Abstract
Spinal cord injury (SCI) leads to reactive inflammation and other harmful events that limit spinal cord regeneration. We propose an approach for studying the mechanisms at the levels of network topology, gene ontology, signaling pathways, and disease inference. We treated inflammatory mediators as toxic chemicals and retrieved the genes and interacting proteins associated with them via a set of biological medical databases and software. We identified >10,000 genes associated with SCI. Tumor necrosis factor (TNF) had the highest scores, and the top 30 were adopted as core data. In the core interacting protein network, TNF and other top 10 nodes were the major hubs. The core members were involved in cellular responses and metabolic processes, as components of the extracellular space and regions, in protein-binding and receptor-binding functions, as well as in the TNF signaling pathway. In addition, both seizures and SCI were highly associated with TNF levels; therefore, for achieving a better curative effect on SCI, TNF and other major hubs should be targeted together according to the theory of network intervention, rather than a single target such as TNF alone. Furthermore, certain drugs used to treat epilepsy could be used to treat SCI as adjuvants.
Similar content being viewed by others
Introduction
Spinal cord injury (SCI) is followed by a complex cascade of inflammatory events, such as reactive astrocytosis, which upregulates many genes1,2 and forms a glial scar3,4,5. This typically results in a permanent loss of neurological function below the injury level3,4. Considerable efforts have been made in research involving SCI-induced inflammatory cytokines, such as neurotoxic reactive astrocytes induced by secreting IL-1α, TNF, and C1q that leads to activation of microglia2, and those represent potential therapeutic targets6,7,8,9,10; however, it is largely unclear how the genes associated with SCI (GAS) interact and which among them play key roles. We consider that a network-based integration and bioinformatic analyses of the data available for genes/proteins associated with disease will reveal possible mechanisms for assessing the effects of GAS on SCI, leading to greater clarity. As is known11, biochemical events driven by electrostatic forces and involving hydrophobic effects are the physical contacts with high specificity through which the gene or protein interaction network is established. In a network of biomolecules, the nodes (vertices) indicate genes or proteins, whereas the links (edges) indicate their physical (direct) or functional (indirect) interactions12. Network analysis has demonstrated an efficient approach for modeling biological systems12,13,14. For example, it could reveal the molecular mechanisms of cancer13,14,15,16,17 and infer the diseases associated with environmental chemicals18,19. During the last decade, with the popularization of RNA sequencing (RNA-seq) technologies and the development of bioinformatics analyses, a wealth of data for constructing a biomolecule network has been available from public databases/resources, such as Online Mendelian Inheritance in Man (OMIM)20, Kyoto Encyclopedia of Genes and Genomes (KEGG)21, Search Tool for the Retrieval of Interacting Genes/Proteins (STRING)22,23, and Comparative Toxicogenomics Database (CTD)24,25,26. OMIM is an updated catalog of human genes and genetic disorders and traits based on selection and review of the published peer-reviewed biomedical literature and has become one of the databases of the National Center for Biotechnology Information (NCBI)27; KEGG is a collection of databases with genomes, diseases, drugs, and chemicals, featuring the products of relevant pathways in biology; STRING includes experimental data, computational prediction methods and public text collections in biology, and links to numerous sources, and is able to generate known and predicted protein–protein interactions (PPIs); and CTD curates data on chemical–gene/protein interactions, chemical–disease and gene–disease relationships from selected literature sources in a structured format, and controlled vocabularies and inference scores, and integrates these data with those from NCBI, OMIM, KEGG, and 8 other databases, and links all the interactions to the original publications to enable users to access the source data for specific details about corresponding experiments. The inflammatory cytokines associated with SCI are neurotoxins2, and the toxicogenomics approaches can be used to identify them. TNF was first discovered in 1968 as a cytotoxic factor induced by lymphocytes and was referred to as a lymphotoxin (LT)28; therefore, it might be favorable for us to use the CTD database for collecting the SCI-gene data, and the CTD in-house scoring system for screening that data. There are thousands of curated genes associated with SCI in CTD that are available for the required bioinformatics analyses26, such as in this study about the effects of GAS on SCI.
In an effort to assess the effects of GAS on SCI, we designed a network-based integration and bioinformatics analysis approach, incorporating the disease-gene toxicogenomics26, PPI networks22,23, and gene ontology (GO) enrichment analysis29,30,31,32 and disease inference26. First, GAS were retrieved from CTD. Subsequently, the protein interactions involved in GASs were integrated from STRING database22,23, and visualized via Cytoscape33,34,35, a popular, open source bioinformatics software platform for network analysis. Finally, by using the interacting proteins, the functions and pathways associated with SCI were inferred. As a result, the most important as well as the top 30 interacting proteins were singled out; affected functions and pathways were identified; and diseases, including neurological and psychological disorders, were predicted, which provided better insight into the influence of GAS on SCI and related diseases. This analysis approach is also expected to be useful for studying neurotrophic factors and nerve growth factors involved in SCI and its consequences. Using biological data in system-level to study disease-gene associations is able to improve our current knowledge of disease relationships, leading to further improvements in disease diagnosis, prognosis and treatment.
Material and Methods
Genes/proteins associated with SCI (GAS)
GAS data were obtained from CTD26 by searching for genes involved in SCI, resulting in a list of 12,824 GAS or their protein products, which were then sorted by the CTD in-house “inference score” in descending order. The top 30 GAS (GAS30) with high scores (47.84-33.7) were taken as the core data in this study (i.e., in this context, the GAS30 represented the genes that were most closely associated with SCI). Then, Cytoscape (version3.4.0, 2016)34 and STRING (version 10.5)23 were conducted to query the protein-protein interactions of the GAS30. STRING is as an application (App, plugin) installed in Cytoscape. The data in STRING are weighted and integrated and a confidence score is calculated for all protein interactions according to the nature and quality of the supporting evidence. As a result, each of these interactions is assigned a confidence score between zero (no interaction) and one (high-confidence interaction), which indicates the probability that the interaction is authentic, given the available evidence. The default cutoff for confidence interactions is 0.418,19. This study utilized this default value to screen PPIs and only the interactions whose confidence scores were >0.4 were considered for network analysis. Of the established PPI network of the GAS30, all nodes were from CTD and with CTD in-house inference scores of >33, and all edges were from STRING and with STRING in-house confidence scores of >0.4. Furthermore, the plugin NetworkAnalyzer34 in Cytoscape was used to visualize molecular interaction networks and integration with gene expression profiles and other state data.
Enrichment analysis of gene ontology, pathway, and disease
Gene ontology (GO)31,32, a controlled vocabulary describing gene products and a useful resource for studying gene functions36, consists of three domains termed cellular components (CC), molecular functions (MFs), and biological processes (BP). Identifying enriched GO terms from a given gene list enables insight into the over-represented functions of genes29. Enrichment analysis of pathways and diseases is also an approach to the further understanding of the implicated pathways and diseases associated with SCI. Several web services such as the BinGO37 plugin of Cytoscape, OmicsBean38 and Set Analyzer26 of CTD can be employed for studying enriched GO terms, pathways and diseases, respectively. Among these services, the pathway-gene relationships for enrichment analysis are from the KEGG21 and REACTOME39 pathway databases, whereas the MEDIC disease vocabulary24 that combines the Medical Subject Headings (MeSH)40 and OMIM20 databases is used for analysis of enriched diseases. Briefly, we input the gene list of GAS30 respectively into BinGO or OmicsBean for GO term analysis; OmicsBean for pathways analysis; and the set analyzer of CTD for diseases analysis, while using a p-value of <0.05.
Results
PPI network analysis
A total of 12,824 genes were identified as associated with SCI using CTD as of December 12, 2017. Among these, the top 30 genes (GAS30) with CTD inference scores >33 are listed in Table 1. After inputting the GAS30 gene list into Cytoscape, assigning a link to STRING, and assuming that the interactions between the molecules were nondirectional and with interacting confidence scores of >0.4 in STRING, we obtained a PPI network for GAS30. Figure 1A shows a GAS30 PPI network that consists of 30 nodes and 232 edges. A node represents a molecule and an edge represents an interaction between two connected nodes. These two nodes are called neighbors. The fact that no edge is connected to a node such as myeloperoxidase (MPO) indicates that interactions between this node and others do not exist (at least their interacting confidence scores were ≤0.4) and it should be deleted from this network. The number of edges/neighbors related to a node is referred to as the degree of the node35,41,42. A node with a number of edges that greatly exceeds the average is referred to as hub and these play crucial roles in the network43. Therefore, molecules in the GAS30 network could be re-sorted by their node degrees. By use of NetworkAnalyzer34, a Cytoscape plugin for network topology analysis, the degrees of each node in the GAS30 network were calculated and the top 10 are listed in Table 2. Among these, the TNF node exhibited the greatest degree, and is termed a major hub. Furthermore, the top 10 interacting molecules become a sub-network of the GAS30 cohort, and are denoted as GAS10 and shown in Fig. 1B. The sub-network GAS10 comprised 10 nodes and 44 edges, leading to an extremely high clustering coefficient41 of 0.978 and an extremely small diameter44 of 2. That is, the 10 nodes were all major hubs and highly interconnected.
Gene ontology (GO) analysis
GO term enrichment analyses of GAS30 proteins were conducted for biological processes (BPs), cellular components (CCs), and MFs using BinGO37 and OmicsBean38. Considering p < 0.05, there were 3,396, 210, and 251 terms for BPs, CCs, and MFs, respectively, enriched in GAS30. The top 10 terms for BPs, MFs, and CCs are listed in Table 3 and shown in Fig. 2.
The enriched BP analysis revealed that GAS30 could interfere with cellular responses and metabolic processes. Specifically, the process of responding to lipopolysaccharide involved 57% of GAS30 members (Fig. 2) and was promoted to the highest GO level in a significant p-value (Table 3). Lipopolysaccharide is a cell wall component of gram-negative bacteria, and is a type of endotoxin45 that is released only when bacterial cells are destroyed or when using an artificial method to kill the microorganisms. Considerable evidence has revealed the influence of lipopolysaccharides on central nervous system (CNS) diseases. For example, lipopolysaccharides can cause learning and memory disorders in rats subsequent to CNS inflammatory responses46,47, which positively supports the outcome of our GO term enrichment analysis for BPs.
The enriched GO terms for CCs of the interacting proteins were mostly related to the extracellular space components, in which the first two (i.e., extracellular space and extracellular region components), exhibited the most significant p-values in CCs (Table 3) and accounted for 50% and 63% of GAS30 members (Fig. 2), respectively. A hierarchical GO tree for CCs enriched in GAS30 is presented in Fig. 3.
The MFs influenced by the interacting GAS30 proteins were mostly related to the protein-binding and receptor-binding functions, according to the enriched GO terms. Notably, protein binding accounted for the highest percentage (93%) in GAS30 in all enriched GO terms as shown in Fig. 2. Figure 4 shows a hierarchical tree of important GO terms for MFs affected by the interacting GAS30 proteins.
Pathway enrichment analysis
To further reveal the pathways affected by interacting GAS30 proteins, analyses were performed using OmicsBean38, a web service for processing biological data and with links to KEGG21 and other public databases. Following the instructions of OmicsBean, Table 4 and Fig. 5 were generated. The top 10 with the most significant p-values are listed in Table 4 and shown in Fig. 5. Specifically, the TNF signaling pathway was ranked at the top of the list, which accounted for 47% of GAS30 members.
Nervous system diseases involving TNF inferred from CTD
The diseases associated with TNF were inferred using the toxicogenomics analyses of CTD, which yielded 570 nervous system diseases associated with TNF. The top 30 diseases sorted by CTD in-house inference scores are listed in Table 5. Notably, among these, seizures had the highest score (No. 1 in Table 5).
Discussion
We considered SCI-induced inflammatory mediators as a type of toxin that inhibits the regeneration of injured tissue/cells, and we identified the associated genes and interacting proteins from known biological medical databases, CTD26, STRING23, and others, and chose the top 30 genes/proteins, GAS30, that were useful for studying their effects on SCI at the levels of network topology, GO, signaling pathways, and disease inference to provide a new visual angle for finding potential methods by which SCI intervenes.
According to this study, more than 10,000 genes associated with SCI and TNF achieved the highest score. TNF, FOS, IL6, and seven other of the top 10 nodes (Table 2) were the major hubs and highly interconnected in the GAS30 PPI network that were identified using CTD, STRING, and related databases. From the perspective of network topology48, such a network allows for a fault-tolerant behavior for which, if a hub-failure occurs, the network will generally not lose its connectedness because of the remaining hubs that will rapidly replace the failing hub. This suggests that although TNF negatively affected SCI repair, all other major hubs, such as FOS IL6, should be targeted simultaneously in the future for the development of new therapeutic approaches, rather than aiming at individual specific genes, one at a time, which might achieve better curative effects.
Furthermore, GAS30 members interfered mainly with cellular responses and metabolic processes, extracellular space and extracellular region components, protein-binding and receptor-binding functions, and TNF signaling pathways as identified by GO and pathway enrichment analyses. Notably, the TNF signaling pathways were promoted to the highest enriched level of GAS30 members and had the most significant p-value (Table 4). Although considerable evidence has revealed the influence of TNF as an inducer of inflammatory cytokines after SCI7,8,9 (e.g., neurotoxic reactive astrocytes induced by activation of microglia through secreting Il-1α, TNF, and C1q2), greater attention should be paid to TNF in the future and consider it to be a major signaling pathway and its use as a crucial and potential therapeutic target for SCI repair.
In addition, seizures were highly associated with TNF by CTD disease inference (Table 5). Clinically, seizures might occur after traumatic brain injury, and interestingly, repeated seizures might develop into post-traumatic epilepsy49,50,51. Seizures were also observed following SCIs52,53. More interestingly, the antiepileptic drug valproate was used as a supplement in stem cell transplantation for a mouse model of SCI, which dramatically enhanced the restoration of hindlimb function54. These suggest that certain drugs used to treat epilepsy could be employed as adjuvants in SCI treatment; however, these observations and suggestions were not directly linked to TNF by the original researchers, and the mechanisms proposed are unclear. Therefore, TNF, which is the most important hub identified in this study, could be further connected to the aforementioned findings and would be a direction for future SCI studies.
There are other aspects of this study that must be mentioned. In addition to CTD, the genes/proteins associated with SCI or other diseases that were searched to construct the PPI networks and subsequent bioinformatics analyses could be from OMIM20,55 or other publicly available databases56, which would generate similar results because nearly all public databases are interconnected to the Internet; therefore, the primary data mostly overlap with each other. As such, CTD includes OMIM and 10 other databases26. We used CTD in this study for its full name (Comparative Toxicogenomics Database) as well as its functions that matched our requirements for treating inflammatory cytokines in SCI as a type of neurotoxin. In recent years, disease genome sequencing and other high-throughput studies of disease genomes have generated many notable discoveries17. Direct data on disease-genes are commonly derived from RNA-seq because it is superior to other high-throughput technologies, such as microarray in accuracy, dynamic range, and differential expression detection, and has nearly completely replaced microarray for conducting genetic tests. The entries curated in OMIM have referenced57,58,59,60 the results from RNA-seq, and NCBI27 online accepts RNA-seq data and shares it with other databases and researchers. In addition, the data in references2,5,8,9,13,14,15,16,17 in this study were primarily from RNA-seq. Furthermore, a combination61 of using RNA-seq approach with PPI network analysis generated that TNF had the largest number of connected edges in the PPI network for contusive SCI in a mouse model, and the top ranked genes in the SCI gene list overlapped considerable with ours, supporting our current study; however, our present method hardly describes the dynamic effects of GAS on SCI. On the other hand, notably, because of its pleiotropic role, TNF shows, for example, a positive effect on regulatory T cells62 and prevents neurons from death/apoptosis by activating NF-κB63,64; therefore, suppressing TNF overexpression might not be a desirable intervention for SCI therapy, and this needs to be observed further.
Furthermore, to more effectively predict the SCI drug targets, the patient-specific signaling networks for reactive inflammation from SCI could be constructed using the concept of “SCI hallmarks” based on individual genomic data and on regulatory functions, just as the signaling networks of “cancer hallmarks”13,14,15,16,17 have been developed and substantially used for revealing molecular mechanisms of cancers and drug targets. This proposed approach to analyses is also expected to be useful for studying neurotrophic factors and nerve growth factors after SCIs.
With a constantly expanding repertoire of techniques, including RNA-seq, together with new information on genes and proteins, the current results will have more possibilities for examination and modification and will advance the current approaches to SCI analysis.
Data Availability
All data generated or analyzed during this study are included in this published article.
References
Zamanian, J. L. et al. Genomic analysis of reactive astrogliosis. J. Neurosci. 32, 6391–6410 (2012).
Liddelow, S. A. et al. Neurotoxic reactive astrocytes are induced by activated microglia. Nature 541, 481–487 (2017).
Fawcett, J. W. & Asher, R. A. The glial scar and central nervous system repair. Brain Res. Bull. 49, 77–391 (1999).
Silver, J. & Miller, J. H. Regeneration beyond the glial scar. Nat. Rev. Neurosci. 5, 146–156 (2004).
Anderson, M. A. et al. Astrocyte scar formation aids central nervous system axon regeneration. Nature 532, 195–200 (2016).
Alizadeh, A., Dyck, S. M. & Kataria, H. Neuregulin-1 positively modulates glial response and improves neurological recovery following traumatic spinal cord injury. GLIA 65, 1152–1175 (2017).
Wu, X. et al. RhoA/Rho Kinase mediates neuronal death through regulating cPLA2 activation. Mol. Neurobiol. 54, 6885–6895 (2017).
Liu, G. et al. FK506 attenuates the inflammation in rat spinal cord injury by inhibiting the activation of NF-kappa B in microglia cells. Cell Mol. Neurobiol. 37, 843–855 (2017).
Shao, A. et al. Hydrogen-rich saline attenuated subarachnoid hemorrhage-induced early brain injury in rats by suppressing inflammatory response: possible involvement of NF-kappa B pathway and NLRP3 inflammasome. Mol. Neurobiol. 53, 3462–3476 (2016).
Zhong, L. et al. Quantitative proteomics study of the neuroprotective effects of B12 on hydrogen peroxide-induced apoptosis in SH-SY5Y cells. Sci. Rep.-UK 6, 22635 (2016).
D L Rivas, J. & Fontanillo, C. Protein-protein interactions essentials: key concepts to building and analyzing interactome networks. PLoSComput. Biol. 6, e1000807 (2010).
Barabási, A. L. & Oltvai, Z. N. Network biology: understanding the cell’s functional organization. Nat. Rev. Genet. 5, 101–113 (2004).
Han, P., Gopalakrishnan, C., Yu, H. & Wang, E. Gene regulatory network rewiring in the immune cells associated with cancer. Genes 2017(8), 308, https://doi.org/10.3390/genes8110308 (2017).
McGee, S. R., Tibiche, C., Mark Trifiro, M. & Wang, E. Network analysis reveals a signaling regulatory loop in PIK3CA-mutated breast cancer predicting survival outcome. Genomics Proteomics & Bioinformatics 15, 121–129 (2017).
Wang, E. et al. Predictive genomics: A cancer hallmark network framework for predicting tumor clinical phenotypes using genome sequencing data. Semin. Cancer Biol. 30, 4–12 (2015).
Zaman, N. et al. Signaling network assessment of mutations and copy number variations predict breast cancer subtype-specific drug targets. Cell Rep. 5, 216–223 (2013).
Wang, E. Understanding genomic alterations in cancer genomes using an integrative network approach. Cancer Lett. 340, 261–269 (2013).
Singh, S. & Li, S. S. Phthalates: toxicogenomics and inferred human diseases. Genomics 97, 148–157 (2011).
Lin, Y. C., Wang, C. C. & Tung, C. W. An in silicotoxicogenomics approach for inferring potential diseases associated with maleic acid. Chem. Biol. Interact. 223, 38–44 (2014).
Amberger, J. S., Bocchini, C. A., Schiettecatte, F., Scott, A. F. & Hamosh, A. OMIM.org: Online Mendelian Inheritance in Man (OMIM®), an online catalog of human genes and genetic disorders. Nucleic Acids Res. 43(Database issue), D789–798 (2015).
Kanehisa, M. et al. KEGG for linking genomes to life and the environment. Nucleic Acids Res. 36, D480–484 (2008).
Szklarczyk, D. et al. STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 43(Database issue), D447–452 (2015).
STRING database homepage, https://string-db.org/ (2017).
Davis, A. P., Wiegers, T. C., Rosenstein, M. C. & Mattingly, C. J. MEDIC: a practical disease vocabulary used at the Comparative Toxicogenomics Database. Database 2012, 1–9 (2012). Article ID bar065.
Davis, A. P. et al. The comparative toxicogenomics database: update 2013. Nucleic Acids Res. 41, D1104–1114 (2013).
Comparative Toxicogenomics Database homepage, http://ctdbase.org/ (2017).
National Center for Biotechnology Information homepage. https://www.ncbi.nlm.nih.gov (2018).
Kolb, W. P. & Granger, G. A. Lymphocyte in vitro cytotoxicity: characterization of human lymphotoxin. Proc. Natl. Acad. Sci. USA 61(4), 1250–1255 (1968).
Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. PNAS 102, 15545–15550 (2005).
Pathan, M. et al. Funrich: An open access standalone functional enrichment and interaction network analysis tool. Proteomics 15, 2597–601 (2015).
Ashburner, M. et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. 25, 25–29 (2000).
Gene ontology homepage, http://www.geneontology.org/
Shannon, P., Markiel, A. & Ozier, O. Cytoscape: A Software environment for integrated models of biomolecular interaction networks. Genome Res. 13, 2498–2504 (2003).
Cytoscape homepage. http://www.cytoscape.org/ (2017).
Doncheva, N. T., Assenov, Y., Domingues, F. S. & Albrecht, M. Topological analysis and interactive visualization of biological networks and protein structures. Nat. Protoc. 7, 670–685 (2012).
Tung, C. W. PupDB: a database of pupylated proteins. BMC Bioinformatics 13, 40 (2012).
Maere, S., Heymans, K. & Kuiper, M. BiNGO: a Cytoscape plugin to assess over representation of Gene Ontology categories in biological networks. Bioinformatics 21, (3448–3449 (2005).
Omicsbean Platform, http://www.omicsbean.com:88/ (2017).
Croft, D. et al. Reactome: a database of reactions, pathways and biological processes. Nucleic Acids Res. 39, D691–697 (2011).
Coletti, M. H. & Bleich, H. L. Medical subject headings used to search the biomedical literature. J. Am. Med. Inform. Assoc. 8, 317–323 (2001).
Albert, R. & Barabási, A. L. Statistical mechanics of complex networks. Rev. Mod. Phys. 74, 47–97 (2002).
Assenov, Y., Ramirez, F., Schelhorn, S. E., Lengauer, T. & Albrecht, M. Computing topological parameters of biological networks. Bioinformatics 24, 282–284 (2008).
Zotenko, E., Mestre, J., O’Leary, D. P. & Przytycka, T. M. Why do hubs in the yeast protein interaction network tend to be essential: reexamining the connection between the network topology and essentiality. Plos Comput. Biol. 4, e1000140 (2008).
Cohen, R. & Havlin, S. Scale-free networks are ultrasmall. Phys. Rev. Lett. 90, 058701 (2003).
Kulp, A. & Kuehn, M. J. Biological functions and biogenesis of secreted bacterial outer membrane vesicles. Annu. Rev. Microbiol. 64, 163–184 (2010).
Zarifkar, A. et al. Agmatine prevents LPS-induced spatial memory impairment and hippocampal apoptosis. Eur. J. Pharmacol. 634, 84–88 (2010).
Hauss-Wegrzyniak, B., Lynch, M. A., Vraniak, P. D. & Wenk, G. L. Chronic brain inflammation results in cell loss in the entorhinal cortex and impaired LTP in perforant path-granule cell synapses. Exp. Neurol. 176, 336–341 (2002).
Zoraghi, R. & Reiner, N. E. Protein interaction networks as starting points to identify novel antimicrobial drug targets. Curr.Opin.Microbiol. 16, 566–572 (2013).
Pagni, C. A. & Zenga., F. Posttraumatic epilepsy with special emphasis on prophylaxis and prevention. ActaNeurochir.Supplementum 93, 27–34 (2005).
Garga, N. & Lowenstein, D. H. Posttraumatic Epilepsy: A Major problem in desperate need of major advances. Epilepsy Curr. 6, 1–5 (2006).
Pitkänen, A., Kharatishvili, I. & Karhunen, H. Epileptogenesis in experimental models. Epilepsia 48(Supplement 2), 13–20 (2007).
Clancy, R. R., Sladky, J. T. & Rorke, L. B. Hypoxic-ischemic spinal cord injury following perinatal asphyxia. Ann. Neurol. 25, 185–189 (1989).
De Louw, A. J. A. et al. Apoptosis in the rat spinal cord during postnatal development; the effect of perinatal asphyxia on programmed cell death. Neuroscience 112, 751–758 (2002).
Abematsu, M. et al. Neurons derived from transplanted neural stem cells restore disrupted neuronal circuitry in a mouse model of spinal cord injury. J. Clin. Invest. 120, 3255–3266 (2010).
Amberger, J., Bocchini, C. & Hamosh, A. A new face and new challenges for Online Mendelian Inheritance in Man (OMIM®). Hum. Mutat. 32, 564–567 (2011).
Sun, K., Goncalves, J. P., Larminie, C. & Pržulj, N. Predicting disease associations via biological network analysis. BMC Bioinformatics 15, 304, http://www.biomedcentral.com/1471-2105/15/304 (2014).
Zhang, L. et al. Microenvironment-induced PTEN loss by exosomal microRNA primes brain metastasis outgrowth. Nature 527, 100–104 (2015).
Zhao, D. et al. Synthetic essentiality of chromatin remodelling factor CHD1 in PTEN-deficient cancer. Nature 542, 484–488 (2017).
Kuchay, S. et al. PTEN counteracts FBXL2 to promote IP3R3- and Ca(2+)-mediated apoptosis limiting tumour growth. Nature 546, 554–558 (2017).
Poliseno, L. et al. A coding-independent function of gene and pseudogene mRNAs regulates tumour biology. Nature 465, 1033–1038 (2010).
Chen, K. et al. RNA-Seq characterization of spinal cord injury transcriptome in acute/subacute phases: a resource for understanding the pathology at the systems level. Plos One 8, e72567, https://doi.org/10.1371/journal.pone.0072567 (2013).
Salomon, B. L. et al. Tumor Necrosis Factor α and Regulatory T Cells in Oncoimmunology. Front. Immunol. 9, 444, https://doi.org/10.3389/fimmu.2018.00444 (2018).
Barger, S. W. et al. Tumor necrosis factors alpha and beta protect neurons against amyloid beta-peptide toxicity: evidence for involvement of a kappa B-binding factor and attenuation of peroxide and Ca2+ accumulation. Proc. Natl. Acad. Sci. USA 92, 9328–9332 (1995).
Mattson, M. P. & Meffert, M. K. Roles for NF-κB in nerve cell survival, plasticity, and disease. Cell Death Differ. 13, 852–860 (2006).
Acknowledgements
This work was supported by the National Natural Sciences Foundation of China Grant No. 31370940.
Author information
Authors and Affiliations
Contributions
Conceptualization: W.Z., X.C. Formal analysis: W.Z., X.C. Investigation: W.Z., X.C., L.N., K.J. Methodology: W.Z., X.C., K.J. Project administration: W.Z. Writing – original draft: W.Z., X.C., L.N., K.J.
Corresponding author
Ethics declarations
Competing Interests
The authors declare no competing interests.
Additional information
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Zhu, W., Chen, X., Ning, L. et al. Network Analysis Reveals TNF as a Major Hub of Reactive Inflammation Following Spinal Cord Injury. Sci Rep 9, 928 (2019). https://doi.org/10.1038/s41598-018-37357-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-018-37357-1
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.