VAT1, ALYREF, and HEXB may be useful in researching new targets to develop novel therapies and individualized treatments to improve the overall survival of patients with IGC.

Authorship contributions

Conception: E.C.S., R.B.

Interpretation or analysis of data: E.C.S., R.B., P.V.F., P.V.F.

Preparation of the manuscript: E.C.S., R.B.

Revision for important intellectual content: E.A.

Supervision: E.A.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-203225.

sj-pdf-1-cbm-10.3233_CBM-203225.pdf - Supplemental material

Supplemental material, sj-pdf-1-cbm-10.3233_CBM-203225.pdf

sj-docx-1-cbm-10.3233_CBM-203225.docx - Supplemental material

Supplemental material, sj-docx-1-cbm-10.3233_CBM-203225.docx

sj-xls-1-cbm-10.3233_CBM-203225.xls - Supplemental material

Supplemental material, sj-xls-1-cbm-10.3233_CBM-203225.xls

Footnotes

Acknowledgments

This work was supported by grants from the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), Ministério da Saúde (MS), and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ). We would like to thank the National Tumor Biobank (BNC) at INCA for providing the tissue samples.

Conflict of interest

The authors declared that they have no competing interests.

References

Ferlay

Colombet

Soerjomataram

Mathers

Parkin

D.M.

Piñeros

Znaor

and Bray

, Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods, Int J Cancer 144(8) (2019), 1941–1953.

Lauren

, The two histological main types of gastric carcinoma: Diffuse and so called intestinal-type carcinoma: An attempt at a histo-clinical classification, Acta PatholMicrobiol 64 (1965), 31–49.

Nagini

, Carcinoma of the stomach: A review of epidemiology, pathogenesis, molecular genetics and chemoprevention, World J Gastrointest Oncol 4(7) (2012), 156–169.

Piazuelo

M.B.

and Correa

, Gastric cáncer: Overview, Colomb Med 44(3) (2013), 192–201.

Malfertheiner

et al., Management of Helicobacter pylori infection – the Maastricht IV/florence consensus report, Gut 61(5) (2012), 646–664.

Lamb

and Chen

L.F.

, Role of the Helicobacter pylori-induced inflammatory response in the development of gastric cancer, J Cell Biochem 114(3) (2013), 491–497.

Correa

et al., Gastric cancer in Colombia. III. Natural history of precursor lesions, J Natl Cancer Inst 57(5) (1976), 1027–1035.

and Heng

, Protein array-based approaches for biomarker discovery in cancer, Genomics, Proteomics & Bioinformatics 15(2) (2017), 73–81.

Lin

L.L.

Huang

H.C.

and Juan

H.F.

, Discovery of biomarkers for gastric cancer: A proteomics approach, J Proteomics 75 (2012), 3081–3097.

10.

Dar

K.B.

Bhat

A.H.

Amin

Anjum

Reshi

B.A.

Zargar

M.A.

Masood

and Ganie

S.A.

, Exploring proteomic drug targets, therapeutic strategies and protein – protein interactions in cancer: Mechanistic view, Curr Cancer Drug Targets 19(6) (2019), 430–448.

11.

Panis

Pizzatti

Herrera

A.C.

Correa

Binato

and Abdelhay

, Label-free proteomic analysis of breast cancer molecular subtypes, Journal of Proteome Research 13(11) (2014), 4752–4772.

12.

Bradford

M.M.

, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal Biochem 72 (1976), 248–254.

13.

Pizzatti

Panis

Lemos

Rocha

Cecchini

Souza

G.H.

and Abdelhay

, Label-free MSE proteomic analysis of chronic myeloid leukemia bone marrow plasma: Disclosing new insights from therapy resistance, Proteomics 12(17) (2012), 2618–2631.

14.

Correa

Panis

Binato

Herrera

A.C.

Pizzatti

and Abdelhay

, Identifying potential markers in breast cancer subtypes using plasma label-free proteomics, J Proteome 151 (2017), 33–42.

15.

Szklarczyk

Franceschini

Wyder

Forslund

Heller

Huerta-Cepas

Simonovic

Roth

Santos

Tsafou

K.P.

et al., STRING v10: Protein-protein interaction networks, integrated over the tree of life, Nucleic Acids Res 43 (2015), 447–452.

16.

Chin

C.H.

Chen

S.H.

H.H.

C.W.

M.T.

and Lin

C.Y.

, cytoHubba: Identifying hub objects and sub-networks from complex interactome, BMC Syst Biol 8 (2014), Suppl 4.

17.

Gao

et al., Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal, Sci Signal 6(269) (2013), pl1.

18.

Nagy

Á.

Lánczky

Menyhárt

and Győrffy

, Validation of miRNA prognostic power in hepatocellular carcinoma using expression data of independent datasets, Sci Rep 8(1) (2018), 11515.

19.

Szász

A.M.

et al., Cross-validation of survival associated biomarkers in gastric cancer using transcriptomic data of 1, 065 patients, Oncotarget 7(31) (2016), 49322–49333.

20.

Lindquist

and Craig

E.A.

, The heat-shock proteins, Annu Rev Genet 22 (1988), 631–677.

21.

Beere

H.M.

, The stress of dying: The role of heat shock proteins in the regulation of apoptosis, J Cell Sci 117 (2004), 2641–2651.

22.

Macario

A.J.

and Conway De Macario

, Molecular chaperones: Multiple functions, pathologies, and potential applications, Front Biosci 12 (2007), 2588–2600.

23.

Liu

Daniels

C.K.

and Cao

, Comprehensive review on the HSC70 functions, interactions with related molecules and involvement in clinical diseases and therapeutic potential, Pharmacol Ther 136(3) (2012), 354–374.

24.

Chatterjee

and Burns

T.F.

, Targeting heat shock proteins in cancer: A promising therapeutic approach, Int J Mol Sci 18(9) (2017), 1978.

25.

Binder

R.J.

, Functions of heat shock proteins in pathways of the innate and adaptive immune system, J Immunol 193(12) (2014), 5765–5771.

26.

Nylandsted

Brand

and Jäättelä

, Heat shock protein 70 is required for the survival of cancer cells, Ann N Y Acad Sci 926 (2000), 122–125.

27.

Murphy

M.E.

, The HSP70 family and cancer, Carcinogenesis 34(6) (2013), 1181–1188.

28.

Liu

Rios

Mei

Lin

and Cao

, Heat shock proteins and cancer, Trends Pharmacol Sci 38 (2017), 226–256.

29.

Wang

X.P.

Liao

Liu

G.Z.

Wang

X.C.

and Shang

H.W.

, Co-expression of heat shock protein 70 and glucose-regulated protein 94 in human gastric carcinoma cell line BGC-823, World J Gastroenterol 11(23) (2005), 3601–3604.

30.

Garg

Kanojia

Saini

Suri

Gupta

Surolia

and Suri

, Germ cell-specific heat shock protein 70-2 is expressed in cervical carcinoma and is involved in the growth, migration, and invasion of cervical cells, Cancer 11 (2010), 3785–3796.

31.

Mestiri

Bouaouina

Ahmed

S.B.

Khedhaier

Jrad

B.B.

Remadi

and Chouchane

, Genetic variation in the tumor necrosis factor-alpha promoter region and in the stress protein hsp70-2: Susceptibility and prognostic implications in breast carcinoma, Cancer 91(4) (2001), 672–678.

32.

Jalbout

Bouaouina

Gargouri

Corbex

Ben Ahmed

and Chouchane

, Polymorphism of the stress protein HSP70-2 gene is associated with the susceptibility to the nasopharyngeal carcinoma, Cancer Lett 193(1) (2003), 75–81.

33.

Jia

Halilou

A.I.

Cai

Liu

and Huang

, Heat shock protein 10 (Hsp10) in immune-related diseases: One coin, two sides, Int J Biochem Mol Biol 2(1) (2011), 47–57.

34.

Fan

S.S.

Feng

Xiao

Fan

and Luo

, Elevated expression of HSP10 protein inhibits apoptosis and associates with poor prognosis of astrocytoma, PLoS One 12(10) (2017), e0185563.

35.

Ghobrial

I.M.

Mccormick

D.J.

Kaufmann

S.H.

Leontovich

A.A.

Loegering

D.A.

Dai

N.T.

Krajnik

K.L.

Stenson

M.J.

Melhem

M.F.

Novak

A.J.

Ansell

S.M.

and Witzig

TE.

, Proteomic analysis of mantle-cell lymphoma by protein microarray, Blood 105(9) (2005), 3722–3730.

36.

B. Têtu Popa

Bairati

L’esperance

Bachvarova

Plante

Harel

and Bachvarov

, Immunohistochemical analysis of possible chemoresistance markers identified by micro-arrays on serous ovarian carcinomas, Mod Pathol 21(8) (2008), 1002–1010.

37.

Huang

Liu

Wang

Lin

Chen

Zeng

Chen

Tao

Wei

and Wu

, Comparative mitochondrial proteomic analysis of hepatocelular carcinoma from patients, Proteomics Clin Appl 7(5–6) (2013), 403–415.

38.

Rappa

Pitruzzella

Marino Gammazza

Barone

Mocciaro

Tomasello

Carini

Farina

Zummo

Conway De Macario

Macario

A.J.

and Cappello

, Quantitative patterns of Hsps in tubular adenoma compared with normal and tumor tissues reveal the value of Hsp10 and Hsp60 in early diagnosis of large bowel cancer, Cell Stress Chaperones 21(5) (2016), 927–933.

39.

Yin

X.M.

, Bid, a BH3-only multi-functional molecule, is at the cross road of life and death, Gene 369 (2006), 7–19.

40.

Binato

Santos

E.C.

Boroni

Demachki

Assumpção

and Abdelhay

, A common molecular signature of intestinal-type gastric carcinoma indicates processes related to gastric carcinogenesis, Oncotarget 9(7) (2017), 7359–7371.

41.

Mao

Chen

Peng

T.L.

Yin

X.F.

Chen

L.Z.

and Chen

M.H.

, Role of trefoil factor 1 in gastric cancer and relationship between trefoil factor 1 and gastrokine 1, Oncol Rep 28(4) (2012), 1257–1262.

42.

Katoh

, Trefoil factors and human gastric cancer (review), Int J Mol Med 12(1) (2003), 3–9.

43.

Aksam

V.K.M.

Chandrasekaran

V.M.

and Pandurangan

, Cancer drug target identification and node-level analysis of the network of MAPK pathways, Netw Model Anal Health Inform Bioinforma 7 (2018), 86–92.

44.

Apic

Ignjatovic

Boyer

and Russell

R.B.

, Illuminating drug discovery with biological pathways, FEBS Lett 579 (2005), 1872–1877.

45.

Farkas

I.J.

et al., Network-based tools for the identification of novel drug targets, Sci Signal 4(173) (2011), pt3.

46.

Yang

et al., Identification of hub genes and therapeutic drugs in esophageal squamous cell carcinoma based on integrated bioinformatics strategy, Cancer Cell Int 19(142) (2019).

47.

Galdiero

M.R.

et al., Roles of neutrophils in cancer growth and progression, J Leukoc Biol 103 (2018), 457–464.

48.

Mollinedo

, Neutrophil Degranulation, Plasticity, and Cancer Metastasis, Trends Immunol 40(3) (2019), 228–242.

49.

Houghton

A.M.

et al., Neutrophil elastase-mediated degradation of IRS-1 accelerates lung tumor growth, Nat Med 16 (2010), 219–223.

50.

A.S.

et al., Neutrophil elastase as a diagnostic marker and therapeutic target in colorectal cancers, Oncotarget 5 (2014), 473–480.

51.

Caruso

J.A.

Akli

Pageon

Hunt

K.K.

and Keyomarsi

, The serine protease inhibitor elafin maintains normal growth control by opposing the mitogenic effects of neutrophil elastase, Oncogene 34 (2015), 3556–3567.

52.

Caron

T.J.

Scott

K.E.

Fox

J.G.

and Hagen

S.J.

, Tight junction disruption: Helicobacter pylori and dysregulation of the gastric mucosal barrier, World J Gastroenterol 21(40) (2015), 11411–11427.

53.

Marques

M.S.

et al., Afadin downregulation by helicobacter pylori induces epithelial to mesenchymal transition in gastric cells, Front Microbiol 9 (2018), 2712.

54.

E Bartova, J Krejci Harnicarova

Galiova

and Kozubek

, Histone modifications and nuclear architecture: A review, J Histochem Cytochem 56 (2008), 711–721.

55.

Bonisch

Nieratschker

S.M.

Orfanos

N.K.

and Hake

S.B.

, Chromatin proteomics and epigenetic regulatory circuits, Expert Rev Proteomics 5 (2008), 105–119.

56.

Zhu

Edwards

R.J.

and Boobis

A.R.

, Increased expression of histone proteins during estrogen-mediated cell proliferation, Environ Health Perspect 117(6) (2009), 928–934.

57.

Li Gantier

et al., Identification of a histone family gene signature for predicting the prognosis of cervical cancer patients, Sci Rep 7 (2017), 16495.

58.

Fehri

L.F.

Rechner

Janssen

Mak

T.N.

Holland

Bartfeld

Bruggemann

and Meyer

T.F.

, Helicobacter pylori-induced modification of the histone H3 phosphorylation status in gastric epithelial cells reflects its impact on cell cycle regulation, Epigenetics 4 (2009), 577–586.

59.

Ding

S.Z.

et al., Helicobacter pylori-induced histone modification, associated gene expression in gastric epithelial cells, and its implication in pathogenesis, PLoS One 5(4) (2010), e9875.

60.

Xia

Schneider-Stock

Diestel

Habold

Krueger

Roessner

Naumann

and Lendeckel

, Helicobacter pylori regulates p21(WAF1) by histone H4 acetylation, Biochem Biophys Res Commun 369 (2008), 526–531.

61.

L.P.

et al., Functional screen of MSI2 interactors identifies an essential role for SYNCRIP in myeloid leukemia stem cells, Nat Genet 49(6) (2017), 866–875.

62.

Dai

et al., Unraveling molecular differences of gastric cancer by label-free quantitative proteomics analysis, Int J Mol Sc 17(1) (2016), 69.

63.

Wang

Shen

Zhao

and Wang

, Nuclear overexpression of the overexpressed in lung cancer 1 predicts worse prognosis in gastric adenocarcinoma, Oncotarget 8(6) (2017), 9442–9450.

64.

Mertsch

Becker

Lichota

Paulus

and Senner

, Vesicle amine transport protein-1 (VAT-1) is upregulated in glioblastomas and promotes migration, Neuropathol Appl Neurobiol 35(4) (2009), 342–352.

65.

Boyne

J.R.

Colgan

K.J.

and Whitehouse

, Recruitment of the complete hTREX complex is required for Kaposi’s sarcoma-associated herpesvirus intronless mRNA nuclear export and virus replication, PLoS Pathog 4(10) (2008), e1000194.

66.

Yang

et al., 5-methylcytosine promotes mRNA export – NSUN2 as the methyltransferase and ALYREF as an m5C reader, Cell Res 27(5) (2017), 606–625.

67.

Zhang

Zheng

and Guo

, Role of m5C-related regulatory genes in the diagnosis and prognosis of hepatocellular carcinoma, Am J Transl Res 12(3) (2020), 912–922.

68.

Welinder

et al., Correlation of histopathologic characteristics to protein expression and function in malignant melanoma, PLoS One 12(4) (2017), e0176167.

The protein-protein interaction network of intestinal gastric cancer patients reveals hub proteins with potential prognostic value

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Material and methods

2.1 Patients and study design

2.3 Protein extraction, concentration and digestion

2.4 Liquid chromatography and mass spectrometry

2.5 Identification, quantification and in silico analysis

2.6 Survival analysis

3. Results

3.1 Proteomic profile evaluation of the IGC tumor samples

Table 1 Functional enrichment pathways related to the 429 differentially expressed proteins in IGC patients

3.3 Protein-protein interaction (PPI) in IGC patients

3.5 Survival analysis of the hub genes in KM plotter

4. Discussion

Authorship contributions

Supplementary data

sj-pdf-1-cbm-10.3233_CBM-203225.pdf - Supplemental material

sj-docx-1-cbm-10.3233_CBM-203225.docx - Supplemental material

sj-xls-1-cbm-10.3233_CBM-203225.xls - Supplemental material

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Functional enrichment pathways related to the 429 differentially expressed proteins in IGC patients