Network and Pathway-Based Prioritization and Analyses of Genes Related to Chronic Obstructive Pulmonary Disease

Xiaojun Sui, Junfang Yu, Jingbo Wu, Lijuan Guo and Xinjie Shi*

Department of Respiration, Weihai Central Hospital, Weihai, Shandong 264200, China

Received January 15, 2018; accepted March 9, 2018

Summary Chronic obstructive pulmonary disease (COPD), characterized by long-term breathing problem and poor airflow, is the fourth cause of death in the United States. Although previous studies have provided insights into the effects of genetic factors on COPD, the molecular mechanism is still unknown. Here, we proposed a network and pathway-based method for prioritization and analysis of COPD-related genes. In brief, we firstly obtained COPD-related single nucleotide polymorphisms (SNPs) from dbGaP and COPD as well as normal lung tissues gene expression profiles from Gene Expression Omnibus (GEO). Combined with protein–protein interaction pairs, the SNPs and gene expression profiles were imported into dmGWAS, a R package designed for subnetwork searching, to identify COPD-specific module. What’s more, we performed functional analysis for genes in COPD-specific module with the combination of WEB-based GEne SeT AnaLysis Toolkit (WebGe- stalt) and KEGG Orthology Based Annotation System (KOBAS). A COPD-specific network module containing 812 gene nodes and 2640 interactions was obtained. While, 450 of the 812 genes were annotated by WebGestalt and/or KOBAS, which were considered as reliable COPD-related genes. The annotated genes were found to be significantly associated with immune and signal transduction processes, which is consistent with the development of COPD. Finally, 427 of the 450 annotated genes formed 1389 interaction pairs, which might contribute COPD progression. This study should be helpful for understanding COPD mechanisms and its targeted therapy.

Key words COPD, dmGWAS, GEO, SNP, Subnetwork.

COPD is defined by the Global Initiative for Chronic Hassan et al. 2014, Postma et al. 2015). Smoking causes Obstructive Lung Disease (GOLD) as “a common pre- bronchial epithelial cilia shorter and irregular and move- ventable and treatable disease,” demonstrating persistent ment disorders. Air pollution causes excessive nitrogen airflow limitation (Zuo et al. 2014). COPD pulmonary and oxygen compounds and ozone in the air, which re- physiology reflects the sum of pathological changes in sult in airway hyperresponsiveness, oxidative stress and large central airways, small peripheral airways, and the inflammation, this may be an important cause of COPD lung parenchyma. COPD is an important cause of mor- (Ko and Hui 2012). Besides, exposure to dust, childhood bidity, mortality, and health-care costs worldwide. By asthma and respiratory tract infections are thought to 2020, The Global Burden of Disease Study shows that contribute the development of COPD (Decramer et al. COPD will be the third leading cause of global death and 2012). the fifth burden of global disease. In recent years, the etiology mechanism of COPD is Risks of COPD are related to the interaction between further studied and some achievements have been made genetic factors and many different environmental expo- in cellular and molecular biology research. COPD is sures, which could also be affected by a comorbid dis- characterized by inflammation of airway lung paren- ease. Serine protease α1 antitrypsin deficiency is the best chyma and pulmonary vascular. Repeated infection is known genetic factor for COPD, which arises in 1–2% of one of the most important causes of aggravating airway COPD patients (Gooptu et al. 2009). Several other genes inflammation. Inflammatory cells such as cytokines and have also been implicated in COPD, including those inflammatory mediators are involved in the regulation of coding transforming growth factor β1, tumor necrosis COPD. The inflammatory cells infiltration in the airway factor α, and microsomal epoxide hydrolase 1 (Keat- wall are mainly lymphocytes, while in the lumen are ings et al. 2000, Celedon et al. 2004, Cheng et al. 2004). neutrophils of COPD patients (Rytila et al. 2006). These Smoking is the most important risk factor for COPD, 80 activated inflammatory cells release a variety of media- to 90% of COPD patients had a history of smoking (Abu tors that destroy lung structure and cause emphysema, inflammation and edema of the peripheral airway, exces- * Corresponding author, e-mail: [email protected] sive mucus secretion and fibrosis cause airway stenosis DOI: 10.1508/cytologia.83.251 and increase airflow resistance. The major cytokines 252 X. Sui et al. Cytologia 83(3) associated with the pathogenesis of COPD are interleu- Gene expression profiles of 111 COPD and 40 nor- kin-6 (IL-6), interleukin-8 (IL-8), tumor necrosis factor mal lung tissues were downloaded from Gene Expres- alpha (TNF-α), etc. (Plopper and Hyde 2008, Castellani sion Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) et al. 2010). MMPs and neutrophil elastase and its inhib- with the accession number of GSE76925 (Morrow itors together constitute the lung protease-antiprotease et al. 2017). We performed normalization analysis firstly system, and the balance of this system plays an impor- with preprocessCore bioconductor package for the gene tant role in maintaining the normal function of the lung expression profiles for the reliability of the following parenchyma. Research of MMPs family found three analysis. Probe sets were annotated at gene level with COPD related MMPs (MMP-2, 9, 12) which increased in illuminaHumanv4.db package and average expression COPD patients and also caused excessive degradation of values were adopted for genes with multiprobes. collagen and elastin. Elastin residues produced by these mechanisms are important chemical inducers of inflam- Screening COPD-specific module matory cells. While, the positive feedback regulation of DmGWAS (Jia et al. 2011) is an R package imple- these mechanisms will eventually lead to sustained de- ments heuristic local search algorithm to recognize can- struction of lung parenchyma (Vijayan 2013, Ishii et al. didate subnetworks or genes for complex diseases by in- 2014, Sinden et al. 2015). Oxidative stress is another im- tegrating the association signal from GWAS datasets and portant mechanism leading to the occurrence and devel- gene co-expression information into PPI network, which opment of COPD. Oxidant/antioxidant imbalance caused has been efficiently applied in detecting disease-asso- by excessive oxidants and/or antioxidants result in tissue ciated signals. Here, we combined the protein–protein or organ damage. The lung tissues of COPD patients interactions pairs from Hu’s study (Hu et al. 2017) with were exposed to endogenous or exogenous oxidants for Menche’s study (Menche et al. 2015), which resulted in a a long time, meanwhile, the activity of antioxidant en- total of 16022 nodes and 228122 non-redundant interac- zymes and the non-enzymatic antioxidants decreased, tion edges for the recognition of COPD-specific module. which resulted in the enhancement of oxidative stress in We imported the combined PPI network, COPD-relat- vivo. Oxidative stress involves in not only local airways, ed SNPs and gene expression profiles into dmGWAS and but also systemic oxidative stress levels in COPD (Kos- evaluated node and edge weights by GWAS signals, i.e., tikas et al. 2003, Santos et al. 2004). The occurrence and the minimal p-value of SNPs within 20 kb immediately development of COPD is related to the cellular immune upstream or downstream of the specific node, and co- dysfunction of the organism (Tzanakis et al. 2004). A expression status between node pairs respectively. The large number of cytokine chemokines destroy autoim- default parameters, i.e., d=1 and r=1, were used for the mune tolerance and stimulate effector cells. In recent screening of primary COPD-specific modules. years, studies have shown that T cell-mediated autoim- mune response is involved in the progression of COPD. Functional enrichment analysis Although various studies have been done on COPD, the We conducted functional enrichment analysis for mechanism remains unclear. genes in COPD-specific modules with WebGestalt In this study, we propose a network and pathway (Wang et al. 2013) and KOBAS (Xie et al. 2011). Brief- based prioritization and analysis method for COPD re- ly, we obtained significantly enriched Gene Ontology lated genes. We obtained a total of 812 COPD-related (GO) terms from WebGestalt online tool with criteria of genes through the combination of SNP, differential Benjamini-Hochberg adjusted p-value <0.01. Then, the expression and protein–protein interaction analysis. A Kyoto Encyclopedia of Genes and Genomes (KEGG) COPD-related network was constructed through the 450 pathways with FDR <0.01 of COPD-related genes were functional annotated COPD-genes, which might contrib- identified from KOBAS. What’s more, to explore asso- ute to COPD progression. This study should be helpful ciations among KEGG pathways significantly enriched for understanding COPD mechanisms and its targeted in COPD-related genes, we investigated associations therapy. between any pair of KEGG pathways with through two coefficients, i.e. Materials and methods ||AB∩ Overlap Coefficient (OC)= COPD-related SNPs and gene expression profiles min(|AB |, | |) COPD-related SNPs were obtained from the database and of Genotypes and Phenotypes (dbGaP, http://www.ncbi. AB∩ Jaccard Coefficient (JC)= nlm.nih.gov/gap and NCBI dbGaP study accession: AB∪ phs000179.v5.p2), which contains 272 COPD cases and 10099 controls. Genotyping of those samples through in which A and B represent genes in the two tested path- the commercial Illumina Omni-1 chip resulted in a total ways. Here, two pathways were considered as with cross- of 25350 SNPs used for the subsequent data analysis. talk only if at least seven overlapping genes were found 2018 Analyses of Chronic Obstructive Pulmonary Disease 253 between those two pathways. Higher OC and JC indicate with specific GO terms or KEGG pathways as impor- closer associations between the two pathways, and this tant. So, we filtered genes without any functional an- process was named as crosstalk analysis in this study. notation from the primary COPD-specific module, as well as their associated edges to obtain the final COPD- Prioritization and evaluation of COPD-specific module specific module. Besides, to determine the reliability of In this study, we only considered genes annotated the final module, we performed non-randomness analy-

Table 1. The signiﬁcantly enriched biological process terms of the COPD genes.

GO Terms Gene count p-Value FDR

GO:0050878~regulation of body fluid levels 48 9.73E-13 7.41E-10 GO:0090130~tissue migration 31 5.67E-12 1.95E-09 GO:0018212~peptidyl-tyrosine modification 39 7.68E-12 1.95E-09 GO:0002764~immune response-regulating signaling pathway 44 1.78E-11 2.79E-09 GO:0048017~inositol lipid-mediated signaling 28 2.04E-11 2.79E-09 GO:0071417~cellular response to organonitrogen compound 45 2.20E-11 2.79E-09 GO:0040017~positive regulation of locomotion 42 4.18E-11 4.55E-09 GO:0018209~peptidyl-serine modification 33 5.33E-11 5.08E-09 GO:0051090~regulation of sequence-specific DNA binding transcription factor activity 37 1.26E-10 1.07E-08 GO:0051272~positive regulation of cellular component movement 40 1.76E-10 1.34E-08 GO:0050817~coagulation 35 4.25E-10 2.94E-08 GO:0043410~positive regulation of MAPK cascade 42 7.97E-10 5.06E-08 GO:0033674~positive regulation of kinase activity 43 9.06E-10 5.14E-08 GO:0007507~heart development 43 9.64E-10 5.14E-08 GO:0019932~second-messenger-mediated signaling 28 1.01E-09 5.14E-08 GO:0002253~activation of immune response 42 1.09E-09 5.21E-08 GO:0001667~ameboidal-type cell migration 33 1.81E-09 8.09E-08 GO:0003013~circulatory system process 42 1.91E-09 8.09E-08 GO:0032844~regulation of homeostatic process 39 3.34E-09 1.34E-07 GO:0051098~regulation of binding 31 4.43E-09 1.69E-07

Table 2. The top ten most signiﬁcantly enriched KEGG pathways of the COPD genes.

Pathways p-Value FDR Genes included in the pathway

Rap1 signaling pathway 1.43E-08 4.22E-06 CNR1; FGFR4; ADCY3; SIPA1L3; PRKCA; RGS14; VAV2; FGF2; LAT; BCAR1; EGFR; EGF; GRIN2A; CSF1; MAPK14; RAF1; IGF1; PLCB1; MET; FGF18; RASGRP3; RAPGEF4; ITGB1; PIK3R1; PIK3CD; PIK3CA; CALM3; RALGDS; PARD6G; ANGPT1; SKAP1; INSR; FGF7; GNAI1; CRK; TEK; TLN2 Inﬂammatory mediator regulation 3.00E-06 2.96E-04 IL1R1; CAMK2G; CAMK2A; CAMK2D; ADCY3; PRKCA; TRPV1; ITPR1; of TRP channels PRKCE; PRKCQ; MAPK14; IGF1; PLCB1; PIK3R1; PIK3CD; PIK3CA; CALM3; PRKACB; HRH1; PLCG2 Oxytocin signaling pathway 7.25E-06 5.37E-04 CAMK2G; CAMK4; CAMK2A; CAMK2D; CAMK1D; ADCY3; PRKCA; MYLK4; ITPR1; PPP3CA; EGFR; RAF1; GUCY1A2; PLCB1; MYLK; CACNB4; PIK3R1; PIK3CD; PIK3CA; CAMKK1; CALM3; KCNJ4; GNAI1; RYR2; RYR3; PRKACB Calcium signaling pathway 9.81E-06 5.78E-04 CAMK2G; CAMK4; CAMK2A; CAMK2D; ADCY3; PRKCA; PDE1A; MYLK4; ITPR1; PPP3CA; EGFR; GRIN2A; EDNRB; PLCB1; GNA15; MYLK; ITPKC; GNA14; CALM3; PDE1C; RYR2; RYR3; PRKACB; HRH1; ATP2A1; PLCG2; PTGER3; SLC8A1 Gastric acid secretion 1.37E-05 5.78E-04 CAMK2G; CAMK2A; CAMK2D; ADCY3; PRKCA; SSTR2; MYLK4; ITPR1; PLCB1; EZR; MYLK; SLC4A2; CALM3; GNAI1; PRKACB; KCNQ1 Focal adhesion 1.22E-05 5.78E-04 VAV3; PXN; DOCK1; FLNB; BCL2; ITGA1; PRKCA; VAV2; MYLK4; BCAR1; EGFR; EGF; ACTN1; ITGB6; RAF1; VAV1; BIRC2; IGF1; ITGA6; MET; PARVA; MYLK; ITGB1; ITGA9; PIK3R1; PIK3CD; PIK3CA; CRK; TNC; TLN2 HIF-1 signaling pathway 2.45E-05 9.05E-04 CAMK2G; CAMK2A; CAMK2D; BCL2; PRKCA; HK1; EDN1; EGFR; EGF; IGF1; IL6R; LTBR; PIK3R1; PIK3CD; PIK3CA; ANGPT1; INSR; PLCG2; TEK PI3K-Akt signaling pathway 2.83E-05 9.31E-04 BCL2; ITGA1; CDK4; FGFR4; PIK3AP1; PRKCA; CASP9; FGF2; OSMR; EGFR; EGF; SYK; GNG2; CSF1; ITGB6; RAF1; JAK2; JAK1; IGF1; ITGA6; PCK2; MET; PRLR; IL6R; YWHAZ; YWHAE; PKN2; PKN1; FGF18; GHR; ITGB1; ITGA9; PIK3R1; PIK3CD; PIK3CA; HSP90AA1; ANGPT1; INSR; FGF7; TEK; TNC; PPP2R5E Proteoglycans in cancer 3.59E-05 1.06E-03 PXN; CAMK2G; CAMK2A; CAMK2D; TGFB2; HSPG2; FLNB; SDC2; ESR1; PRKCA; VAV2; FGF2; ITPR1; EGFR; MAPK14; RAF1; IGF1; MET; EZR; ITGB1; CD44; PIK3R1; PIK3CD; PIK3CA; ANK2; PRKACB; TWIST2; PLCG2; CBLC Glioma 4.94E-05 1.22E-03 CAMK2G; CAMK2A; CAMK2D; CDK4; PRKCA; EGFR; EGF; RAF1; IGF1; PIK3R1; PIK3CD; PIK3CA; CALM3; PLCG2 254 X. Sui et al. Cytologia 83(3)

Fig. 1. Pathway crosstalk of COPD-related genes’ enriched pathways. Nodes represent pathways, and edges represent associations between pathways. Node size corresponds to the number of COPD-related genes contained in the correspond pathway. Larger size indicates more genes included. Edge-width corresponds to the score, i.e., combination of OC and JC, of specific pathway pair. Larger edge-width indicates the higher score. sis for the module with network python module. Briefly, with regulation of immune system process (e.g., activa- 1000 random networks with the same size with COPD- tion of an immune response, regulation of inflamma- specific module were generated and average clustering tory response and immune response-regulating signaling coefficients of random networks were calculated. Then, pathway) and regulation of phosphorylation (e.g., protein the reliability of COPD-specific module was determined autophosphorylation and positive regulation of kinase by the significance of differences between average clus- activity). The top 20 most significantly enrichment GO tering coefficients of random networks and clustering terms were shown in Table 1. coefficient of COPD-specific module by calculating the Pathway analysis of COPD-related genes obtained a empirical p-value. total of 37 significantly enriched pathways and Table 2 shows the top ten most significantly enriched pathways. Results In line with previous studies, several pathways, e.g., Calcium signaling pathway, Gastric acid secretion, Focal Primary COPD-specific module adhesion, HIF-1 signaling pathway, PI3K-Akt signaling By taking advantage of dmGWAS, candidate disease pathway, and Platelet activation, were enriched in the modules related to COPD were deduced by integrating COPD-related genes. Moreover, inflammatory media- the association signal from GWAS and gene co-expres- tor regulation of TRP channels, Fc gamma R-mediated sion information into PPI network. There were 591 dis- phagocytosis, T cell receptor signaling pathway and ease modules identified. After merging these modules cAMP signaling pathway were also significantly en- and excluding the redundant, we obtained a network riched, which were consistent with the prior knowledge containing 812 nodes and 2640 edges which referred as of pathological process concerning COPD. Crosstalk a primary COPD-specific module. analysis of significantly enriched pathways obtained two mainly pathway clusters that closely associated with im- Enriched functions of COPD-related genes mune and signal transduction respectively (Fig. 1), this Functional enrichment analysis unraveled a more should be an important clue for studying COPD progres- specific function pattern of the COPD-related genes, sion. In Fig. 1, larger node size indicates more genes i.e., genes contained in COPD-specific module. The sig- contained in the pathway and the thicker line indicates nificantly enriched GO terms included those associated more overlapping genes between two KEGG pathways. 2018 Analyses of Chronic Obstructive Pulmonary Disease 255

Fig. 2. Summary plot for expression changes of COPD-related genes, the COPD-related genes smallest p-value of SNP contained in the COPD-related genes and interactions among COPD-related genes. Circular tracks from outside to inside: genome positions by chromosomes (black lines are cytobands), expression heat map of the 450 COPD-related genes whose locations are indicated by elbow connectors, corresponding SNP smallest p-value of the 450 COPD-related genes (larger dot size indicates smaller p-value), interactions within the COPD-speciﬁc network.

Final COPD-specific module the prior knowledge of pathological process concerning Among the 812 nodes in the primary COPD-specific COPD. TRP channels are cation channel superfam- module, 450 of them were found to be annotated with at ily (Venkatachalam and Montell 2007). Their activation least one GO term or KEGG pathway which was thought may be modulated by endogenous stimuli, exogenous to be more reliable. Figure 2 illustrates the expression stimuli, extracellular mediators, intracellular media- changes between COPD and normal lung tissues of the tors, membrane depolarization and physical stimuli such 450 annotated genes. A total of 1389 interaction pairs as osmotic stress and temperature. Different members were formed among the 450 annotated genes and Fig. of the TRP channel family are activated by different 3 illustrates the final COPD-specific module. In the chemical and physical stimuli. Once activated, the tetra- non-randomness validation step, the average cluster- meric TRP channel forms a pore through which cations ing coefficient of the random networks was 0.02, which such as Ca2+ may pass, thus eliciting a wide range of statistically significantly smaller than that of the COPD- cellular functions. Many TRP channels are involved or specific network (clustering coefficient, 0.20; empirical implicated in disease processes of COPD, as well as in p<0.001). Hence, we concluded that the COPD-specific other inflammatory diseases and normal inflammatory network was a non-random network. processes, such as the generation of airway inflammation. The expression of TRPs, TRPA1, TRPC6, TRPM2, Discussion TRPM8, TRPV1 and TRPV4 in key cell-types is important in the development and progression of COPD. There Pathway analysis of COPD-related genes obtained a are reports about the expression of TRPA1 channels in total of 37 significantly enriched pathways. Inflamma- CD4+ and CD8+ T cells, B cells and mast cells, and tory mediator regulation of TRP channels, PI3K-Akt these cells are closely related to COPD (Banner et al. signaling pathway and T cell receptor signaling pathway 2011). TRPA1 is activated by cigarette smoke, acro- were significantly enriched, which were consistent with lein and crotonaldehyde, which are key inflammatory 256 X. Sui et al. Cytologia 83(3)

Fig. 3. COPD-specific network. COPD-specific network contains 427 nodes and 1389 edges. Node color corresponds to its degree in the COPD-specific network. Darker color indicates higher degree. components of the primary cause of COPD. TRPA1 is flammatory mediator release. In addition, activation of known to be an important mediator of cough, a promi- PI3K signaling, decreased PTEN expression may also nent and problematic symptom of COPD. TRPC6 is pre- be important in corticosteroid resistance and acceler- dominantly expressed in macrophages, lymphocytes and ated aging, as well as the increased risk of lung cancer in neutrophils and is also found in the airway epithelium COPD (Barnes 2016). and participate in inflammatory lung diseases. TRPM2 Crosstalk analysis of significantly enriched pathways is implicated in inflammatory responses to oxidative obtained two mainly pathway clusters that closely as- stress. TRPM8 functional variant expressed in human sociated with immune and signal transduction respec- epithelial cells, which promotes endoplasmic reticulum tively, this should be an important clue for studying Ca2+ release leading to increased inflammatory cytokine COPD progression. We found that the two clusters are transcription (Sabnis et al. 2008). TRPV1 is phosphory- connected by Focal adhesion pathway. Focal adhesion lated by protein kinase A (PKA), protein kinase C (PKC) kinase (FAK) is a non-receptor tyrosine protein kinase and other kinases, thought to be important in regulating that could mediate signal transduction pathways, such immune function (Premkumar and Ahern 2000). TRPV4 as FAK-Ras-MAPK, FAK-PI3K and, FAK-Rho/ROCK, has also been shown to contribute to mild temperature and is involved in cell proliferation, migration, apoptosis and ATP-induced increase in ciliary beat frequency in and other processes. Macrophages are the most abundant vitro, and may therefore play a key role in the trans- immune cells in the lung defense system of patients and duction of physical and chemical stimuli into a Ca2+ widely distributed in the alveolar and airway surfaces of signal which regulates mucociliary clearance (Lorenzo COPD. Macrophages regulate and initiate local immune et al. 2008). The PI3K signaling pathway is an impor- inflammatory responses by phagocytosis of inflammato- tant signal cascade that may be activated by oxidative ry cells and secretion of cytokines. Studies have shown stress (Ito et al. 2007), and its prolonged/elevated and that COPD patients are affected by many factors, such as inappropriate activation is associated with various pul- macrophage phagocytosis of airway epithelial cells, par- monary diseases, such as lung cancer (Lee et al. 2006, ticulate matter and the decreased ability of pathogenic Arcaro and Guerreiro 2007), interstitial lung disease microorganisms, resulting in impaired innate immunity (Nho and Hergert 2014), and COPD (Mitani et al. 2016). in the lungs, thereby contributing to the exacerbation The oxidative stress inhibits the protein levels of PTEN of COPD (Leus et al. 2016). Proline-rich tyrosine ki- in patients with COPD, which results in the persistent nase2 (Pyk2) is a new member of the focal adhesion activation of the PI3K/Akt pathway and resultant proin- kinase family. Pyk2 regulates the activity of three major 2018 Analyses of Chronic Obstructive Pulmonary Disease 257 members of the MAPKs family, including extracellular sociated with COPD, we obtained COPD-specific genes signal-regulated kinase (ERK), Jun-N-terminal kinase and network. This network and pathway-based method (JNK) and P38 protein kinase (P38MAPK), and plays a should provide important clues to prioritize and analyze key role in the G protein-mediated activation of MAPK. COPD-related genes. Pyk2 activation exists during phagocytosis, and recent studies have found that Pyk2 plays a pivotal role in the Acknowledgements IL-8-mediated chemotaxis of human neutrophils (Di Ci- occio et al. 2004). I am greatly indebted to Xinjie Shi, for her valuable In the COPD-specific network, there are several instructions and suggestions on my thesis as well as her nodes with higher degrees (deep color nodes). EGFR careful reading of the manuscript. My heartfelt thanks and HSP90AA1 are two of the higher degree nodes. also go to Junfang Yu and Jingbo Wu for their help in Growth factors mediate various cellular responses to the making of this thesis. Then, I also want to thank environmental stimuli. Specifically, exposure of lung Lijuan Guo for the precious help. This research did not epithelium to oxidative stress induced by cigarette receive any specific grant from funding agencies in the smoke could stimulate aberrant epidermal growth fac- public, commercial, or not-for-profit sectors. tor receptor (EGFR, also known as ERBB) family activation. EGFR activity is promoted in COPD patients References (Wang et al. 2000, West et al. 2003). Cigarette smoke directly stimulates the epidermal growth factor recep- Abu Hassan, H., Abd Aziz, N., Hassan, Y. and Hassan, F. 2014. Does tor (EGFR), leading to the replication of mucin gene the duration of smoking cessation have an impact on hospital admission and health-related quality of life amongst COPD pa- MUC5A and excessive secretion of mucus, resulting tients? Int. J. Chron. Obstruct. Pulmon. Dis. 9: 493–498. in the reduction of mucociliary clearance and repeated Arcaro, A. and Guerreiro, A. S. 2007. The phosphoinositide 3-kinase exacerbation of inflammation. TGF-β1 and EGFR are pathway in human cancer: genetic alterations and therapeutic involved in the activation and proliferation of fibroblasts, implications. Curr. Genomics 8: 271–306. leading to fibrosis of peribronchial and airway obstruc- Banner, K. H., Igney, F. and Poll, C. 2011. TRP channels: emerging targets for respiratory disease. Pharmacol. Ther. 130: 371–384. tion. Cigarette smoking activates PI-3 kinase activity via Barnes, P. J. 2016. Kinases as novel therapeutic targets in asthma EGFR activation, and increased Akt phosphorylation is and chronic obstructive pulmonary disease. Pharmacol. Rev. 68: seen in a bronchial tissue of smokers. Phosphorylation 788–815. of both EGFR and Akt are increased by oxidative stress, Castellani, M. L., De Lutiis, M. A., Toniato, E., Conti, F., Felaco, P., which has been implicated in the pathogenesis of COPD. Fulcheri, M., Theoharides, T. C., Caraffa, A., Antinolfi, P., Conti, Shyamala’s research identified biochemical networks P., Cuccurullo, C., Ciampoli, C., Felaco, M., Orso, C., Salini, V., Cerulli, G., Kempuraj, D., Tete, S. and Shaik, B. 2010. Impact in airway epithelial cells (AEC) of COPD patients, by of RANTES, MCP-1 and IL-8 in mast cells. J. Biol. Regul. Ho- which activation of EGFR, PI-3 kinase and Akt was meost. Agents 24: 1–6. persistently maintained, and results in increased airway Celedon, J. C., Lange, C., Raby, B. A., Litonjua, A. A., Palmer, L. J., inflammation, cytoplasmic retention of FoxO3A and in- DeMeo, D. L., Reilly, J. J., Kwiatkowski, D. J., Chapman, H. A., creasing levels of neutrophil chemokines IL-8 (Ganesan Laird, N., Sylvia, J. S., Hernandez, M., Speizer, F. E., Weiss, S. T. and Silverman, E. K. 2004. The transforming growth factor- et al. 2013). beta1 (TGFβ1) gene is associated with chronic obstructive pul- Heat shock protein in (HSP) family plays an impor- monary disease (COPD). Hum. Mol. Genet. 13: 1649–1656. tant role in cell growth, development, differentiation, Cheng, S. L., Yu, C. J., Chen, C. J. and Yang, P. C. 2004. Genetic protein synthesis, gene transcription and so on. In re- polymorphism of epoxide hydrolase and glutathione S-transfer- cent years, HSP has been found to be closely related to ase in COPD. Eur. Respir. J. 23: 818–824. Chung, K. F. 2005. Inflammatory mediators in chronic obstructive chronic airway inflammation. The study has shown that pulmonary disease. Curr. Drug Targets Inflamm. Allergy 4: HSP27, HSP70, and HSP90α were significantly altered 619–625. in patients suffering from COPD (Ganesan et al. 2013). Decramer, M., Janssens, W. and Miravitlles, M. 2012. Chronic ob- HSP70 plays a critical role in the process of inflamma- structive pulmonary disease. Lancet 379: 1341–1351. tion and innate immune response under environmental Di Cioccio, V. R., Strippoli, R., Bizzarri, C., Troiani, G., Cervellera, M. N., Gloaguen, I., Colagrande, A., Cattozzo, E. M., Pagliei, S., stress (Chung 2005). HSP70 expression was increased in Santoni, A., Colotta, F., Mainiero, F. and Bertini, R. 2004. Key lung tissues of COPD patients and current smokers. At role of proline-rich tyrosine kinase 2 in interleukin-8 (CXCL8/ the same time, HSP70 is also involved in the secretion IL-8)-mediated human neutrophil chemotaxis. Immunology 111: of airway mucin. Extracellular HSP70 could regulate 407–415. chemokine productions and EGFR phosphorylation and Ganesan, S., Unger, B. L., Comstock, A. T., Angel, K. A., Mancuso, plays an important role in the CSE-induced inflamma- P., Martinez, F. J. and Sajjan, U. S. 2013. Aberrantly activated EGFR contributes to enhanced IL-8 expression in COPD airways tory and innate immune responses in bronchial epithelial epithelial cells via regulation of nuclear FoxO3A. Thorax 68: cells. 131–141. In conclusion, through the combination of the genetic Gooptu, B., Ekeowa, U. I. and Lomas, D. A. 2009. Mechanisms of polymorphisms and differential expression genes as- emphysema in alpha1-antitrypsin deficiency: molecular and cel- 258 X. Sui et al. Cytologia 83(3)

lular insights. Eur. Respir. J. 34: 475–488. type I collagen matrix-induced cell death by suppressing low au- Hu, Y. S., Xin, J., Hu, Y., Zhang, L. and Wang, J. 2017. Analyzing the tophagy via aberrant Akt/mTOR kinases. PLoS ONE 9: e94616. genes related to Alzheimer’s disease via a network and pathway- Plopper, C. G. and Hyde, D. M. 2008. The non-human primate as a based approach. Alzheimer’s Res. Ther. 9: 29. model for studying COPD and asthma. Pulm. Pharmacol. Ther. Ishii, T., Abboud, R. T., Wallace, A. M., English, J. C., Coxson, H. 21: 755–766. O., Finley, R. J., Shumansky, K., Pare, P. D. and Sandford, A. J. Postma, D. S., Bush, A. and van den Berge, M. 2015. Risk factors and 2014. Alveolar macrophage proteinase/antiproteinase expression early origins of chronic obstructive pulmonary disease. Lancet in lung function and emphysema. Eur. Respir. J. 43: 82–91. 385: 899–909. Ito, K., Caramori, G. and Adcock, I. M. 2007. Therapeutic potential of Premkumar, L. S. and Ahern, G. P. 2000. Induction of vanilloid recep- phosphatidylinositol 3-kinase inhibitors in inflammatory respira- tor channel activity by protein kinase C. Nature 408: 985–990. tory disease. J. Pharmacol. Exp. Ther. 321: 1–8. Rytila, P., Plataki, M., Bucchieri, F., Uddin, M., Nong, G., Kinnula, V. Jia, P., Zheng, S., Long, J., Zheng, W. and Zhao, Z. 2011. dmGWAS: L. and Djukanovic, R. 2006. Airway neutrophilia in COPD is not dense module searching for genome-wide association studies in associated with increased neutrophil survival. Eur. Respir. J. 28: protein–protein interaction networks. Bioinformatics 27: 95–102. 1163–1169. Keatings, V. M., Cave, S. J., Henry, M. J., Morgan, K., O’Connor, C. Sabnis, A. S., Reilly, C. A., Veranth, J. M. and Yost, G. S. 2008. In- M., FitzGerald, M. X. and Kalsheker, N. 2000. A polymorphism creased transcription of cytokine genes in human lung epithelial in the tumor necrosis factor-alpha gene promoter region may pre- cells through activation of a TRPM8 variant by cold tempera- dispose to a poor prognosis in COPD. Chest 118: 971–975. tures. Am. J. Physiol. Lung Cell. Mol. Physiol. 295: L194–L200. Ko, F. W. and Hui, D. S. 2012. Air pollution and chronic obstructive Santos, M. C., Oliveira, A. L., Viegas-Crespo, A. M., Vicente, L., pulmonary disease. Respirology 17: 395–401. Barreiros, A., Monteiro, P., Pinheiro, T. and Bugalho De Almei- Kostikas, K., Papatheodorou, G., Psathakis, K., Panagou, P. and da, A. 2004. Systemic markers of the redox balance in chronic Loukides, S. 2003. Oxidative stress in expired breath condensate obstructive pulmonary disease. Biomarkers 9: 461–469. of patients with COPD. Chest 124: 1373–1380. Sinden, N. J., Baker, M. J., Smith, D. J., Kreft, J. U., Dafforn, T. R. Lee, S. Y., Kang, E. J., Hur, G. Y., Jung, K. H., Jung, H. C., Lee, S. Y., and Stockley, R. A. 2015. alpha-1-antitrypsin variants and the Kim, J. H., Shin, C., In, K. H., Kang, K. H., Yoo, S. H. and Shim, proteinase/antiproteinase imbalance in chronic obstructive pul- J. J. 2006. Peroxisome proliferator-activated receptor-gamma monary disease. Am. J. Physiol. Lung Cell. Mol. Physiol. 308: inhibits cigarette smoke solution-induced mucin production in L179–L190. human airway epithelial (NCI-H292) cells. Am. J. Physiol. Lung Tzanakis, N., Chrysofakis, G., Tsoumakidou, M., Kyriakou, D., Tsili- Cell. Mol. Physiol. 291: L84–L90. gianni, J., Bouros, D. and Siafakas, N. M. 2004. Induced sputum Leus, N. G., van der Wouden, P. E., van den Bosch, T., Hooghiemstra, CD8+ T-lymphocyte subpopulations in chronic obstructive pul- W. T., Ourailidou, M. E., Kistemaker, L. E., Bischoff, R., Gos- monary disease. Respir. Med. 98: 57–65. ens, R., Haisma, H. J. and Dekker, F. J. 2016. HDAC 3-selective Venkatachalam, K. and Montell, C. 2007. TRP channels. Annu. Rev. inhibitor RGFP966 demonstrates anti-inflammatory properties Biochem. 76: 387–417. in RAW264.7 macrophages and mouse precision-cut lung slices Vijayan, V. K. 2013. Chronic obstructive pulmonary disease. Indian J. by attenuating NF-kappaB p65 transcriptional activity. Biochem. Med. Res. 137: 251–269. Pharmacol. 108: 58–74. Wang, J., Duncan, D., Shi, Z. and Zhang, B. 2013. WEB-based gene Lorenzo, I. M., Liedtke, W., Sanderson, M. J. and Valverde, M. A. set analysis toolkit (WebGestalt): update 2013. Nucleic Acids 2008. TRPV4 channel participates in receptor-operated calcium Res. 41 W1: W77-83. entry and ciliary beat frequency regulation in mouse airway Wang, X., McCullough, K. D., Franke, T. F. and Holbrook, N. J. 2000. epithelial cells. Proc. Natl. Acad. Sci. U.S.A. 105: 12611–12616. Epidermal growth factor receptor-dependent Akt activation Menche, J., Sharma, A., Kitsak, M., Ghiassian, S. D., Vidal, M., by oxidative stress enhances cell survival. J. Biol. Chem. 275: Loscalzo, J. and Barabasi, A. L. 2015. Disease networks. Un- 14624–14631. covering disease–disease relationships through the incomplete West, K. A., Brognard, J., Clark, A. S., Linnoila, I. R., Yang, X., interactome. Science 347: 1257601. Swain, S. M., Harris, C., Belinsky, S. and Dennis, P. A. 2003. Mitani, A., Ito, K., Vuppusetty, C., Barnes, P. J. and Mercado, N. Rapid Akt activation by nicotine and a tobacco carcinogen mod- 2016. Restoration of corticosteroid sensitivity in chronic obstruc- ulates the phenotype of normal human airway epithelial cells. J. tive pulmonary disease by inhibition of mammalian target of Clin. Invest. 111: 81–90. rapamycin. Am. J. Respir. Crit. Care Med. 193: 143–153. Xie, C., Mao, X., Huang, J., Ding, Y., Wu, J., Dong, S., Kong, L., Gao, Morrow, J. D., Zhou, X., Lao, T., Jiang, Z., DeMeo, D. L., Cho, M. G., Li, C. Y. and Wei, L. 2011. KOBAS 2.0: a web server for an- H., Qiu, W., Cloonan, S., Pinto-Plata, V., Celli, B., Marchetti, N., notation and identification of enriched pathways and diseases. Criner, G. J., Bueno, R., Washko, G. R., Glass, K., Quackenbush, Nucleic Acids Res. 39: W316–W322. J., Choi, A. M., Silverman, E. K. and Hersh, C. P. 2017. Func- Zuo, L., He, F., Sergakis, G. G., Koozehchian, M. S., Stimpfl, J. N., tional interactors of three genome-wide association study genes Rong, Y., Diaz, P. T. and Best, T. M. 2014. Interrelated role of are differentially expressed in severe chronic obstructive pulmo- cigarette smoking, oxidative stress, and immune response in nary disease lung tissue. Sci. Rep. 7: 44232. COPD and corresponding treatments. Am. J. Physiol. Lung Cell. Nho, R. S. and Hergert, P. 2014. IPF fibroblasts are desensitized to Mol. Physiol. 307: L205–L218.