Abstract
Background
To explore the similar and different pathogenesis between non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC).
Methods
This study used bioinformatics methods, including functional enrichment analysis, compared the topological features of SCLC and NSCLC in the human protein interaction network in a system aspect, and analyzed the highly intense modules from an integrated network.
Results
This study included 5082 and 2781 significantly different expression genes for NSCLC and SCLC, respectively. The differently expressed genes of NSCLC are mainly distributed in the extracellular region and synapse. By contrast, the genes of SCLC are located in the organelle, macromolecular complex, membrane-enclosed lumen, cell part, envelope, and synapse. Compared with SCLC, the differently expressed genes of NSCLC act in the biological regulation, multicellular organismal process, and viral reproduction and locomotion, which show that NSCLC is more likely to cause a wide range of cancer cell proliferation and virus infection than SCLC. The network topological properties of SCLC and NSCLC are similar, except the average shortest path length, which indicates that most of the genes of the two lung cancers play a similar function in the entire body. The commonly expressed genes show that all of the genes in the module may also cause NSCLC and SCLC, simultaneously.
Conclusions
The proteins in module will involve the same or similar biological functions and the interactions among them induce the occurrence of lung cancer. Moreover, a potential biomarker of SCLC is the interaction between APIP and apoptotic protease activating factor (APAF)1, which share a common module.
Keywords: Bioinformatics, lung cancer, pathology
Introduction
Lung cancer is the leading cause of cancer related death with millions of deaths worldwide each year,1 and is divided into two main types, namely, small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC).2 SCLC accounts for approximately 20%3,4 and NSCLC makes up the majority of lung cancers, including three sub-types: squamous cell carcinoma, large cell carcinoma, and adenocarcinoma.5,6 NSCLC responds to treatment differently to SCLC, and it is often difficult to discern which type the patient has.7 Lung cancer therapy is based on clinical stages, but 80% of patients with lung cancer at the time of diagnosis are already in advanced stage. Consequently, the treatment of lung cancer is more aggressive in the early stages and more palliative in the late stages, because the cancer must be stopped in the early stages to have a better chance of eradication. Compared with SCLC, the development and growth of NSCLC is relatively slow and the spread is relatively late.8 Understanding the difference between the two types of cancer can enhance the therapeutic effects and alleviate the suffering of patients. Therefore, it is necessary to distinguish the occurrence mechanism of the two cancers, in view of a more effective and individualized treatment.9
Carcinogenesis is a gradual process, involving multiple genetic mutations, and the occurrence and development of lung cancer is a complex pathological process involving multiple molecules and stages.10 Based on system biology, the occurrence of lung carcinoma may be inhibited or guided by a regulation system composed of genes, proteins, and other molecules, and raise the differentially expressed genes in the core position in the system.11 In the study of liver cancer, multiple related molecules may cause tumors, performing as a “group.”12 In theory, molecular pathology change takes place throughout the process of lung cancer and can be used as a monitoring index of lung cancer development. Although there is much research on the occurrence mechanism of lung cancer, molecular functions are yet to be tested. Therefore, it is important to study the development of lung cancer based on systemic and integrated methods. By constructing the NSCLC related protein network, research has confirmed that the occurrence of lung cancer is associated with a complex system formed by molecular groups that are connected with tumors, rather than a single or a few genes or proteins.13 Identifying the method of complex networks has been significant in the research of cancer in recent years and has attracted researchers from various areas, including mathematics, physics, and biology.14 The approach of network analysis has been proven useful to organize high throughput biological datasets and extract significant information, which is mainly based on the relationship among nodes including genetic regulatory interactions, gene co-expression, physical and/or chemical interactions or other shared property between nodes.15 From a global perspective, network analysis approaches are essential to discover non-obvious, but intrinsically important nodes, such as the predicted proteins and functional modules.16,17 Because of the complexity of lung cancer occurrence, it is necessary to study the different mechanisms for SCLC and NSCLC at a systems level.
Based on the interaction of proteins in the biological system, it is important to understand the principle of the protein in biological systems, the biological signal and reaction mechanism of energy metabolism in diseases and other special physiological states, and the relationship between proteins. Here, we studied the expression of different genes for SCLC and NSCLC and the different genes that are commonly expressed in the two kinds of lung cancers. Utilizing the GO and KEGG databases for significantly different expressed genes, we performed enrichment analyses of molecular function, cell component, biological process, and pathway. We constructed the lung cancer protein interaction network and compared this to the network topological properties of the two different genes, and mined the module to study the possible role and molecular mechanism of the module genes, which provides a new research theory for lung cancer research.
Material and method
Data
The gene expression profile GSE40275, published on 25 August 2012 from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/), focused on the differences in gene expression profiles among SCLC, NSCLC, and normal lung samples regarding the expression of gene encoding for proteins with G protein-coupled receptor activity. We grouped the adenocarcinoma, squamous cell, and large cell carcinomas into NSCLC. In our research, there were 84 samples, including 25 SCLC, 16 NSCLC, and 43 normal human lung ribonucleic acid (RNA) samples and the gene expression analysis was performed using affymetrix microarrays. We downloaded the data and analyzed the mechanisms of SCLC and NSCLC based on a global view. Using a t-test for differential expression analysis, we compared the gene expression profiles of SCLC, NSCLC, and normal tissues. We acquired significantly different expression genes using a P-value of less than 0.05 as the signature. Based on the GPL15947, we translated the probe number of genes into gene symbols. Consequently, we obtained 9682 different expression genes of NSCLC and 7381 different expression genes of SCLC, with 4600 genes shared by NSCLC and SCLC. We downloaded the micro (mi)RNA from miRBase (Release 20: June 2013, http://www.mirbase.org/), and selected the miRNAs that targeted the different expressed genes of SCLC and NSCLC.
The gene expression patterns of non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC)
We removed the 4600 shared genes and extracted the significant expression genes between NSCLC and SCLC. To analyze the change of the gene expression pattern in NSCLC and SCLC and the metabolic pathway the differential expression genes are involved in, we enriched the significantly different expression genes into GO, KEGG pathway, and UP-tissue.
Building the protein interaction network
We downloaded human protein interactions from the Human Protein Reference Database (HPRD, Release 9: 13 April 2010, http://www.hprd.org) and translated protein names into gene symbols. As a result, we obtained 39 184 human protein interactions, including 9617 genes. We mapped the signatures of NSCLC and SCLC into the protein interactions and obtained 3057 protein interactions of lung cancer, including 6512 different expression genes of NSCLC, 4221 different expression genes of SCLC, and 1964 common genes. Using Cytoscape software, we constructed a protein interaction network and extracted 10 network topological properties, mainly divided into two categories: single proteins, such as degree, clustering coefficient, topological coefficient; and the remainder demonstrate the relationship of other proteins, such as average shortest path length. All of the nodes in network topology properties are calculated. We then extracted the network topological properties of NSCLC and SCLC different expressed genes and used the Wilcox rank sum to scan the different characteristics.
Degree
The degree is the basic topological property of a single node and describes the number of connected nodes in a network. More specifically, the degree of node i is the number of nodes which are directly linked to node i and recorded by ki.
Studies have shown that in the biological molecular network, such as the protein interaction network, the necessary genes or translation products that support basic life activities occur more frequently in central nodes than in general nodes. In the human protein interaction network, the central nodes enriched the genes, which are related to the genetic disorder.18
Average shortest path length
The average shortest path length shows the global property of the network and measures compactness.19
Average distance to protein (ADT):
 |
where M is the number of total proteins; and dij is the length of the shortest path between node i and node j.
Betweenness
Betweenness measures the proportion of shortest path where a node occurs in other nodes, and shows the role of a node in connecting other nodes.20 The higher the betweenness, the more important that nodes maintain network connectivity:
 |
where σkj denotes the shortest paths between node pairs k and j, and σkij denotes that path through the node i.
Closeness centrality
Closeness is a measure of how long a node spreads information from node i to all other nodes sequentially,21 and is measured like the random walk closeness centrality i that measures the speed with which randomly walking messages reach a vertex from elsewhere in the network – a sort of random walk version of closeness centrality.
Clustering coefficient
In the amount of network, if node i is linked with node j and node j connects node k, then it is possible that node i connects with node k. This phenomenon indicates that there is an intensive connection property for some part of nodes and we can present this property with a clustering coefficient (CC).22
Clustering coefficient:
 |
where ni is the number of links connecting the ki neighbors of node i to each other. The range of CC is from 0 to 1. When all of the neighbors of node i are connected intensively, the value of CC is 1. On the contrary, the value of CC is 0 when there is no connection among the neighbors.
Eccentricity
Eccentricity shows the maximum non-limited length of a shortest path between n and another node in the network. If n is an isolated node, the value of this eccentricity is 0.
Neighborhood connectivity
The neighborhood connectivity of node n is the average connectivity of all neighbors of n, and its distribution gives the average of the neighbourhood connectivities of all nodes n with k neighbors for k = 0,1, ….
Radiality
By subtracting the average shortest path length of node n from the diameter of the connected component plus 1, the radiality is computed, which is a node centrality index. Because the diameter of the connected component divides the radiality of each node, it is a number between 0 and 1.
Stress
Stress is the number of shortest paths passing through a node. If a node is passed by a high number of paths, then its stress is relatively high. This parameter is defined only for networks without multiple edges.
Topological coefficient
The topological coefficient reflects the proportion that the neighbors of a node are shared by other nodes.23
Topological coefficient:
 |
where Cij is the number of nodes that are connected by node i and node j; and mi is the node set that shares the neighbours of node i. We extracted these topological properties and then screened the significant different properties using the Wilcoxon rank sum test.
Constructing the protein interaction network of lung cancer
Earlier studies have proposed that the occurrence of cancer is based on molecular group. Thus, we investigated whether the occurrence of SCLC and NSCLC is related to molecular group. First, we obtained the protein interaction pairs from HPRD and converted them into the gene symbols. Using Cytoscape software, we then constructed the PPI network and acquired the topology parameters of all nodes for the entire network through network analysis. Based on the protein interaction relationship, we reflected the different expressed genes of SCLC and NSCLC into the PPI network and acquired the network topological properties of the two lung cancers, and then extracted the remarkable and different network topological properties using the Wilcox rank sum test. Subsequently, we extracted the protein interactions of SCLC and NSCLC in PPI and combined the two kinds of protein interaction pairs, which were utilized to construct the network of lung cancer different expressed genes using Cytoscape. In the network, the four node attributes are: yellow for NSCLC; green for SCLC; red represents both gene expression differences; and dark blue represents the lung cancer genes. From this point onward, we excavated the highly relevant modules using the plugin, MINE, and enriched the genes of modules for GO function.
Results
Gene function enrichment of SCLC and NSCLC
After testing, the number of different expressed genes for NSCLC and SCLC were 9682 and 7381, respectively, and there were 4600 different genes shared by both diseases. After eliminating the commonly expressed genes using the DAVID online tool, we analyzed the function of the two cancers from four aspects: molecular function, cell composition, biological process, and metabolic pathways.
Cellular component
We extracted the feature sets of the P-values that were less than 0.05. The enrichment shows that the different expressed genes of NSCLC are mainly distributed in the extracellular region and synapse, while the genes of SCLC are mainly located in the organelle, macromolecular complex, membrane-enclosed lumen, cell part, envelope, and synapse.
From the comparison, there are more extracellular components in NSCLC than in SCLC, which indicates that NSCLC is more likely to spread than SCLC. The NSCLC related gene plays an important role in cell interaction, while SCLC related genes are mainly distributed in the organelles and the cell membrane. There are a large number of genes for the two cancers located in the synapse distribution, suggesting that the disease will appear neurologically after the cancer gene of the synapse is activated. Research on the components of the two types of lung cancer cells and their distribution assists doctors to choose a treatment program in order to achieve the goal of personalized therapy.
Biology process
Research on the biology process of disease will make pathogenesis of the disease clearer; therefore, we enriched the differentially expressed genes of NSCLC and SCLC in the biological process (Table 1). From the enrichment of the biological process, both types of cancers play a pivotal role in the regulation of cell proliferation and programmed cell death. There are more differently expressed genes in NSCLC than in SCLC in biological regulation, multicellular organismal process, viral reproduction and locomotion, which shows that NSCLC is more likely to cause a wide range of cancer cell proliferation and virus infection than SCLC. The enriched processes of different expressed genes for SCLC are cellular component biogenesis and metabolic process, identifying that SCLC could synthesize the cancer cell component quickly, and, therefore, it is difficult to eradicate.
Table 1.
Comparison of biology process in NSCLC and SCLC
| NSCLC | SCLC |
|---|---|
| Multicellular organismal process | Cellular process |
| Developmental process | Cellular component organization |
| Response to stimulus | Metabolic process |
| Biological regulation | Cellular component biogenesis |
| Immune system process | Death |
| Localization | Developmental process |
| Locomotion | |
| Establishment of localization | |
| Multi-organism process | |
| Reproductive process | |
| Reproduction | |
| Biological adhesion | |
| Rhythmic process | |
| Cellular process | |
| Death | |
| Cellular component organization | |
| Cell killing | |
| Viral reproduction |
NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
Molecular function
NSCLC and SCLC play the same role in binding and transcription regulator activity. Compared to SCLC (Table 2), NSCLC acts in enzyme regulator activity, molecular transducer activity, and structural molecule activity, and the activation of these materials means that NSCLC is more likely to spread and is difficult to eliminate. However, the catalytic activity in SCLC is more powerful than in NSCLC; although the SCLC cell is small, the reaction is strong. While radiotherapy on SCLC will not cause as much damage to normal tissues/cells, it is not especially effective.
Table 2.
Comparison of molecular function in NSCLC and SCLC
| NSCLC | SCLC |
|---|---|
| Binding | Binding |
| Enzyme regulator activity | Catalytic activity |
| Transporter activity | Transcription regulator activity |
| Molecular transducer activity | |
| Structural molecule activity | |
| Transcription regulator activity |
NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
Pathway
In the enrichment of KEGG pathways(Table 3), the different expressed genes of SCLC play a pivotal role in the pathways of gene expression, such as spliceosome, DNA replication, RNA degradation, RNA polymerase, base excision repair, and homologous recombination.
Table 3.
Comparison of pathways in NSCLC and SCLC
| NSCLC | SCLC |
|---|---|
| Cytokine-cytokine receptor interaction | Spliceosome |
| Adherens junction | Aminoacyl-tRNA biosynthesis |
| Autoimmune thyroid disease | Base excision repair |
| Calcium signaling pathway | Cell cycle |
| Chemokine signaling pathway | Citrate cycle (TCA cycle) |
| Complement and coagulation cascades | DNA replication |
| Drug metabolism | Fructose and mannose metabolism |
| Endocytosis | Glycolysis/Gluconeogenesis |
| Galactose metabolism | Homologous recombination |
| GnRH signaling pathway | Huntington's disease |
| Jak-STAT signaling pathway | Lysine degradation |
| Long-term potentiation | N-Glycan biosynthesis |
| Metabolism of xenobiotics by cytochrome P450 | Nucleotide excision repair |
| Neuroactive ligand-receptor interaction | Oocyte meiosis |
| NOD-like receptor signaling pathway | Parkinson's disease |
| Pathogenic Escherichia coli infection | Pentose phosphate pathway |
| Progesterone-mediated oocyte maturation | |
| Proteasome | |
| Pyrimidine metabolism | |
| Ribosome | |
| RNA degradation | |
| RNA polymerase | |
| Systemic lupus erythematosus | |
| Ubiquitin mediated proteolysis | |
| Valine, leucine and isoleucine degradation |
GnRH, gonadotropin-releasing hormone; Jak-STAT, Janus kinase/signal transducers and activators of transcription; NOD, nucleotide-binding oligomerization; NSCLC, non-small cell lung cancer; tRNA, transfer ribonucleic acid; SCLC, small cell lung cancer.
Moreover, the different expressions of SCLC genes are involved in the metabolism, cell cycle, and some diseases including Huntington's disease, Parkinson's disease, and systemic lupus erythematosus. NSCLC gene expression is mainly indicated in the signaling pathways, including the calcium, chemokine, gonadotropin-releasing hormone (GnRH), Janus kinase/signal transducers and activators of transcription (Jak-STAT), and nucleotide-binding oligomerization (NOD)-like receptor signaling pathways. The different expressed genes of NSCLC act in the Cytokine-cytokine receptor and neuroactive ligand-receptor interactions and are involved in autoimmune thyroid disease.
Comparison of the network topology properties for NSCLC and SCLC
As shown in the Table 4, the difference in most network topological properties in NSCLC and SCLC is slight, except the for the average shortest path length (P < 0.05). The average shortest-path length is a concept in network topology that is defined as the average number of steps along the shortest paths for all possible pairs of network nodes. It is a measure of the efficiency of information or mass transport in a network. The significant difference between NSCLC and SCLC in the average shortest path shows the distinction in their interaction with other proteins. The average shortest path is a significant measurement of network tightness and indicates the average value of the shortest paths between any two proteins.
Table 4.
Comparison of network topology properties in NSCLC and SCLC
| Topological feature | P-value |
|---|---|
| Average shortest path length | 0.004259 |
| Betweenness centrality | 0.999997 |
| Closeness centrality | 0.999734 |
| Clustering coefficient | 0.877423 |
| Degree | 0.999886 |
| Eccentricity | 0.88478 |
| Neighborhood connectivity | 0.999388 |
| Radiality | 0.998826 |
| Stress | 0.999998 |
| Topological coefficient | 0.393612 |
NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.
Lung cancer related protein network
The difference in the average shortest path of NSCLC and SCLC is obvious, suggesting that the interaction with other proteins in the two types of cancer is significantly varied. Therefore, we constructed the protein interaction network of lung cancer related protein and mined the stable module in the network. We extracted the protein interaction pairs for the different expressed genes of NSCLC and SCLC, and combined the two types of lung cancer gene pairs. We then constructed the protein interaction network of lung cancer genes using Cytoscape software, where commonly and special expressed genes are marked by different colors (Fig 1).
Figure 1.
The lung cancer related protein network and the modules of the lung cancer related protein network. Orange represents non-small cell lung cancer (NSCLC); greens represents small cell lung cancer (SCLC); red denotes commonly expressed genes; and dark blue denotes non-lung cancer genes.
We utilized the Cytoscape plugin, MINE, to reveal the modules of the lung cancer protein network and obtained 28 modules. These modules show that genes from the same family share a similarly expressed pattern and proteins interact with other proteins to participate in the biological processes, though not through a single protein, and tend to function in a module.
Discussion
The morbidity and mortality of a malignant tumor is a serious threat to health. Clinically, based on histopathology, lung cancer is mainly divided into SCLC and NSCLC. However, the molecular mechanism of the two is still not clear, which makes it difficult for early diagnosis and effective treatment. Histopathology, used to identify the type of lung cancer, can result in false positives and is inconvenient to patients in that it is a difficult or uncomfortable process. In view of this, we used bioinformatics methods to study the similarities and differences of the pathogenesis between these two kinds of lung cancer. The functional annotation for the differentially expressed data of lung cancer genes shows that NSCLC and SCLC share similar occurrence mechanisms, such as binding and transcription regulator activity. NSCLC related genes play an important role in cell interaction and signaling pathways, while SCLC related genes are mainly distributed in the organelles and the cell membrane and act in the cell cycle, including in RNA degradation and DNA replication. The differences of two lung cancer genes were mapped to the human protein interaction network and the significant topological feature of the network was the average shortest path. The average shortest path of NSCLC is longer than SCLC, which indicated that more genes are enrolled in the occurrence of NSCLC than SCLC and are, therefore, more likely to cause a wider range of cancer cell proliferation and virus infection in the form of molecular modules.
The proteins in a same module will involve the same or similar biological functions and the interaction among them promotes the occurrence of lung cancer. Thus, we take three modules (modules 2, 4, and 12) as examples to study the biological function of the different expressed genes SCLC and NSCLC. In modules 2 and 4, the non-commonly expressed genes of NSCLC and SCLC are shared in a module and we studied these modules to identify the common function that SCLC and NSCLC genes are involved in. In module 12, we obtained the different expression and interaction genes in SCLC to acquire the possible oncogene sets.
In module 2, there are four genes: APOA1BP, APOA2, APOA1 and APOF. APOA1BP is significantly expressed in SCLC, while the other three genes are differently expressed in NSCLC.
The APOA1BP gene plays a role in sperm capacitation, encodes the apolipoprotein (APO) A-I-binding protein, and binds to APOA1, APOA2, and high-density lipoprotein (HDL). APO A-II encoded by APOA2, is the second most abundant protein of the HDL particles and acts in plasma as a monomer, homodimer, or heterodimer with APOD.24 Defects in APOA2 may result in hypercholesterolemia and APO A-II deficiency.25 APOAPOAPO A-I is a protein that, in humans, is encoded by the APOA1 gene, has a specific role in lipid metabolism, and is the major component of HDL in plasma.26,27 From tissues to the liver for excretion, APO A-I promotes fat efflux, including cholesterol. APO A-I could be isolated as a prostacyclin-stabilizing factor and has an anticlotting effect. Defects in APOA1 are associated with systemic non-neuropathic amyloidosis and with HDL deficiencies, including Tangier disease.28 APOF encodes the APO F protein that is one of the minor APOs found in plasma and forms complexes with lipoproteins.29 The APO F protein may be involved in the transport and/or esterification of cholesterol.
Based on the whole annotation as a reference set, we utilized the Cytoscape plugin, Bingo, to research the genes in module 2 using a hyper geometric test, Benjamin & Hochberg False Discovery Rate (FDR) correction, and the selected significance level (P < 0.05). We annotated the genes from cellular components, molecular function, and biological process of GO. In the biology process, APOA2 and APOA1 interacted with each other, involving the negative regulation of very-low-density lipoprotein particle remodeling and cytokine secretion involved in immune response, HDL particle assembly, and regulation, such as intestinal cholesterol absorption. APOA2, APOA1, and APOF take part in the cholesterol metabolic process, localization, and in the sterol metabolic process. In molecular function, APOA2, APOA1 and APOF, play a role in cholesterol binding, sterol binding, and lipid transporter activity. All of the genes in the module are located in the extracellular region.
In module 4, there are five genes: AP1S1, AP1G2, AP1M2, AFTPH, and AP1G1. AP1S1 is the different expression gene in SCLC, AP1G1 is the different expression gene in NSCLC, and AFTPH is the common gene shared by SCLC and NSCLC.
The AP-1 complex subunit sigma-1A is encoded by the AP1S1 gene30,31 and is part of the clathrin coat assembly complex linking clathrin to receptors in coated vesicles, which are involved in endocytosis and Golgi processing. AP-1 complex subunit sigma-1A, as well as beta-prime-adaptin, gamma-adaptin, and the medium (mu) chain AP47, form the AP-1 assembly protein complex located at the Golgi vesicle. The mutation of AP1S1 causes the rare familial MEDNIK syndrome – a genetic disorder.32 The proteins of AP1G2 and AP1G1 belong to the adaptor complex's large subunit family. AP1G2 encodes AP-1 complex subunit gamma-like 2, and AP1G1 encodes the AP-1 complex subunit gamma-1. These proteins act as the functions at a trafficking step in the complex pathways between the trans-Golgi network and the cell surface.33 AP-1 complex subunit mu-2 is encoded by the AP1M2 gene,34 which belongs to the adaptor complex's medium subunit family and interacts with tyrosine-based sorting signals. Aftiphilin is encoded by the AFTPH gene35 and forms a stable complex with p200 and synergin gamma, which contains a clathrin box with two identified clathrin-binding motifs and is involved in vesicle-trafficking. The protein is found in many eukaryotes. The commonly expressed genes show that all of the genes in the module may also cause NSCLC and SCLC, simultaneously. The genes in the module are annotated in 27 GO terms in biological processes, including protein transport, establishment of protein localization, and macromolecule localization, and 43 GO terms in cellular components, including clathrin adaptor complex, AP-type membrane coat adaptor complex, and clathrin coat.
In module 12, there are two genes: APIP and (APAF)1, and the APAF1 is significantly expressed in SCLC. APAF1 is a human homolog of the C. elegans CED-4 gene, and encodes a cytoplasmic protein that is the central hub in the apoptosis regulatory network.36 APAF1 expression is lower in lung cancerous tissues and is involved in the p53 signaling pathway, apoptosis, and some disease pathways, including SCLC and Huntington's disease. The protein encoded by the APIP gene is an enzyme with methylthioribulose 1-phosphate dehydratase activity37 that is involved in the methionine salvage pathway and interacts with the APAF1 encoding protein, which acts as a negative regulator of ischemic/hypoxic injury. APIP deficiency will induce cell death and cancer, including NSCLC and cystic fibrosis. Related super-pathways are metabolic, S-methyl-5-thio-alpha-D-ribose 1-phosphate degradation I, and the methionine salvage pathways, which play a significant role in microbial proliferation, cancer, apoptosis, and inflammation. According to the GO annotations related to APIP, we acquired the significant biological process, including identical protein binding and methylthioribulose 1-phosphate dehydratase activity. APIP is an APAF1-interacting protein that acts as a negative regulator of ischemic/hypoxic injury, and both of these genes could be the potential biomarkers to identify SCLC from NSCLC. The proteins in a same module involve the same or similar biological functions and the interaction among them promotes the occurrence of lung cancer.
Conclusion
With the proliferation of environmental pollution and the influence of human lifestyle, lung cancer incidence has increased year by year and has become an inevitable problem, a serious threat to human health. The different mechanisms of cancer occurrence decide the diagnosis scheme. The treatment for SCLC patients is chemotherapy and surgery is not effective for these patients. Surgical treatment is mainly suitable for patients with NSCLC. Relying only upon clinical pathological tissue to diagnose the type of lung cancer causes false positives, errors, and physical and mental trauma to patients. Thus, it is critical to research the occurrence mechanism of NSCLC and SCLC.
A system and integrated view is an effective way to clarify the mechanism of lung cancer and, in our research, we utilized bioinformatics methods to explore the difference between NSCLC and SCLC according to differentially expressed genes, the protein interaction network, and function enrichment. The results reveal that the different expressed genes for SCLC and NSCLC have similar topological features in the whole human protein interaction, indicating that the two cancers share a common interaction tendency with other proteins and a similar biological function. We, therefore, believe that it is critical to research the occurrence mechanisms of SCLC and NSCLC based on a module approach.
Acknowledgments
This research is supported by a grant from Tianjin Health Bureu (No. 2011KZ116, no. 2010KZ106), Tianjin Natural Science Foundation (No. 10JCYBJC23700), National Natural Science Foundation (No. 81370183).
Disclosure
No authors report any conflict of interest.
References
- Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61:69–90. doi: 10.3322/caac.20107. . (Published erratum appears in CA Cancer J Clin 2011; 134) [DOI] [PubMed] [Google Scholar]
- National Cancer Institute/National Institutes of Health. 2013. Cancer Staging. [Cited 25 Jan 2014.] Available from URL: http://www.cancer.gov/cancertopics/factsheet/detection/staging.
- National Cancer Institute/National Institutes of Health. 2013. General Information About Small Cell Lung Cancer (Patient Version). [Cited 25 Jan 2014.] Available from URL: http://www.cancer.gov/cancertopics/pdq/treatment/small-cell-lung/Patient.
- National Cancer Institute/National Institutes of Health. 2013. General Information About Small Cell Lung Cancer (Health Professional Version). [Cited 21 Feb 2014.] Available from URL: http://www.cancer.gov/cancertopics/pdq/treatment/small-cell-lung/healthprofessional.
- Medline Plus, National Institutes of Health. 2011. Lung cancer – non-small cell. [Cited 25 Jan 2014.] Available from URL: http://www.nlm.nih.gov/medlineplus/ency/article/007194.htm.
- American Cancer Society. 2013. Non-small cell lung cancer survival rates by stage. [Cited 26 Jan 2014.] Available from URL: http://www.cancer.org/cancer/lungcancer-non-smallcell/detailedguide/non-small-cell-lung-cancer-survival-rates.
- Cancer Reseach UK. 2014. About chemotherapy for lung cancer. [Cited 27 Jan 2014.] Available from URL: http://www.cancerresearchuk.org/cancer-help/type/lung-cancer/treatment/chemotherapy/about-chemotherapy-for-lung-cancer.
- University of Rochester Medical Center. 2014. Health Encyclopedia: Lung Cancer Introduction. [Cited 28 Jan 2014.] Available from URL: http://www.urmc.rochester.edu/encyclopedia/content.aspx?ContentTypeID=34&ContentID=16333-16331.
- Schiller JH, Gandara DR, Goss GD, Vokes EE. Non-small-cell lung cancer: then and now. J Clin Oncol. 2013;31:981–983. doi: 10.1200/JCO.2012.47.5772. [DOI] [PubMed] [Google Scholar]
- Stampfer MR, Yaswen P. Human epithelial cell immortalization as a step in carcinogenesis. Cancer Lett. 2003;194:199–208. doi: 10.1016/s0304-3835(02)00707-3. [DOI] [PubMed] [Google Scholar]
- Wachi S, Yoneda K, Wu R. Interactome-transcriptome analysis reveals the high centrality of genes differentially expressed in lung cancer tissues. Bioinformatics. 2005;21:4205–4208. doi: 10.1093/bioinformatics/bti688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhou HJ, Liu YK, Li Z, Yun D, Shun QL, Guo K. Analysing protein–protein interaction networks of human liver cancer cell lines with diverse metastasis potential. J Cancer Res Clin Oncol. 2007;133:663–672. doi: 10.1007/s00432-007-0218-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen Q. 2009. Research on molecular mechanism of non-small cell lung cancer (NSCLC) based on bioinformatics. CNKI.
- Ostlund G, Lindskog M, Sonnhammer EL. Network-based identification of novel cancer genes. Mol Cell Proteomics. 2010;9:648–655. doi: 10.1074/mcp.M900227-MCP200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karpinets TV, Park BH, Uberbacher EC. Analyzing large biological datasets with association networks. Nucleic Acids Res. 2012;40:e131. doi: 10.1093/nar/gks403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Doncheva NT, Assenov Y, Domingues FS, Albrecht M. Topological analysis and interactive visualization of biological networks and protein structures. Nat Protoc. 2012;7:670–685. doi: 10.1038/nprot.2012.004. [DOI] [PubMed] [Google Scholar]
- Sharan R, Ulitsky I, Shamir R. Network-based prediction of protein function. Mol Syst Biol. 2007;3:88. doi: 10.1038/msb4100129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barabási AL, Oltvai ZN. Network biology: understanding the cell's functional organization. Nat Rev Genet. 2004;5:101–113. doi: 10.1038/nrg1272. [DOI] [PubMed] [Google Scholar]
- Mao G, Zhang N. Analysis of average shortest-path length of scale-free network. Journal of Applied Mathematics. 2013;2013:865643. [Google Scholar]
- Newman MEJ. A measure of betweenness centrality based on random walks. Soc Networks. 2005;27:39–54. [Google Scholar]
- Okamoto K, Chen W, Li X-Y. Ranking of closeness centrality for large-scale social networks. Frontiers in Algorithmics. 2008;5059:186–195. [Google Scholar]
- Zhou T, Yan G, Wang BH. Maximal planar networks with large clustering coefficient and power-law degree distribution. Phys Rev. 2005;71:046141. doi: 10.1103/PhysRevE.71.046141. [DOI] [PubMed] [Google Scholar]
- Xu J, Li Y. Discovering disease-genes by topological features in human protein–protein interaction network. Bioinformatics. 2006;22:2800–2805. doi: 10.1093/bioinformatics/btl467. [DOI] [PubMed] [Google Scholar]
- Breslow JL, Ross D, McPherson J, et al. Isolation and characterization of cDNA clones for human apolipoprotein A-I. Proc Natl Acad Sci U S A. 1982;79:6861–6865. doi: 10.1073/pnas.79.22.6861. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Welch CL, Bretschger S, Wen PZ, et al. Novel QTLs for HDL levels identified in mice by controlling for Apoa2 allelic effects: confirmation of a chromosome 6 locus in a congenic strain. Physiol Genomics. 2004;17:48–59. doi: 10.1152/physiolgenomics.00124.2003. [DOI] [PubMed] [Google Scholar]
- Tsao YK, Wei CF, Robberson DL, Gotto AM, Jr, Chan L. Isolation and characterization of the human apolipoprotein A-II gene. Electron microscopic analysis of RNA:DNA hybrids, nucleotide sequence, identification of a polymorphic MspI site, and general structural organization of apolipoprotein genes. J Biol Chem. 1985;260:15222–15231. [PubMed] [Google Scholar]
- Arinami T, Hirano T, Kobayashi K, Yamanouchi Y, Hamaguchi H. Assignment of the apolipoprotein A-I gene to 11q23 based on RFLP in a case with a partial deletion of chromosome 11, del(11)(q23.3—qter) Hum Genet. 1990;85:39–40. doi: 10.1007/BF00276323. [DOI] [PubMed] [Google Scholar]
- Wasan KM, Brocks DR, Lee SD, Sachs-Barrable K, Thornton SJ. Impact of lipoproteins on the biological activity and disposition of hydrophobic drugs: implications for drug discovery. Nat Rev Drug Discov. 2008;7:84–99. doi: 10.1038/nrd2353. [DOI] [PubMed] [Google Scholar]
- Day JR, Albers JJ, Gilbert TL, Whitmore TE, McConathy WJ, Wolfbauer G. Purification and molecular cloning of human apolipoprotein F. Biochem Biophys Res Commun. 1994;203:1146–1151. doi: 10.1006/bbrc.1994.2302. [DOI] [PubMed] [Google Scholar]
- Peyrard M, Parveneh S, Lagercrantz S, et al. Cloning, expression pattern, and chromosomal assignment to 16q23 of the human gamma-adaptin gene (ADTG) Genomics. 1998;50:275–280. doi: 10.1006/geno.1998.5289. [DOI] [PubMed] [Google Scholar]
- Takatsu H, Sakurai M, Shin HW, Murakami K, Nakayama K. Identification and characterization of novel clathrin adaptor-related proteins. J Biol Chem. 1998;273:24693–24700. doi: 10.1074/jbc.273.38.24693. [DOI] [PubMed] [Google Scholar]
- Montpetit A, Côté S, Brustein E, et al. Disruption of AP1S1, causing a novel neurocutaneous syndrome, perturbs development of the skin and spinal cord. PLoS Genet. 2008;4:e1000296. doi: 10.1371/journal.pgen.1000296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lewin DA, Sheff D, Ooi CE, et al. Cloning, expression, and localization of a novel gamma-adaptin-like molecule. FEBS Lett. 1998;435:263–268. doi: 10.1016/s0014-5793(98)01083-7. [DOI] [PubMed] [Google Scholar]
- Ohno H, Tomemori T, Nakatsu F, et al. Mu1B, a novel adaptor medium chain expressed in polarized epithelial cells. FEBS Lett. 1999;449:215–220. doi: 10.1016/s0014-5793(99)00432-9. [DOI] [PubMed] [Google Scholar]
- Mattera R, Ritter B, Sidhu SS, McPherson PS, Bonifacino JS. Definition of the consensus motif recognized by gamma-adaptin ear domains. J Biol Chem. 2004;279:8018–8028. doi: 10.1074/jbc.M311873200. [DOI] [PubMed] [Google Scholar]
- Zou H, Henzel WJ, Liu X, Lutschg A, Wang X. Apaf-1, a human protein homologous to C. elegans CED-4, participates in cytochrome c-dependent activation of caspase-3. Cell. 1997;90:405–413. doi: 10.1016/s0092-8674(00)80501-2. [DOI] [PubMed] [Google Scholar]
- Mary C, Duek P, Salleron L, et al. Functional identification of APIP as human mtnB, a key enzyme in the methionine salvage pathway. PLoS ONE. 2012;7:e52877. doi: 10.1371/journal.pone.0052877. [DOI] [PMC free article] [PubMed] [Google Scholar]