Schizophrenia (SZ) is a chronic and devastating mental illness that affects around 20 million individuals worldwide. Cognitive deficits and structural and functional changes of the brain, abnormalities of brain ECM components, chronic neuroinflammation, and devastating clinical manifestation during SZ are likely etiological factors shown by affected individuals. However, the pathophysiological events associated with multiple regulatory pathways involved in the brain of this complex disorder are still unclear. This study aimed to develop a pipeline based on bioinformatics and machine learning approaches for identifying potential therapeutic targets involving possible biological mechanisms from SZ patients and healthy volunteers. 420 overlapping DEGs from three RNA-seq datasets were identified. GO, and pathways analysis showed several biological mechanisms enriched by the commonly shared DEGs, including ECM organization, collagen fibril organization, integrin signaling pathway, inflammation mediated by chemokines and cytokines signaling pathway, and GABA-B receptor II and IL4 mediated signaling. 15 hub genes (FN1, COL1A1, COL3A1, COL1A2, COL5A1, COL2A1, COL6A2, COL6A3, MMP2, THBS1, DCN, LUM, HLA-A, HLA-C, and FBN1) were discovered by comprehensive analysis, which was mainly involved in the ECM organization and inflammatory signaling pathway. Furthermore, the miRNA target of the hub genes was analyzed with the random-forest-based approach software miRTarBase. In addition, the transcriptional factors and protein kinases regulating overlapping DEGs in SZ, namely, SUZ12, EZH2, TRIM28, TP53, EGR1, CSNK2A1, GSK3B, CDK1, and MAPK14, were also identified. The results point to a new understanding that the hub genes (fibronectin 1, collagen, matrix metalloproteinase-2, and lumican) in the ECM organization and inflammatory signaling pathways may be involved in the SZ occurrence and pathogenesis.
Research Article
FN1, Collagen, MMP-2, and Lumican Associated with ECM Organization with Multiple Signaling Pathways Dysfunction in Schizophrenia Pathogenesis by Network Biology Approaches
https://doi.org/10.21203/rs.3.rs-2070137/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Schizophrenia (SZ)
Bioinformatics and machine learning
Extracellular matrix organization (EMC)
inflammatory signaling pathway
SZ affects around 2.5% of the global population and is one of the top 10 causes of disability globally1. SZ is a neurodevelopmental condition that appears early but is compensated for until the average age of clinical onset in young adulthood2, 3. The growing prevalence of SZ illness is a severe global problem for the scientific community to overcome. The scientific community has undertaken efforts in recent decades to develop effective therapeutics for the devastating pathophysiology of SZ. In genome-wide association and gene-expression profiling, investigations into the impacts of SZ pathogenesis observed that numerous genes connected to the immune system, cytoskeletal development, and neuroinflammation had been implicated4, 5. However, the precise pathological mechanisms that promote SZ progression are often ambiguous and poorly understood. The influence of environmental, genetic, neurodevelopmental, immunological, and neuroinflammatory factors on the onset and course of SZ has led to several theories to explain the pathophysiological mechanisms underpinning SZ6–8.
The neuroinflammatory process has been linked to SZ-associated genes that have a variety of functions in the brain9. Accumulating recent evidences reported that the cytotoxic consequences of chronic microglial activation might promote the progression of SZ and enhance secondary neuronal degeneration, reduced neurogenesis, and synaptic dysfunction10–12. A growing body of research suggests that microglia play a novel role in regulating the ECM remodeling during normal brain homeostasis, and these activities may, in turn, become dysfunctional in SZ13, 14. The ECM and interstitial environment are essential regulators of neuroinflammation and the immune response, just as they are for glial cells-mediated scar formation15. Damage-associated molecular patterns (DAMPs) are structural components of the ECM and products of its degradation that stimulate or inhibit microglial reactivity by signaling through pattern recognition receptors (e.g., Toll-like receptors) 13, 16, 17. For example, in vitro culturing of microglia on a CSPG substrate induces microglial activation, proliferation, and expression of IGF-1, MMP-2, and MMP-9, whereas pharmacological inhibition of CSPG production with xyloside after spinal cord injury alters neuroinflammation and cytokine production differently depending on treatment timing18. It has been reported that repeated social defeat caused an increase in biomarkers for EMC remodeling and blood-brain barrier (BBB) leakage, thus resulting in enhanced BBB permeability associated with microglial phagocytosis19, 20.
MMPs are a family of structurally similar proteolytic enzymes that remodeling of the ECM to keep synaptic functioning and BBB intact21. MMP-mediated ECM dysfunction plays a crucial role in SZ pathophysiology22. Dysregulation of MMPs by neuroinflammation causes ECM abnormalities, which change neuronal processes such as synaptic plasticity and BBB disruption directly or indirectly22, 23. Growth factors, cytokines including TNF-α, and IL-1β, chemical agents, physical stress, and, most crucially, cell-to-cell or cell-to-ECM interaction are all known to trigger MMP genes24. This study suggested that TNF-α, and IL-1β induce transcription of MMP-3 and MMP-9, which may be involved in neuroinflammation. MMP-9 can also be activated by other proteases (MMP-3) and free radicals (nitric oxide, which works by N-nitrosylation) 25. MMPs have a role in the degradation of CSPGs, a key component of ECM26. Decreased levels of CSPGs have been discovered in the brains of SZ patients27. Alterations complicate the pathophysiology of SZ in MMP expression caused by glial cells22. MMP hydrolyzes ECM, which is involved in neural control, neurotransmitter signaling, and synaptic plasticity28. In the schizophrenic brain, upregulation of MMPs and an imbalance between MMP and TIMP are linked to various ECM abnormalities22. MMP expression can be upregulated by oxidative stress and neuroinflammation, leading to tissue degradation, neuronal death, and ECM abnormalities in SZ29, 30.
This is why well-established bioinformatics methods and analytical approaches were employed to investigate the therapeutic targets as well as molecular signaling pathways that could play a critical role in the SZ pathogenesis. Firstly, we identified shared overlapping DEGs using RNA sequencing data from SZ patients and healthy controls. Secondly, GO and functional enrichment pathway analyses of shared overlapping genes has investigated to uncover the significant biological mechanisms and signaling pathways. Then, the PPI network was used to identify the hub genes which are impacting the pathophysiology of SZ patients. Moreover, this work was predicted transcriptional factors and PKs that may be involved in the SZ pathogenesis. In addition, the miRNA-gene regulatory network of the selected hub genes was predicted by miRNA databases. Finally, this study suggested that the hub genes (FN1, Collagen, MMP-2, and Lumican) dysregulation linked with EMC organization and inflammatory signaling pathways may be involved in the SZ occurrence and pathogenesis.
Literature search and datasets collection
In order to collect the RNA sequencing and transcriptomes gene expression raw data samples, this work was used to search two public repositories GEO RNA-seq database (NCBI Gene Expression Omnibus and GREIN). The keyword “Schizophrenia” extracted three RNA-seq datasets with accession numbers GSE92874, GSE63738, and GSE121376. In order to acquire RNA-seq datasets, the criteria were used to follow inclusion: (1) Original experiments studies were conducted for screening RNA expression levels between control and schizophrenia, (2) the RNA-seq datasets were identified only for the organism “Homo sapiens” (3) RNA-seq raw data samples were deposited in the NCBI GEO for publicly accessible use, (4) high throughput sequencing Illumina 2500 and 6000 platforms for extracting RNA-seq, and (5) RNA-seq datasets submitted in the year range between 2007–2021. The search excluded datasets obtained from the healthy control and SZ patients according to earlier criteria.
Rna-seq Data Processing And Identification Of Common Degs
The GREIN31 web application was used to identify DEGs from three RNA-seq datasets. GREIN employed a negative binomial generalized linear model executed in edgeR to find DEGs between control and SZ patients. Datasets were normalized using a trimmed mean of M-values. GREIN used Bioconductor package RNA-seq power of the following formula given below:
Where α represents target false positive rate, β determines target false-negative rate or 1-β is power, n is the sample data size, Δ shows the effect size, µ represents average sequencing depth, and σ defines BCOV measured as the square root of the dispersion. Furthermore, GREIN used the FDR correction formula for adjustment to correct DEGs with multiple testing and performed the following formula:
Where α* represents the significance level of sample sizes, r1 determines the anticipated number of true positives within m1 rejected null hypotheses, m0 indicates the number of genes when null hypotheses are true, and f defines the desired FDR level.
In order to determine deregulated DEGs, algorithm methods employed Log2 transformation using the LIMMA package in R with the Benjamini-Hochberg method to control the rate of false discovery of multiple correction methods32. Then, the statistically significant DEGs were filtered from three RNA-seq datasets using the cutoff criteria as mathematically employed below:
Here, DEGi is estimating the ith DEG either it determines upregulated or downregulated p-values based on the criteria mentioned in the above equation. Furthermore, using Bioinformatics & Evolutionary Genomics Venn diagram web application tool was identified common significant DEGs from all three RNA-seq datasets (GSE92874, GSE63738, and GSE121376).
Go And Signaling Pathway Enrichment Analyses Of Common Degs
GO and pathway enrichment studies were used to determine the biological characteristics of mutual DEGs in SZ. GO enrichment analysis was performed on FunRich bioinformatics software33 as integrated through 1.5 million gene annotations based on genomics and proteomics databases. Different statistically significant results for BP, MF, and CC were identified with a p-value < 0.01.
Then, to determine the molecular signaling pathways functional enrichment analysis, Fisher’s exact test employed the combined DEGs using EnrichR34 (A comprehensive gene enrichment analysis web-based tool). Then, using different datasets suggested putative signaling pathways and gene annotation ontologies related to common DEGs. The overrepresentation analysis determines the biological significance of the often expressed genes concerning the participation of cell signaling pathways in SZ. Four databases were utilized in this computational approach to investigate biological signaling pathways of commonly shared DEGs in SZ, namely: Reactome 2016, BioPlanet 2019, Panther 2016, and NCI Nature 2016. We chose the top-listed pathways depending on the standard criteria of a P-value of 0.01.
Construction Of Protein-protein Interaction Network And Screening Of Hub Genes
To further determine the functional interactions between these common DEGs, the protein interaction network was mapped using experimentally validated interactome database STRING (Version 11.5) and a set of parameters criteria that included active interaction sources, text mining, experiments, databases, co-expression, neighbourhood, gene fusion, and co-occurrence as well as functional and physical protein associations. Studies were also limited to “Homo sapiens” with a high confidence score of > 0.7. Then, using the Cytoscape software (Version 3.9.0)35, import a tab-separated values (TSV) file to visualize a protein-protein interaction network, where the number of nodes denotes proteins, and the edges indicate their interactions. Based on degree methods with a high interconnected score, the Cytohubba plugin of the Cytoscape program was used to identify the top 15 hub genes. Using Cytoscape software was also able to identify the hub genes interaction network.
Identification Of A Transcription Factor Linked To The Protein Network
Transcription factors (TFs) play an essential role in several biological pathways by acting in the extensive protein complexes network generated by a protein-protein interaction network that initiates and controls genetic code transcription. Based on the hypergeometric p-value, the X2K is a web-based algorithm bioinformatics tool36 (regulatory networks platform) from the ChIP-seq experiments (ChEA) database to determine the top transcriptional factors. The X2K online tool generates inferred transcription factor networks linked to PPI, resulting in upstream regulatory pathways based on commonly shared DEG signatures. The Genes2Networks (G2N) algorithm37 was used to identify transcriptional factors by identifying proteins that interact physically with these transcription factors. G2N is a powerful command-line and web-based program that interprets DEGs based on experimentally established protein-protein interactions network or protein complexes in genomic and proteomic data. This tool aids scientists in filtering transcription factors with protein network complex connections to better understand the cell signaling cascade.
KEA determines the relationship between kinases, proteins, and transcription factors.
PTKs are enzymes that turn on the cell of phosphorylated targeted proteins to control signaling proteins dynamically. The KEA module of X2K36 has been used to identify protein kinase. KEA38 is a command-line program that can match shared DEGs lists of mammalian proteins to the PTKs expected to phosphorylate them. The upstream signaling network depended on a regulatory kinase-substrate interaction within the extended subnetwork incorporating protein kinases, protein-protein interactions, and transcription factors with phosphorylation using the HPRD, PhosphoSite, phospho.ELM, NetworKIN, and Kinexus databases.
Analysis Of Microrna-hub Genes Regulatory Network
The microRNA-hub gene interaction network identified the regulatory molecules which could be implicated in SZ pathogenesis. Based on miRTarBase database plugging, the miRNA-hub gene regulatory networks were revealed using the Network Analyst software, and then Cytoscape constructed the miRNA-hub genes network. Then, the determination of miRNA-hub genes enrichment analysis depended on the adjusted p-value of 0.5 and interaction threshold 3 using MIENTURNET39 (interactive web tool). Further, the miRTarBase database determined the predicted miRNA targets with set an inclusion threshold criteria minimum ≥ 5 interactions and an adjusted p-value of 0.5.
Identification Of Potential Biological Signaling Pathways
To identify the potential biological pathway connected with common DEGs, the FunRich interactive bioinformatics tool33 (http://www.funrich.org) was employed against the human genomic and proteomic database as integrated with the FunRich background.
Utilizing Hub Genes To Infer The Network Biology Of Emc Architecture
Network biology approaches identified scored-based hub genes and EMC organization-associated interaction through data integration and quality control with several-fold more interactions (> 500,000) network biology approach. The InWeb_InBioMap databases40 was utilized with a relevance score cutoff of 0.85 to design hub genes associated with EMC organization and integrin cell surface interactions network pathway. Finally, hub genes are also screened by sharing EMC organization, and integrin cell surface interactions networks link using these same databases.
Design Emc-receptor Interaction Complex Pathway
ShinyGO v0.74141 was used to create a diagram of the EMC-receptor interaction based on KEGG pathway databases. A threshold of P-value of 0.05 and the organism Homo sapiens were used to evaluate the significance level of KEGG pathway enrichment for drawing this diagram.
Identification of overlapping common DEGs
This study collected the gene expression profiles of three RNA-seq datasets (GSE92874, GSE63738, and GSE121376) from the NCBI GEO database to identify and investigate differentially expressed gene expression patterns that may impact SZ development. Firstly, the DEGs were obtained by analyzing three RNA-seq datasets using the GREIN interactive web platform. DEGs were observed in a total of, 25213, 22214 and 21849 from the datasets of GSE92874, GSE63738, and GSE121376 respectively. Then, a total of 3880 (2453 upregulated and 1427 downregulated genes), 1953 (1080 upregulated and 873 downregulated genes), and 5787 (2922 upregulated and 2865 downregulated genes) deregulated DEGs were identified from the GSE92874, GSE63738, and GSE121376 respectively based on the criteria of p-value < 0.5 and log2FC ≥ 1 or ≤ − 1 (supplementary data 1). Figure 1(A, C, E) illustrates an interactive heatmap of the top 100 upregulated and downregulated genes. Figure 1 (B, D, F) shows all the DEGs with a log2FC versus the – log10 (p-value) between the SZ patients and healthy groups in a MA plot, and positive Log2FC represents upregulated genes, whereas negative Log2FC indicates downregulated genes. Four hundred twenty overlapping DEGs were identified using Venn diagrams among, GSE92874, GSE63738, and GSE121376datasets (Figure F), including 290 upregulated genes and 130 downregulated genes. Then, using bioinformatics and machine learning tools were employed to analyze the functional annotation of genes, enrichment analysis of interaction networks, and molecular and biological pathways using those commonly deregulated DEGs.
Gene Ontological Analysis Of Overlapping Degs
The GO is a comprehensive conceptual model for describing the functional characteristics of DEGs. GO analyzed the functional annotation of overlapping DEGs into three functional categories: BP, MF, and CC. In the aspect of BP, the overlapping DEGs were enriched in cell communication, signal transduction, cell growth and maintenance, regulation of immune responses, cell-matrix adhesion, immune response, peptide metabolism, regulation of the biological process, cell differentiation, cell surface receptor linked signal transduction, regulation of gene expression, epigenetic, calcium-mediated signaling, transport, cell adhesion and transcription (Fig. 2A). In terms of MF (Fig. 2B), the commonly shared DEGs were mainly involved in EMC structural constitute, MHC receptor class I receptor activity, MHC receptor class II receptor activity, transcription factor activity, cell adhesion molecule activity, metallopeptidase activity, receptor binding, motor activity, growth factor activity, amino acid transporter activity, GTPase activator activity, ATPase activity, and calcium ion binding. For the CC category, overlapping expressed genes were significantly enriched in extracellular, EMC, extracellular space, plasma membrane, nucleus, cytoplasm, fibrinogen complex, integral to the plasma membrane, proteinaceous EMC, collagen type I, collagen type V, integrin complex, microfibril, TAP complex, inclusion body, intermediate filament, and basement membrane (Fig. 2C). These findings suggest that EMC-related processes are essential in the SZ development mechanisms.
Identification Of Significant Signaling Pathway Enrichment Analysis
Functional enrichment pathways analysis determines how complicated the underlying biological processes cause SZ pathogenesis. Using web-based bioinformatics tools like EnrichR, including four pathways databases such as Reactome, Panther, BioPlanets, and NCI nature, performed gene set enrichment analysis of overlapping DEGs to find critical signaling pathways that may relate to SZ pathogenesis (Fig. 3). The top 15 signaling pathways were chosen based on the significance of the adjusted p-value of less than 0.1 as a criterion for pathway analysis. Critical biological pathways were identified which may be implicated in SZ using the Reactome 2016 pathway database, including EMC organization, collagen biosynthesis and modifying enzymes, assembly of collagen fibrils and other multimeric structures, collagen formation, elastic fibre formation, integrin cell surface interactions, muscle contraction, antigen processing-cross presentation, transmission across chemical synapses, endosomal/Vacuolar pathway, neuronal System, ER-phagosome pathway, neurotransmitter receptor binding, ECM proteoglycans, and fibronectin matrix formation (Fig. 3A). Panther 2016 pathway revealed integrin signaling pathway, ionotropic glutamate receptor pathway, nicotinic acetylcholine receptor signaling pathway, 5-Hydroxytryptamine degradation, Alzheimer disease-presenilin pathway, TGF-beta signaling pathway, cadherin signaling pathway, cytoskeletal regulation by Rho GTPase, inflammation mediated by chemokine and cytokine signaling pathway, p53 pathway T cell activation, dopamine receptor-mediated signaling pathway, adrenaline and noradrenaline biosynthesis, GABA-B receptor II signaling, and JAK/STAT signaling pathway (Fig. 3B). Results from the BioPlanet 2019 pathway database identified the most significant biological function associated with signaling pathways, including beta-1 integrin cell surface interactions, TGF-beta regulation of EMC, collagen biosynthesis and modifying enzymes, EMC organization, syndecan-1 pathway, ECM-receptor interaction, integrins in angiogenesis, antigen processing and presentation, beta-3 integrin cell surface interactions, interleukin-1 regulation of EMC, neural crest differentiation, inflammatory response pathway, integrin cell surface interactions, integrin beta-5 pathway, and neuronal system (Fig. 3C). The pathway analysis of the NCI 2016 nature database revealed beta1 integrin cell surface interactions, syndecan-1-mediated signaling events, integrins in angiogenesis, beta3 integrin cell surface interactions, beta5 beta6 beta7 and beta8 integrin cell surface interactions, VEGFR3 signaling in lymphatic endothelium, Plexin-D1 Signaling, Direct p53 effectors, syndecan-4-mediated signaling events, IL4-mediated signaling, BMP receptor signaling, ALK1 signaling events, integrin family cell surface interactions, Arf6 trafficking events, and syndecan-2-mediated signaling events (Fig. 3D). We showed pathway enrichment analysis along with gene name, adjusted p value and combined score (supplementary data 2). These pathway enrichment analysis findings suggest that EMC and integrin cell surface contacts may have played a role in the progression of SZ.
Protein-protein Interaction Network And Hub Protein Identification
To investigate a correlation between overlapping DEGs and SZ disorder that focused on the PPIs network. The PPIs network of DEGs was generated using the STRING database (version 11.5), with combined scores more significant than 0.7 (interaction score: high confidence), and visualized using Cytoscape software, with 183 nodes and 384 edges (Fig. 4A). Proteins are represented as nodes in the PPI network, connected by undirected edges, implying a relationship between two proteins. Furthermore, the PPIs network is employed for hub protein discovery, which may aid in the identification of a therapeutic biomarker for the disease comorbidities.
To analyze hub protein across the whole PPIs network, using MNC and MCC approaches employed the Cytoscape plug-in CytoHubba, which was generated Hub proteins network. The PPI networks selected the top 15 hub proteins using the techniques above. The therapeutic targets were determined SZ pathogenesis and progression by identifying the top 15 hub proteins (FN1, COL1A1, COL3A1, COL1A2, COL5A1, COL2A1, COL6A2, COL6A3, MMP2, THBS1, DCN, LUM, HLA-A, HLA-C, and FBN1), which have significantly higher degree interactions (Fig. 4B). We demonstrated the highest rank score of hub genes (supplementary data 3). The found hub proteins are potential biomarkers that might lead to new SZ therapeutic targets and play critical roles in EMC construction during SZ development.
Analysis Of Transcription Factors
Transcription factors (TFs) are essential molecules that regulate gene expression and maintain gene regulatory networks. DEGs are controlled by the activation or repression of transcription factors (TFs), which are required for various biological and cellular processes and have been linked to the development of neurological disorders. The X2K online tool identified transcriptional factors for controlling overlapping DEGs expression in SZ using the ChEA database. Transcription factor enrichment analysis (TFEA) was performed based on the hypergeometric p-value and selected the top 20 TF candidates (Fig. 5A). SUZ12, EZH2, IRF8, TP63, TRIM28, TP53, FOSL2, EGR1, PPARD, KLF4, NANOG, POU5F1, REST, TCF3, SALL4, SOX2, GATA1, and GATA2.The Genes2Networks (G2N) method is employed to find proteins that physically interact with PPIs and TFs to investigate their relationships. Based on the degree of the nodes, the regulatory network of linked TFs and their functionally and physically interacting proteins was depicted (Fig. 5B).
Identification Of Protein Kinase With Upstream Regulatory Network
The mRNA translation, cell proliferation and survival regulation, and nuclear genomic response to cellular stressors are regulated by protein kinases, which comprise sensors and effectors from signal transduction cascades. Protein kinase activity dysregulation has been linked to various disorders, ranging from inflammatory to neurodevelopment disorders. Using the KEA module of X2K, the most critical PTKs linked with SZ were identified to examine possible neurotherapeutic kinase targets. The results of KEA revealed that CSNK2A1, GSK3B, CDK1, MAPK14, ATM, HIPK2, CDK2, DNAPK, MAPK1, MAPK8, CDK4, CK2ALPHA, MAPK3, ABL1, AKT1, ERK1, IKKALPHA, JNK1, PKR, ERK2, PKBALPHA are the most critical PTKs in intracellular signaling pathways related with SZ (Fig. 6A). Then, the human protein reference database (HPRD), PhosphoSite, phospho.ELM, NetworKIN, and Kinexus were used to build a kinase-substrate network. Finally, these databases constructed a regulatory kinase-substrate network in which active PTKs phosphorylate substrates inside the enlarged subnetwork of transcription factors and intermediary proteins (Fig. 6B).
Prediction Of Mirnas Targets With Hub Genes
Using network analyst, the miRNA-hub gene regulatory network associated with the development of SZ was constructed. According to the miRNA-hub gene-targeted network, the network included hsa-miR-29c-3p, hsa-miR-29b-3p, hsa-miR-29a-3p, hsa-miR-143-3p, hsa-miR-767-5p, hsa-miR-6825-5p, hsa-miR-29b-1-5p, hsa-let-7g-5p, hsa-miR-544a, hsa-miR-1331-3p which had targeted relationships with the hub genes (Fig. 7A). Finally, using MitarBase employed to identify the miRNA-target enrichment analysis result based on significant adjusted p-value and target hub genes interaction (Fig. 7B).
The FunRich software examined possible biological pathways related to SZ (Fig. 8). These findings revealed that SZ's major potential biological pathways had been compromised of beta3 integrin cell surface interactions, integrin cell surface interactions, integrin family cell surface interactions, beta1 integrin cell surface interactions, VEGFR3 signaling in lymphatic endothelium, dopamine and serotonin degradation pathway, transmission across chemical synapses, and neuronal system. According to the biological pathways, analysis results might be helpful for future studies to investigate the involvement the SZ pathogenesis.
Interconnection Of Hub Genes Associated With Emc Organization And Integrin Cell Surface Interactions Mechanism
Interactions were annotated as genetic interactions, co-localizations, or neighbouring reactions, and only extracted the small subset of data that described direct protein-protein interactions or data on experimentally resolved protein complexes from the pathway databases (NetPath, Reactome, and WikiPathways). Connectome analysis was observed a network of signaling molecules, receptors, plasma membrane-bound transcription factors, kinases, enzymes, and structural proteins interconnected with hub genes to determine the potential biological signal of EMC organization and integrin cell surface interactions network pathway. Figure 9A has shown that FBN1-“governs as a signaling molecule” interacts with MMP-2/13/8/12 that regulates structural proteins networks and also maintains plasma membrane stability. Similarly, MIA, itself also signaling molecule, is tightly bound with EMCs proteins such as FN1, THBS1, COL5A1, COL3A1, COL1A2, COL2A1, COL2A1, COL5A1, and COL6A3, thus may be modulated integrin activity. Then, the hub genes association with shared EMC organization and integrin cell surface interactions network pathway was determined. Our hub genes identification result described that EMC organization shared their interaction with integrin cell surface molecules as well as their signaling pathways (Fig. 9B). However, it needs to be confirmed and validated this computational analyses results by experimentally. Finally, hub genes used overlapping DEGs of SZ to create a schematic diagram of the EMC-receptor interaction complex pathway based on the KEGG database. In Fig. 9C, we tried to identify how EMC-associated genes interact with integrin receptor and proteoglycan molecules based on the KEGG pathway analysis by bioinformatics method. We observed that fibronectin, collagen, and THBS interact with integrin dimer (α1, α2, α3, α8, αV, αIIb, β1, and β3), thus resulting in maintaining their cell adhesion and signal strength.
Identified Targets Validation With Literature Review
According to GO and functional enrichment analysis, the commonly shared overlapping dysregulated genes were mainly implicated in EMC organization, signal transduction, collagen fibril organization, antigen processing and presentation, and ECM-receptor interaction process. These suggested results might be crucial in the progression of SZ. ECM molecules govern GABAergic function, synaptic pruning and plasticity during childhood and adolescence, neuronal migration, proliferation, differentiation, neurodevelopment modulation, neuroplasticity, axon guidance, and neurite outgrowth42. Previously, it has been reported that altered ECM organization has been associated with the pathophysiology of SZ and neurodegenerative diseases43, 44. A small peptide generated from a collagen protein may protect the brain from SZ by stimulating the formation of neural connections45. These studies suggested that abnormalities of collagen fibril organization may be linked to the pathogenesis of SZ. Evidence reported that signal transduction anomalies caused by changes in the kinase activity network are the foundation of SZ fundamental symptoms46. The results of functional enrichment analysis were consistent with earlier research. Pathway enrichment analysis was used to identify the top 15 pathways associated with the pathogenesis of SZ. Previously, it has been reported that SZ patient-derived cells are sensitive to ECM proteins that bind integrin receptors, thus resulting in increases in focal adhesion number and size in response to ECM proteins and alterations in cell shape and cytoskeleton47,. Cadherins have a critical role in developing brain circuitry and mature synaptic function48; hence genes encoding members of the cadherin superfamily play a vital role in the pathophysiology of neuropsychiatric illnesses. In the mammalian brain, the cadherin signaling system regulates adhesion molecules that are important in cell-cell contact49. The TGF-beta signaling pathway is crucial for the use-dependent control of GABAA synaptic transmission and dendritic homeostasis; moreover, a disruption in the excitatory-inhibitory balance in the hippocampal network may induce psychiatric-like behavior50. In human mental illnesses like SZ, the TGF -β signaling pathway components have been altered in the hippocampus51. Syndecan signaling from the ECM to various cytoplasmic components and their distribution to different membrane compartments rely on oligomerization; hence, the syndecan homodimer is a crucial functional unit52. However, future research needs to investigate the bidirectional link between Syndecan signaling and ECM organization in the pathophysiology of SZ. Mounting studies have reported that inflammatory signaling pathways have a significant role in depression and SZ, and chronic inflammation is linked to immunological dysregulation in SZ53–55.
FN1 mutations may increase SZ susceptibility, suggesting that this gene is an SZ-susceptible gene56. Moreover, in patients with SZ, FN1 gene expression levels involved complement pathway activators (C1qA) and mediators (C3 or C4) enhanced in the midbrain parenchyma near dopamine cell bodies, especially in individuals with a high inflammatory biotype57. This study also discovered that peripheral macrophages were significantly increased in the midbrain of SZ cases related to high inflammatory potential. However, it remains unclear how FN1 triggered inflammatory signaling pathways in the progression of SZ pathogenesis. Evidence reported that the collagen chain gene (COL1A1) expression might be associated with the EMC ligands, thus resulting in the inactivation of DDR1 with a time-dependent decrease in the SZ patients58. In addition, evidence reported that COL1A1 and COL1A2 genes expression levels were significantly decreased in patients with SZ compared to control59. One of the most increased genes, COL3A1, is detected, which encodes the pro-1 chains of type III collagen and exhibits a de novo mutation in SZ patients60. COL5A1 may be associated with genetic variations of psychotic experiences in SZ61. In distinct brain areas, the expression patterns of collagen family genes (COL1A1, COL3A1, COL1A2, COL5A1, COL2A1, COL6A2, COL6A3) encoding proteins that are fundamental components of ECM62. Chronic agonistic interactions may cause abnormal collagen genes expression, which might indicate specific ECM abnormalities in the brain areas of mice with different social experiences63. MMP-2 levels were significantly increased in the CSF of SZ patients, thus may be connected to neuroinflammation in the brain64. This study postulated that the pathophysiology of SZ seems to be influenced by state-dependent changes in MMP-2 and activation of MMP-2, -7, and − 10 cascades. A clinical study reported that THBS1 gene might have a critical role in the development of SZ patients65. The involvement of the THBS1 gene in SZ pathogenesis will need to be investigated further in the future.
This investigation revealed that transcriptional factors (including SUZ12, EZH2, TRIM28, TP53, and EGR1) are associated with the disease progression and development of SZ pathogenesis. SUZ12 binds a location 1.5 kb downstream of the GAD1 TSS, and the H3K27me3 mark is enhanced at the GAD1 promoter region in the prefrontal cortex of individuals with SZ, along with a significant number of the discovered DMRs within the GAD1 regulatory network66, 67. EZH2 may play a role in SZ risk at two different times: during development, when it causes neurodevelopmental abnormalities, and during adulthood, when it causes abnormal expression reactivation68. TRIM28 suppresses ERVs in the gene regulatory network that governs gene expression of protein-coding transcripts essential for brain development69. An essential link between TP53 and SZ was discovered by haplotype analysis70. These findings suggest that TP53 may have a role in SZ etiology. Consequently, transcriptional factors might be useful to develop potential therapeutic targets for SZ. Identification of PKs such as CSNK2A1, GSK3B, CDK1, and MAPK14 were reported to have a critical role in the pathogenesis of SZ71, 72.
This work integrated three RNA sequencing data from SZ patients using bioinformatics and machine learning algorithms to uncover possible therapeutic targets for SZ prevention. Firstly, this study identified 420 DEGs with statistically significant overlap in SZ, with 290 upregulated and 130 downregulated genes. The multi-step integrative bioinformatics methods used to look at the functional annotation of commonly shared DEGs, GO enrichment pathway analysis, interactomics analysis, upstream regulatory molecules, kinase enrichment analysis, miRNA-hub genes regulation network, biological signaling pathways analysis, as well as possible potential therapeutic targets.
Based on the pathway enrichment analysis results, overlapping genes were enriched in ECM remodeling, integrin signaling pathway, inflammation mediated by chemokines and cytokines signaling pathway, and GABA-B receptor II and IL4 mediated signaling. Numerous studies have reported that dysfunction of ECM remodeling contributes to pathological features of SZ, such as disruption of synapses connectivity, change of neuronal migration and axonal guidance, and abnormalities of neurotransmission, including GABAergic, glutamatergic, dopaminergic systems as well as neuroinflammatory process 73–75. Given that many cognitive symptoms of SZ are thought to be linked to a mismatch between inhibitory GABA and excitatory glutamate neurotransmission in the dorsolateral prefrontal cortex. SZ cognitive deficits may be partly caused by MCHTAP expression or function dysregulation that disrupts GABAergic/glutamatergic balance 76–78. ECM proteins are involved in neuronal cell migration, axonal growth, myelination, and the formation and maintenance of the neuromuscular junction at the BBB79. The ECM-integrin signaling pathway may regulate cell adhesion and signaling, endosomal trafficking, CSK dynamics, and gene expression at all stages of nervous system development, maintenance, degeneration, and regeneration 80, 81. Alterations of the GABAergic system have been implicated in the pathology of SZ 82. In SZ, the GluRl and GluR2 subunits of kainic acid/amino-3-hydroxy5-methyl-4-isoxazole propionic acid (KA/AMPA) glutamate receptors are attenuated in the hippocampus 83. Due to the loss of GABAergic hippocampal interneurons, GABAA receptors are increased, and GABA-uptake sites are diminished 84, 85. Furthermore, serotonergic 5-HT1A and 5-HT2 receptors are upregulated in the hippocampus in SZ, although 5-HT uptake sites remain constant86. Reduced GABA release and uptake at synaptic terminals, decreased expression of the GABA-synthesizing enzyme glutamic acid decarboxylase (GAD), altered expression of GABAA receptors, and a reduction in axon cartridges of GABAergic chandelier neurons have all been reported as anomalies of GABAergic interneurons87. Multiple GABA-related proteins, including GAD65, SYNPR, DBI, GAT3, SN1, and CPT1A, were dramatically downregulated in adolescence88. To maintain normal GABA homeostasis, these molecules participate in the GABA cycle, which includes GABA synthesis, release, reuptake, and replenishment. Dysregulation of the GABAergic system is mainly affected in the prefrontal cortex, thalamus, and hippocampus during childhood and adolescence 89.
Though complicated biological mechanisms but linked, hypotheses for SZ pathology include GABAergic system and EMC remodeling hypotheses are being proposed here. Abnormalities in the ECM would affect GABAergic synaptic transmission, which is essential in the pathophysiology of neuropsychiatric disorders, including SZ90. It has been reported that the enhanced and persistent GABA activity that balances the reduced density of parvalbumin neurons might be connected to changes in the ECM mediated by gene-environment interactions 91, 92. However, it remains unexplored how ECM-associated proteins, including FN1, Collagen, MMP-2, and Lumican, consequently affect GABAergic system function during childhood. Is an altered epigenetic regulation of gene expression the molecular mechanism mediating transcription controlling pathways, thus resulting in a significant impact on the downstream pathways of ECM remodeling and GABAergic system in SZ?
The bidirectional communication between the major histocompatibility complex (MHC) and the GABAergic system is hugely complicated during brain development and maintaining synaptic plasticity. Neurite outgrowth, synaptic transmission, the development and maintenance of cortical connections, activity-dependent refinement in the visual system, and long-term and homeostatic plasticity are all regulated by MHCI molecules93. The balance of excitation to inhibition (E/I balance) on cortical neurons is controlled by MHCI, which affects glutamatergic vs. GABAergic synapses differently94. Because of the MHCI molecule's capacity to influence glutamatergic and GABAergic synapse density in both directions, it is in a prime position to contribute to alterations in neural circuitry throughout brain development that might explain prodromal symptoms and enhance susceptibility to the second hit in SZ95. It has been demonstrated that MHCI regulates synapse density in the developing brain in an age-dependent manner96, 97. MHCI levels are affected by genetic variations, gene expression alterations, or immunological dysregulation, and as in SZ in humans, synapse density should also be affected 93, 98. Further studies need to be clarified how MHC and GABAergic systems interact after a chronic immune-dysregulated state in offspring, thus causing SZ.
In our study, MMP-2 identified as hub gene based on bioinformatics analysis which may be play a critical role in the pathogenesis of SZ. MMPs are zinc-dependent endopeptidases that play a key role in BBB integrity and brain disorders99. Recent study have reported that MMP-9 is linked to the language and fluency aspects of cognition, and it raises the likelihood of cognitive impairment in SZ100. MMP-9, an extracellular network protease involved in glutamatergic signaling pathway, may have a role in SZ pathogenesis101. MMP-9 is involved in the maturation of inhibitory neurons that contain the calcium-binding protein parvalbumin, the embryonic production of the specialized extracellular matrix structure perineuronal net, synaptic pruning, and myelination, all of which are considered to be impaired in SZ102. On the flip side, MMP-9 has a role in the brain development as well as ability to control synaptogenesis, axonal pathfinding and myelination103. Clinical study demonstrated that upregulated MMP-9 expression levels are directly linked with cognitive impairment in SZ patients104. It has been reported that overactivation of the MMP9/RAGE pathway causes redox dysregulation and neuroinflammation, which leads to an inhibitory/excitatory imbalance in SZ pathophysiology105. An increase levels of MMP-2 and MMP-9 have been observed in the pathophysiology of neuropsychiatric diseases106. The extracellular matrix is degraded by MMP proteins, which govern vascular remodeling107. Extracellular components such as collagen, gelatin, and laminin are cleaved by MMP-1108. MMP-1 can also activate other MMPs, including as MMP2 and MMP9, which together can lower the expression of tight junction proteins such TJP1, OCLN, and CLDN5109. MMP-2/-9 is capable of degrading tight junction proteins and basement membranes, two BBB components, directly, resulting in enhanced BBB permeability110. Microglial activation has been shown to degrade BBB by releasing MMP-2/-9. α--synuclein increased MMP-9 expression and gene transcripts in primary microglia from rats in a dose-dependent manner111. The link between MMP-2 and MMP-9 might also be assumed to have a potential biological mechanisms of SZ pathogenesis. Therefore, we can postulate that MMP-2 and MMP- 9 may have therapeutic importance in improvement of SZ disease progression and pathology.
However, there are some limitations and drawbacks of this present study. We utilized several bioinformatics methods and network biology approaches for functional analyses. Firstly, the bioinformatics methods and network biology approaches have certain limitations, such as the fact that it only considers the number of genes or DEGs value. In addition, an artificial threshold was required to get the genes of interest or differential expression. Furthermore, it often employs the most significant genes while ignoring others with no significant alterations, which may result in the loss of genes with lower significance but a more critical function, resulting in lowered detection sensitivity. Secondly, RNA-seq data were received from the GEO database, not conducted by the authors. Finally, the predicted results should be validated by experimental data.
The primary purpose of this study lies in exploring potential therapeutic targets in the progression of SZ pathogenesis based on bioinformatics and network biology approaches. The gene ontology, biological signaling pathways, protein interatomic networks, transcription, upstream regulatory pathway protein kinases, miRNA-hub genes network, and the most critical hub genes could be therapeutically targeted for SZ pathogenesis using RNA-seq data were analyzed. Eventually, the essential hub genes like fibronectin 1, collagen, matrix metalloproteinase-2, and lumican were identified. Then their suitability for biological targets to regulate SZ progression was then checked, and potential hub genes for therapeutic intervention were also predicted. Finally, the gold benchmark databases, including dbGap, OMIM, and OMIM, were used to validate the role of the therapeutic targets and molecular pathways in the development of SZ disease.
SZ = Schizophrenia
ECM = Extracellular matrix
CAMs = Cell adhesion molecules
CSPGs =Chondroitin sulfate proteoglycans
MMPs = Matrix metalloproteases
NMDA = N-methyl-D-aspartate
GWAS = Genome-wide association study
ASDs = Autism spectrum disorders
DEGs = Differentially expressed genes
PPI = Protein–protein interaction
NCBI = National Center for Biotechnology Information
GREIN = GEO RNA-seq Experiments Interactive Navigator
BCOV = Biological coefficient of variation
FDR = False discovery rate
BP = Biological process
MF = Molecular function
CC = Cellular component
TSV = Tab-separated values
TFs = Transcription factors
ChEA = ChIP-seq experiments
PTKs = Protein kinases
KEA = Kinase enrichment analysis
HPRD = Human protein reference database
KEGG = Kyoto Encyclopedia of Genes and Genomes
GO = Gene ontology
MHC = Major histocompatibility complex
TAP = Transporter associated with antigen processing
Rho = Ras homologous
GABA = Gamma-amino butyric Acid
JAK/STAT =Janus kinase/signal transducers and activators of transcription
TGF = Transforming growth factor
BMPs = Bone morphogenetic proteins
FN1 = fibronectin 1
COL1A1 = collagen type I alpha 1 chain
COL3A1 = collagen type III alpha 1 chain
COL1A2 = collagen type I alpha 2 chain
COL5A1 = collagen type V alpha 1 chain
BMP4 = bone morphogenetic protein 4
MMP2 = matrix metallopeptidase 2
CDH1 = cadherin 1
POSTN = periostin
THBS1 = thrombospondin 1
COL6A2 = collagen type VI alpha 2 chain
DCN = decorin
FBN1 = fibrillin 1
COL2A1 = collagen type II alpha 1 chain
LUM = lumican
SUZ12 = SUZ12 Polycomb Repressive Complex 2 Subunit
EZH2 = Enhancer of Zeste 2 Polycomb Repressive Complex 2 Subunit
IRF8 = Interferon Regulatory Factor 8
TP63 = Tumor Protein P63
TRIM28 = Tripartite Motif Containing 28
TP53 = Tumor protein P53
FOSL2 = FOS Like 2, AP-1 Transcription Factor Subunit
EGR1= Early Growth Response 1
PPARD = Peroxisome Proliferator Activated Receptor Delta
KLF4 = Kruppel Like Factor 4
NANOG = Nanog Homeobox
POU5F1 = POU Class 5 Homeobox 1
REST = RE1-silencing transcription factor
TCF3 = Transcription Factor 3
SALL4 = Spalt Like Transcription Factor 4
SOX2 = SRY-Box Transcription Factor 2
GATA1 = GATA Binding Protein 1
GATA2 = GATA Binding Protein 2
CSNK2A1 = Casein kinase 2 alpha 1
GSK3B = Glycogen synthase kinase 3 beta
CDK1= Cyclin dependent kinase 1
MAPK14 = Mitogen-activated protein kinase 14
HIPK2 = Homeodomain interacting protein kinase 2
CDK2 = Cyclin dependent kinase 2
CDK4 = Cyclin dependent kinase 4
MAPK3 = Mitogen-activated protein kinase 3
ABL1 = ABL proto-oncogene 1, non-receptor tyrosine kinase
PKR= Pyruvate kinase L/R
ERK2= Mitogen-activated protein kinase 1
CNS= Central nervous system
Gad1 = Glutamate Decarboxylase 1
Dbi = Diazepam binding inhibitor
MFS = Marfan syndrome
GAD1 = Glutamate Decarboxylase 1
H3K27me3 = Trimethylation of histone H3 lysine 27
ERVs= Endogenous retroviruses
BBB = Blood-brain-barrier
Authors’ contributions
PB, ZS and YQ were designed this study and also involved in the acquisition, analysis, and interpretation of the data. PB, and AK were involved in the literature search, data processing, software, data interpretation, and data analysis. PB wrote the draft and edited manuscript. YQ and ZS supervised the whole study and were involved in the study design, acquisition, data analysis and editing the draft manuscript.
Funding
This study was supported by the grants from the National Natural Science Foundation of China (No 81671387).
Availability of data and materials
The Gene Expression Omnibus (GEO) database from NCBI (https://www.ncbi.nlm.nih.gov/geo/) was used to access the GSE92874, GSE63738, and GSE121376 datasets.
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare that there is no conflict of interests regarding this study.
Acknowledgements
All authors would like to thank NCBI GEO and GREIN team for freely accessible RNA-seq datasets.
- Fleischhacker, W. W., Arango, C., Arteel, P., Barnes, T. R., Carpenter, W., Duckworth, K., Galderisi, S., Halpern, L., Knapp, M., Marder, S. R., Moller, M., Sartorius, N., & 628 Woodruff, P. (2014).Schizophrenia--time to commit to policy change. Schizophrenia 629 bulletin, 40 Suppl 3(Suppl 3), S165–S194.
- Owen, M. J., & O'Donovan, M. C. (2017). Schizophrenia and the neurodevelopmental continuum: evidence from genomics. World psychiatry: official journal of the World Psychiatric Association (WPA), 16(3), 227–235.
- Owen, M. J., O'Donovan, M. C., Thapar, A., & Craddock, N. (2011). Neurodevelopmental hypothesis of schizophrenia. The British journal of psychiatry: the journal of mental science, 198(3), 173–175.
- Radhakrishnan, R., Kaser, M., & Guloksuz, S. (2017). The Link between the Immune System, Environment, and Psychosis. Schizophrenia bulletin, 43(4), 693–697.
- Bennett M. R. (2011). Schizophrenia: susceptibility genes, dendritic-spine pathology and gray matter loss. Progress in neurobiology, 95(3), 275–300. 638
- Reale, M., Costantini, E., & Greig, N. H. (2021). Cytokine Imbalance in Schizophrenia. From Research to Clinic: Potential Implications for Treatment. Frontiers in psychiatry, 12, 536257.
- Vilain, J., Galliot, A. M., Durand-Roger, J., Leboyer, M., Llorca, P. M., Schürhoff, F., & Szöke, A. (2013). Les facteurs de risque environnementaux de la schizophrénie [Environmental risk factors for schizophrenia: a review]. L'Encephale, 39(1), 19–28.
- Henriksen, M. G., Nordgaard, J., & Jansson, L. B. (2017). Genetics of Schizophrenia: Overview of Methods, Findings and Limitations. Frontiers in human neuroscience, 11, 322.
- Comer, A. L., Carrier, M., Tremblay, M. È., & Cruz-Martín, A. (2020). The Inflamed Brain in Schizophrenia: The Convergence of Genetic and Environmental Risk Factors That Lead to Uncontrolled Neuroinflammation. Frontiers in cellular neuroscience, 14, 274.
- Gober, R., Ardalan, M., Shiadeh, S., Duque, L., Garamszegi, S. P., Ascona, M., Barreda, A., Sun, X., Mallard, C., & Vontell, R. T. (2022). Microglia activation in postmortem brains with schizophrenia demonstrates distinct morphological changes between brain regions. Brain pathology (Zurich, Switzerland), 32(1), e13003.
- Parellada, E., & Gassó, P. (2021). Glutamate and microglia activation as a driver of dendritic apoptosis: a core pathophysiological mechanism to understand schizophrenia. Translational psychiatry, 11(1), 271.
- Monji, A., & Mizoguchi, Y. (2022). Neuroinflammation in Late-Onset Schizophrenia: Viewing from the Standpoint of the Microglia Hypothesis. Neuropsychobiology, 81(2), 98–103.
- Crapser, J. D., Arreola, M. A., Tsourmas, K. I., & Green, K. N. (2021). Microglia as hackers of the matrix: sculpting synapses and the extracellular space. Cellular & molecular immunology, 18(11), 2472–2488.
- Nguyen, P. T., Dorman, L. C., Pan, S., Vainchtein, I. D., Han, R. T., Nakao-Inoue, H., Taloma, S. E., Barron, J. J., Molofsky, A. B., Kheirbek, M. A., & Molofsky, A. V. (2020). Microglial Remodeling of the Extracellular Matrix Promotes Synapse Plasticity. Cell, 182(2), 388–403.e15.
- Dzyubenko, E., Manrique-Castano, D., Kleinschnitz, C., Faissner, A., & Hermann, D. M. (2018). Role of immune responses for extracellular matrix remodeling in the ischemic brain. Therapeutic advances in neurological disorders, 11, 1756286418818092.
- Ghorbani, S., & Yong, V. W. (2021). The extracellular matrix as modifier of neuroinflammation and remyelination in multiple sclerosis. Brain: a journal of neurology, 144(7), 1958–1973.
- Gaudet, A. D., & Popovich, P. G. (2014). Extracellular matrix regulation of inflammation in the healthy and injured spinal cord. Experimental neurology, 258, 24–34.
- Rolls, A., Shechter, R., London, A., Segev, Y., Jacob-Hirsch, J., Amariglio, N., Rechavi, G., & Schwartz, M. (2008). Two faces of chondroitin sulfate proteoglycan in spinal cord repair: a role in microglia/macrophage activation. PLoS medicine, 5(8), e171.
- Stankiewicz, A. M., Goscik, J., Majewska, A., Swiergiel, A. H., and Juszczak, G. R. (2015). The effect of acute and chronic social stress on the hippocampal transcriptome in mice. PLoS One 10:e0142195.
- Lehmann, M. L., Weigel, T. K., Cooper, H. A., Elkahloun, A. G., Kigar, S. L., and Herkenham, M. (2018). Decoding microglia responses to psychosocial stress reveals blood-brain barrier breakdown that may drive stress susceptibility. Sci. Rep. 8:11240.
- Rempe, R. G., Hartz, A., & Bauer, B. (2016). Matrix metalloproteinases in the brain and blood-brain barrier: Versatile breakers and makers. Journal of cerebral blood flow and metabolism: official journal of the International Society of Cerebral Blood Flow and Metabolism, 36(9), 1481–1507.
- Chopra, K., Baveja, A., & Kuhad, A. (2015). MMPs: a novel drug target for schizophrenia. Expert opinion on therapeutic targets, 19(1), 77–85.
- Cabral-Pacheco, G. A., Garza-Veloz, I., Castruita-De la Rosa, C., Ramirez-Acuña, J. M., Perez-Romero, B. A., Guerrero-Rodriguez, J. F., Martinez-Avila, N., & Martinez-Fierro, M. L. (2020). The Roles of Matrix Metalloproteinases and Their Inhibitors in Human Diseases. International journal of molecular sciences, 21(24), 9739.
- Kim, Y. S., & Joh, T. H. (2012). Matrix metalloproteinases, new insights into the understanding of neurodegenerative disorders. Biomolecules & therapeutics, 20(2), 133–143.
- Rosenberg G. A. (2009). Matrix metalloproteinases and their multiple roles in neurodegenerative diseases. The Lancet. Neurology, 8(2), 205–216.
- Haylock-Jacobs, S., Keough, M. B., Lau, L., & Yong, V. W. (2011). Chondroitin sulphate proteoglycans: extracellular matrix proteins that regulate immunity of the central nervous system. Autoimmunity reviews, 10(12), 766–772.
- Sethi, M. K., & Zaia, J. (2017). Extracellular matrix proteomics in schizophrenia and Alzheimer's disease. Analytical and bioanalytical chemistry, 409(2), 379–394.
- Fujioka, H., Dairyo, Y., Yasunaga, K., & Emoto, K. (2012). Neural functions of matrix metalloproteinases: plasticity, neurogenesis, and disease. Biochemistry research international, 2012, 789083.
- Devanarayanan, S., Nandeesha, H., Kattimani, S., & Sarkar, S. (2016). Relationship between matrix metalloproteinase-9 and oxidative stress in drug-free male schizophrenia: a case control study. Clinical chemistry and laboratory medicine, 54(3), 447–452.
- De Luca, C., & Papa, M. (2017). Matrix Metalloproteinases, Neural Extracellular Matrix, and Central Nervous System Pathology. Progress in molecular biology and translational science, 148, 167–202.
- Mahi, N. A., Najafabadi, M. F., Pilarczyk, M., Kouril, M., & Medvedovic, M. (2019). GREIN: An Interactive Web Platform for Re-analyzing GEO RNA-seq Data. Scientific reports, 9(1), 7580.
- Ferreira J. A. (2007). The Benjamini-Hochberg method in the case of discrete test statistics. The international journal of biostatistics, 3(1)
- Pathan, M., Keerthikumar, S., Ang, C. S., Gangoda, L., Quek, C. Y., Williamson, N. A., Mouradov, D., Sieber, O. M., Simpson, R. J., Salim, A., Bacic, A., Hill, A. F., Stroud, D. A., Ryan, M. T., Agbinya, J. I., Mariadason, J. M., Burgess, A. W., & Mathivanan, S. (2015). FunRich: An open access standalone functional enrichment and interaction network analysis tool. Proteomics, 15(15), 2597–2601.
- Xie, Z., Bailey, A., Kuleshov, M. V., Clarke, D., Evangelista, J. E., Jenkins, S. L., Lachmann, A., Wojciechowicz, M. L., Kropiwnicki, E., Jagodnik, K. M., Jeon, M., & Ma'ayan, A. (2021). Gene Set Knowledge Discovery with Enrichr. Current protocols, 1(3), e90.
- Shannon, P., Markiel, A., Ozier, O., Baliga, N. S., Wang, J. T., Ramage, D., Amin, N., Schwikowski, B., & Ideker, T. (2003). Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome research, 13(11), 2498–2504.
- Clarke, D., Kuleshov, M. V., Schilder, B. M., Torre, D., Duffy, M. E., Keenan, A. B., Lachmann, A., Feldmann, A. S., Gundersen, G. W., Silverstein, M. C., Wang, Z., & Ma'ayan, A. (2018).eXpression2Kinases (X2K) Web: linking expression signatures to upstream cell signaling networks. Nucleic acids research, 46(W1), W171–W179.
- Berger, S. I., Posner, J. M., & Ma'ayan, A. (2007). Genes2Networks: connecting lists of gene symbols using mammalian protein interactions databases. BMC bioinformatics.
- Lachmann, A., & Ma'ayan, A. (2009). KEA: kinase enrichment analysis. Bioinformatics (Oxford, England), 25(5), 684–686.
- Licursi, V., Conte, F., Fiscon, G., & Paci, P. (2019). MIENTURNET: an interactive web tool for microRNA-target enrichment and network-based analysis. BMC bioinformatics, 20(1), 545.
- Li, T., Wernersson, R., Hansen, R. B., Horn, H., Mercer, J., Slodkowicz, G., Workman, C. T., Rigina, O., Rapacki, K., Stærfeldt, H. H., Brunak, S., Jensen, T. S., & Lage, K. (2017). A scored human protein-protein interaction network to catalyze genomic interpretation. Nature methods, 14(1), 61–64.
- Ge, S. X., Jung, D., & Yao, R. (2020). ShinyGO: a graphical gene-set enrichment tool for animals and plants. Bioinformatics (Oxford, England), 36(8), 2628–2629.
- Khandaker, G. M., Cousins, L., Deakin, J., Lennox, B. R., Yolken, R., & Jones, P. B. (2015). Inflammation and immunity in schizophrenia: implications for pathophysiology and treatment. The lancet. Psychiatry, 2(3), 258–270.
- Berretta S. (2012). Extracellular matrix abnormalities in schizophrenia. Neuropharmacology, 62(3), 1584–1597.
- Ma, J., Ma, C., Li, J., Sun, Y., Ye, F., Liu, K., & Zhang, H. (2020). Extracellular Matrix Proteins Involved in Alzheimer's disease. Chemistry (Weinheim an der Bergstrasse, Germany), 26(53), 12101–12110.
- Maynard, T. M., Sikich, L., Lieberman, J. A., & LaMantia, A. S. (2001). Neural development, cell-cell signaling, and the "two-hit" hypothesis of schizophrenia. Schizophrenia bulletin, 457–476.
- McGuire, J. L., Depasquale, E. A., Funk, A. J., O'Donnovan, S. M., Hasselfeld, K., Marwaha, S., Hammond, J. H., Hartounian, V., Meador-Woodruff, J. H., Meller, J., & McCullumsmith, R. E. (2017). Abnormalities of signal transduction networks in chronic schizophrenia. NPJ schizophrenia, 3(1), 30.
- Wu, X., & Reddy, D. S. (2012). Integrins as receptor targets for neurological disorders. Pharmacology & therapeutics, 134(1), 68–81.
- Redies, C., Hertel, N., & Hübner, C. A. (2012). Cadherins and neuropsychiatric disorders. Brain research, 1470, 130–144.
- Punovuori, K., Malaguti, M., & Lowell, S. (2021). Cadherins in early neural development. Cellular and molecular life sciences: CMLS, 78(9), 4435–4450.
- Sun, M., Gewirtz, J. C., Bofenkamp, L., Wickham, R. J., Ge, H., & O'Connor, M. B. (2010). Canonical TGF-beta signaling is required for the balance of excitatory/inhibitory transmission within the hippocampus and prepulse inhibition of acoustic startle. The Journal of neuroscience: the official journal of the Society for Neuroscience, 30(17), 6025–6035.
- Sun, M., Gewirtz, J. C., Bofenkamp, L., Wickham, R. J., Ge, H., & O'Connor, M. B. (2010). Canonical TGF-beta signaling is required for the balance of excitatory/inhibitory transmission within the hippocampus and prepulse inhibition of acoustic startle. The Journal of neuroscience: the official journal of the Society for Neuroscience, 30(17), 6025–6035.
- Klass, C. M., Couchman, J. R., & Woods, A. (2000). Control of extracellular matrix assembly by syndecan-2 proteoglycan. Journal of cell science, 113 (Pt 3), 493–506.
- Feng, T., Tripathi, A., & Pillai, A. (2020). Inflammatory Pathways in Psychiatric Disorders: The case of Schizophrenia and Depression. Current behavioral neuroscience reports, 7(3), 128–138.
- Müller, N., Weidinger, E., Leitner, B., & Schwarz, M. J. (2015). The role of inflammation in schizophrenia. Frontiers in neuroscience, 9, 372.
- Müller, N., & Schwarz, M. J. (2008). A psychoneuroimmunological perspective to Emil Kraepelins dichotomy: schizophrenia and major depression as inflammatory CNS disorders. European archives of psychiatry and clinical neuroscience, 258 Suppl 2, 97–106.
- Nakata, K., Ujike, H., Sakai, A., Takaki, M., Imamura, T., Tanaka, Y., & Kuroda, S.(2003). Association study between the fibronectin gene and schizophrenia. American journal of medical genetics. Part B, Neuropsychiatric genetics: the official publication of the International Society of Psychiatric Genetics, 116B (1), 41–44.
- Purves-Tyson, T. D., Robinson, K., Brown, A. M., Boerrigter, D., Cai, H. Q., Weissleder, C., Owens, S. J., Rothmond, D. A., & Shannon Weickert, C. (2020). Increased Macrophages and C1qA, C3, C4 Transcripts in the Midbrain of People with Schizophrenia. Frontiers in immunology, 11, 2002.
- Muntané, G., Chillida, M., Aranda, S., Navarro, A., & Vilella, E. (2021). Coexpression of the discoidin domain receptor 1 gene with oligodendrocyte-related and schizophrenia risk genes in the developing and adult human brain. Brain and behavior, 11(8), e2309.
- Xu, R., Liang, J., Luo, Y., Wan, X., Li, K., Qi, L., Yuan, W., Chen, J., Wu, Z., Wang, M., Zhou, J., Xie, Y., Zhou, S., & He, F. (2018). Mass spectrometry identification of potential biomarker proteins in the 150-kD electrophoretic band in patients with schizophrenia. Medicine, 97(51), e13553.
- Huo, C., Liu, X., Zhao, J., Zhao, T., Huang, H., & Ye, H. (2018). Abnormalities in behaviour, histology and prefrontal cortical gene expression profiles relevant to schizophrenia in embryonic day 17 MAM-Exposed C57BL/6 mice. Neuropharmacology, 140, 287–301.
- Zammit, S., Hamshere, M., Dwyer, S., Georgiva, L., Timpson, N., Moskvina, V., Richards, A., Evans, D. M., Lewis, G., Jones, P., Owen, M. J., & O'Donovan, M. C. (2014). A population-based study of genetic variation and psychotic experiences in adolescents. Schizophrenia bulletin, 40(6), 1254–1262.
- Kular, J. K., Basu, S., & Sharma, R. I. (2014). The extracellular matrix: Structure, composition, age-related differences, tools for analysis and applications for tissue engineering. Journal of tissue engineering, 5, 2041731414557112.
- Smagin, D. A., Galyamina, A. G., Kovalenko, I. L., Babenko, V. N., & Kudryavtseva, N. N. (2019). Aberrant Expression of Collagen Gene Family in the Brain Regions of Male Mice with Behavioral Psychopathologies Induced by Chronic Agonistic Interactions. BioMed research international, 2019, 7276389.
- Omori, W., Hattori, K., Kajitani, N., Tsuchioka, M. O., Boku, S., Kunugi, H., Okamoto, Y., & Takebayashi, M. (2020). Increased matrix metalloproteinases in cerebrospinal fluids of patients with major depressive disorder and schizophrenia. The international journal of neuropsychopharmacology, 23(11), 713–720.
- Park, H. J., Kim, S. K., Kim, J. W., Kang, W. S., & Chung, J. H. (2012). Association of thrombospondin 1 gene with schizophrenia in Korean population. Molecular biology reports, 39(6), 6875–6880.
- Huang, H. S., Matevossian, A., Whittle, C., Kim, S. Y., Schumacher, A., Baker, S. P., & Akbarian, S. (2007). Prefrontal dysfunction in schizophrenia involves mixed-lineage leukemia 1-regulated histone methylation at GABAergic gene promoters. The Journal of neuroscience: the official journal of the Society for Neuroscience, 27(42), 11254–11262
- Ruzicka, W. B., Subburaju, S., & Benes, F. M. (2015). Circuit- and Diagnosis-Specific DNA Methylation Changes at γ-Aminobutyric Acid-Related Genes in Postmortem Human Hippocampus in Schizophrenia and Bipolar Disorder. JAMA psychiatry, 72(6), 541–551.
- Billingsley, K. J., Manca, M., Gianfrancesco, O., Collier, D. A., Sharp, H., Bubb, V. J., & Quinn, J. P. (2018). Regulatory characterisation of the schizophrenia-associated CACNA1C proximal promoter and the potential role for the transcription factor EZH2 in schizophrenia aetiology. Schizophrenia research, 199, 168–175.
- Brattås, P. L., Jönsson, M. E., Fasching, L., Nelander Wahlestedt, J., Shahsavani,M., Falk, R., Falk, A., Jern, P., Parmar, M., & Jakobsson, J. (2017). TRIM28 Controls a Gene Regulatory Network Based on Endogenous Retroviruses in Human Neural Progenitor Cells. Cell reports, 18(1), 1–11.
- Ni, X., Trakalo, J., Valente, J., Azevedo, M. H., Pato, M. T., Pato, C. N., & Kennedy, J. L. (2005). Human p53 tumor suppressor gene (TP53) and schizophrenia: case-control and family studies. Neuroscience letters, 388(3), 173–178.
- Rees, E., Creeth, H., Hwu, H. G., Chen, W. J., Tsuang, M., Glatt, S. J., Rey, R., Kirov, G., Walters, J., Holmans, P., Owen, M. J., & O'Donovan, M. C. (2021). Schizophrenia, autism spectrum disorders and developmental disorders share specific disruptive coding mutations. Nature communications, 12(1), 5353.
- Jope, R. S., & Roh, M. S. (2006). Glycogen synthase kinase-3 (GSK3) in 1034psychiatric diseases and therapeutic interventions. Current drug targets, 7(11), 1421–1434. 1035
- Rodrigues-Amorim, D., Rivera-Baltanás, T., Fernández-Palleiro, P., Iglesias-Martínez-Almeida, M., Freiría-Martínez, L., Jarmardo-Rodriguez, C., Del Carmen Vallejo-Curto, M., Álvarez-Ariza, M., López-García, M., de Las Heras, E., García-Caballero, A., Olivares, J. M., & Spuch, C. (2021). Changes in the Brain Extracellular Matrix Composition in schizophrenia: A Pathophysiological Dysregulation and a Potential Therapeutic Target. Cellular and molecular neurobiology, 10.1007/s10571-021-01073-8. Advance online publication.
- Pantazopoulos, H., Woo, T. U., Lim, M. P., Lange, N., & Berretta, S. (2010). Extracellular matrix-glial abnormalities in the amygdala and entorhinal cortex of subjects diagnosed with schizophrenia. Archives of general psychiatry, 67(2), 155–166.
- Matuszko, G., Curreli, S., Kaushik, R., Becker, A., & Dityatev, A. (2017). Extracellular matrix alterations in the ketamine model of schizophrenia. Neuroscience, 350, 13–22.
- Su, J., Chen, J., Lippold, K., Monavarfeshani, A., Carrillo, G. L., Jenkins, R., & Fox, M. A. (2016). Collagen-derived matricryptins promote inhibitory nerve terminal formation in the developing neocortex. The Journal of cell biology, 212(6), 721–736.
- Lewis, D. A., & Moghaddam, B. (2006). Cognitive dysfunction in schizophrenia: convergence of gamma-aminobutyric acid and glutamate alterations. Archives of neurology, 63(10), 1372–1376.
- Kehrer, C., Maziashvili, N., Dugladze, T., & Gloveli, T. (2008). Altered Excitatory-Inhibitory Balance in the NMDA-Hypofunction Model of Schizophrenia. Frontiers in molecular neuroscience.
- Baeten, K. M., & Akassoglou, K. (2011). Extracellular matrix and matrix receptors in blood-brain barrier formation and stroke. Developmental neurobiology, 71(11), 1018–1039.
- Reichardt, L. F., & Prokop, A. (2011). Introduction: the role of extracellular matrix in nervous system development and maintenance. Developmental neurobiology, 71(11), 883–888.
- Khadilkar, R. J., Ho, K., Venkatesh, B., & Tanentzapf, G. (2020). Integrins Modulate Extracellular Matrix Organization to Control Cell Signaling during Hematopoiesis. Current biology: CB, 30(17), 3316–3329.e5.
- Bullock, W. M., Cardon, K., Bustillo, J., Roberts, R. C., & Perrone-Bizzozero, N. I. (2008). Altered expression of genes involved in GABAergic transmission and neuromodulation of granule cell activity in the cerebellum of schizophrenia patients. The American journal of psychiatry, 165(12), 1594–1603.
- Harrison, P. J., McLaughlin, D., & Kerwin, R. W. (1991). Decreased hippocampal expression of a glutamate receptor gene in schizophrenia. Lancet (London, England), 337(8739), 450–452.
- Reynolds, G. P., Czudek, C., & Andrews, H. B. (1990). Deficit and hemispheric asymmetry of GABA uptake sites in the hippocampus in schizophrenia. Biological psychiatry, 27(9), 1038–1044.
- Simpson, M. D., Slater, P., Deakin, J. F., Royston, M. C., & Skan, W. J. (1989). Reduced GABA uptake sites in the temporal lobe in schizophrenia. Neuroscience letters, 107(1-3), 211–215.
- Joyce, J. N., Shane, A., Lexow, N., Winokur, A., Casanova, M. F., & Kleinman, J. E. (1993). Serotonin uptake sites and serotonin receptors are altered in the limbic system of schizophrenics. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology, 8(4), 315–336.
- de Jonge, J. C., Vinkers, C. H., Hulshoff Pol, H. E., & Marsman, A. (2017). GABAergic Mechanisms in Schizophrenia: Linking Postmortem and In Vivo Studies. Frontiers in psychiatry, 8, 118.
- Wang, X., Hu, Y., Liu, W., Ma, Y., Chen, X., Xue, T., & Cui, D. (2021). Molecular Basis of GABA Hypofunction in Adolescent Schizophrenia-Like Animals. Neural plasticity, 2021, 9983438.
- Caballero, A., & Tseng, K. Y. (2016). GABAergic Function as a Limiting Factor for Prefrontal Maturation during Adolescence. Trends in neurosciences, 39(7), 441–448.
- Jahangir, M., Zhou, J. S., Lang, B., & Wang, X. P. (2021). GABAergic System Dysfunction and Challenges in Schizophrenia Research. Frontiers in cell and developmental biology, 9, 663854.
- Kegeles, L. S., Mao, X., Stanford, A. D., Girgis, R., Ojeil, N., Xu, X., Gil, R., Slifstein, M., Abi-Dargham, A., Lisanby, S. H., & Shungu, D. C. (2012). Elevated prefrontal cortex γ-aminobutyric acid and glutamate-glutamine levels in schizophrenia measured in vivo with proton magnetic resonance spectroscopy. Archives of general psychiatry, 69(5), 449–459.
- Wang, T., Sinha, A. S., Akita, T., Yanagawa, Y., & Fukuda, A. (2018). Alterations of GABAergic Neuron-Associated Extracellular Matrix and Synaptic Responses in Gad1-Heterozygous Mice Subjected to Prenatal Stress. Frontiers in cellular neuroscience, 12, 284.
- Elmer, B. M., & McAllister, A. K. (2012). Major histocompatibility complex class I proteins in brain development and plasticity. Trends in neurosciences, 35(11), 660–670.
- Glynn, M. W., Elmer, B. M., Garay, P. A., Liu, X. B., Needleman, L. A., El-Sabeawy, F., & McAllister, A. K. (2011). MHCI negatively regulates synapse density during the establishment of cortical connections. Nature neuroscience, 14(4), 442–451.
- Maynard, T. M., Sikich, L., Lieberman, J. A., & LaMantia, A. S. (2001). Neural development, cell-cell signaling, and the "two-hit" hypothesis of schizophrenia. Schizophrenia bulletin, 27(3), 457–476.
- Neumann, H., Cavalié, A., Jenne, D. E., & Wekerle, H. (1995). Induction of MHC class I genes in neurons. Science (New York, N.Y.), 269(5223), 549–552.
- Goddard, C. A., Butts, D. A., & Shatz, C. J. (2007). Regulation of CNS synapses by neuronal MHC class I. Proceedings of the National Academy of Sciences of the United States of America, 104(16), 6828–6833.
- Michel, M., Schmidt, M. J., & Mirnics, K. (2012). Immune system gene dysregulation in autism and schizophrenia. Developmental neurobiology, 72(10), 1277–1287.
- Rempe, R. G., Hartz, A., & Bauer, B. (2016). Matrix metalloproteinases in the brain and blood-brain barrier: Versatile breakers and makers. Journal of cerebral blood flow and metabolism: official journal of the International Society of Cerebral Blood Flow and Metabolism, 36(9), 1481–1507.
- Keshri, N., Nandeesha, H., Rajappa, M., & Menon, V. (2021). Matrix metalloproteinase-9 increases the risk of cognitive impairment in schizophrenia. Nordic journal of psychiatry, 75(2), 130–134.
- Schoretsanitis, G., de Filippis, R., Ntogka, M., Leucht, S., Correll, C. U., & Kane, J. M. (2021). Matrix Metalloproteinase 9 Blood Alterations in Patients with Schizophrenia Spectrum Disorders: A Systematic Review and Meta-Analysis. Schizophrenia bulletin, 47(4), 986–996.
- Bitanihirwe, B., & Woo, T. W. (2020). A conceptualized model linking matrix metalloproteinase-9 to schizophrenia pathogenesis. Schizophrenia research, 218, 28–35.
- Reinhard, S. M., Razak, K., & Ethell, I. M. (2015). A delicate balance: role of MMP-9 in brain development and pathophysiology of neurodevelopmental disorders. Frontiers in cellular neuroscience, 9, 280.
- Kudo, N., Yamamori, H., Ishima, T., Nemoto, K., Yasuda, Y., Fujimoto, M., Azechi, H., Niitsu, T., Numata, S., Ikeda, M., Iyo, M., Ohmori, T., Fukunaga, M., Watanabe, Y., Hashimoto, K., & Hashimoto, R. (2020). Plasma levels of matrix metalloproteinase-9 (MMP-9) are associated with cognitive performance in patients with schizophrenia. Neuropsychopharmacology reports, 40(2), 150–156.
- Dwir, D., Giangreco, B., Xin, L., Tenenbaum, L., Cabungcal, J. H., Steullet, P., Goupil, A., Cleusix, M., Jenni, R., Chtarto, A., Baumann, P. S., Klauser, P., Conus, P., Tirouvanziam, R., Cuenod, M., & Do, K. Q. (2020). MMP9/RAGE pathway overactivation mediates redox dysregulation and neuroinflammation, leading to inhibitory/excitatory imbalance: a reverse translation study in schizophrenia patients. Molecular psychiatry, 25(11), 2889–2904.
- Beroun, A., Mitra, S., Michaluk, P., Pijet, B., Stefaniuk, M., & Kaczmarek, L. (2019). MMPs in learning and memory and neuropsychiatric disorders. Cellular and molecular life sciences: CMLS, 76(16), 3207–3228.
- Suzuki, Y., Nagai, N., & Umemura, K. (2016). A Review of the Mechanisms of Blood-Brain Barrier Permeability by Tissue-Type Plasminogen Activator Treatment for Cerebral Ischemia. Frontiers in cellular neuroscience, 10, 2.
- Inflammation in Heart Failure. (Elsevier, 2015). https://doi.org/10.1016/B978-0-12-800039-7.00004-9
- Rempe, R. G., Hartz, A., Soldner, E., Sokola, B. S., Alluri, S. R., Abner, E. L., Kryscio, R. J., Pekcec, A., Schlichtiger, J., & Bauer, B. (2018). Matrix Metalloproteinase-Mediated Blood-Brain Barrier Dysfunction in Epilepsy. The Journal of neuroscience: the official journal of the Society for Neuroscience, 38(18), 4301–4315.
- Yang, C., Hawkins, K. E., Doré, S., & Candelario-Jalil, E. (2019). Neuroinflammatory mechanisms of blood-brain barrier damage in ischemic stroke. American journal of physiology. Cell physiology, 316(2), C135–C153.
- Joo, S. H., Kwon, K. J., Kim, J. W., Kim, J. W., Hasan, M. R., Lee, H. J., Han, S. H., & Shin, C. Y. (2010). Regulation of matrix metalloproteinase-9 and tissue plasminogen activator activity by alpha-synuclein in rat primary glial cells. Neuroscience letters, 469(3), 352–356.
Supplementary Data 1-3 are not available with this version