Gene expression profile series analysis
The GSE series GSE1297, GSE28416, and GSE29378 consisted of 22283 genes with 29 groups (Control: 9 and diseased: 20), 54675 genes with 30 groups (Control: 8 and diseased: 22), and 48803 genes with 63 groups (Control: 32 and diseased: 31) respectively (Table 1). After probe ids-gene symbol matching and duplicacy removal, a total of 13310, 22185, and 25158 genes remained in GSE1297, GSE28416, and GSE29378 respectively. The datasets of each GSE series were mapped and 10834 common genes (Fig. 3A) among these GSE series were retrieved utilizing the InteractiVenn tool (http://www.interactivenn.net/). Figure 3B exhibits a principal component analysis (PCA) plot illustrating the distribution of clusters and the variance of genes between control vs diseased (AD) groups along the first two principal components.
Table 1
List of GSE series as retrieved from the GEO database.
Series | Sample | Normal | Disease | Genes | Upregulated | Downregulated | Tissue | Platform | Reference |
GSE1297 | 29 | 9 | 20 | 22283 | 2 | 4 | Hippocampus | GPL96 | (Blalock, Geddes, et al. 2004) |
GSE28146 | 30 | 8 | 22 | 54675 | 2 | 10 | Hippocampus | GPL570 | (Blalock, Buechel et al. 2011) |
GSE29378 | 63 | 32 | 31 | 48803 | 227 | 516 | Hippocampus | GPL6947 | (Miller, Woltjer et al. 2013) |
DESeq2 analysis
The expression profile datasets of each GSE series were further normalized in RStudio using DESeq2 module and we obtained 6 (upregulated: 2 and downregulated: 4), 12 (upregulated: 2 and downregulated: 10), and 743 (upregulated: 227 and downregulated: 516) DEGs from GSE1297, GSE28146, and GSE29378 respectively as shown in MA plot (Fig. 3C-E). After log2FC and p-value adjustment, 5 (upregulated: 2 and downregulated: 3), 7 (upregulated: 1 and downregulated: 6), and 334 (upregulated: 100 and downregulated: 234) significant genes from GSE1297, GSE28146, and GSE29378 remained as displayed in the density plot (Fig. 3F-H). Further, non-significant genes were discarded as their expression in the tissue was negligible. The annotated heatmap plots illustrate the significant genes obtained from the datasets GSEGSE1297, GSE28416, and GSE29378 (Fig. 3I-K) and the PCA plot shows 103 upregulated vs 243 downregulated genes (Fig. 3L).
PPI network construction
A cluster of genetically similar proteins is referred to as a module, which serves as a pivotal component in BP. These proteins execute their function through a precise phenomenon known as PPIs, which take place temporally and spatially [41]. The entire set of 346 genes that exhibited significance were put into the STRING database to construct a network having 346 nodes, 577 edges, a node degree of 3.38, a PPI enrichment p-value of 0.961, and a clustering coefficient of 0.376. Moreover, the network file obtained from STRING was subjected to analysis in Cytoscape 3.10.0 to construct native networks and identification of hub genes. Out of the total 346 nodes in the network, a subset of 95 nodes were found to have no interactions with any other nodes. Consequently, these 95 nodes were excluded from the network. The network representing the differentially expressed genes (DEGs) that are either upregulated or downregulated consists of 246 nodes and 503 edges. The average number of neighboring nodes for each node in the network is 4.089, indicating a moderate level of connectivity. The clustering coefficient of the network is 0.211, suggesting that there is a tendency for nodes to form clusters or groups within the network. The identification of complex network modules was conducted using the CytoHubba plugin integrated into the Cytoscape 3.10.0 software. Within the comprehensive network, we have employed 11 distinct clustering techniques to identify hub nodes. From the primary network, we have extracted the top 50 nodes using these clustering methods, each of which yielded varying edges: MCC (171 edges), DMNC (69 edges), MNC (168 edges), Degree (157 edges), EPC (180 edges), BottleNeck (113 edges), EcCentricity (121 edges), Closeness (168 edges), Radiality (167 edges), Betweenness (122 edges), and Stress (129 edges). A total of 50 key genes were found among which 7 genes (TBP, SOX2, CCND1, RELA, CAT, HDAC2, and TNF) were common among these 11 methods. Native networks for all significant DEGs based on Log2FC value and subnetwork of the top 50 hub genes for all 11 network topologies methods were constructed (Fig. 4). Lastly, the common genes from each of the 11 subnetworks were mapped, and 50 hub genes were retrieved using Calculate and Draw custom Venn diagram tool (CDCVD) [30] (Table 2).
Table 2
Overlapped key genes based on various topological methods.
Topological methods | Count | Genes |
Betweenness Bottleneck Closeness Degree EPC EcCentricity MCC MNC Radiality Stress | 7 | TBP SOX2 CCND1 RELA CAT HDAC2 TNF |
Betweenness Bottleneck Closeness Degree EPC MCC MNC Radiality Stress | 1 | EGF |
Betweenness Bottleneck Closeness Degree EcCentricity MNC Radiality Stress | 1 | RPS6 |
Bottleneck Closeness Degree EPC EcCentricity MCC MNC Radiality | 1 | IRF4 |
Closeness Degree EPC EcCentricity MCC MNC Radiality Stress | 1 | NOS3 |
Betweenness Bottleneck Closeness Degree EcCentricity Radiality Stress | 1 | YWHAZ |
Closeness Degree EPC EcCentricity MNC Radiality | 1 | PAX5 |
Closeness Degree EPC EcCentricity Radiality Stress | 1 | H2AFZ |
Betweenness Bottleneck Closeness Radiality Stress | 1 | STK11 |
Closeness EPC EcCentricity Radiality | 2 | ETS1 SOCS3 |
Degree EPC MCC MNC | 3 | BUB1 FCGR2B CCR7 |
Betweenness Bottleneck Radiality | 1 | RHEB |
Betweenness Bottleneck Stress | 2 | DPYSL3 TXNRD2 |
Closeness EPC MNC | 1 | FUT4 |
EPC EcCentricity MNC | 1 | IRF3 |
EPC MCC MNC | 1 | PDCD1LG2 |
Betweenness Bottleneck | 1 | CAD |
Betweenness Stress | 3 | KRT8 TBL1X ITCH |
Closeness Radiality | 2 | KNG1 CYP19A1 |
Degree MCC | 2 | MCM3 E2F1 |
Degree MNC | 1 | CS |
MCC MNC | 1 | PDCD1 |
Betweenness | 2 | RAD23A FKBP4 |
Bottleneck | 4 | TXNIP MYH9 AXIN1 MYO1C |
EcCentricity | 4 | SPO11 SIX3 FABP7 NUP93 |
MCC | 3 | TNFSF4 MCM5 AOC3 |
Stress | 2 | AGT RPGR |
GSEA, GO terms, and KEGG pathway analysis
The gene ontologies and KEGG pathway analysis for 50 key genes was done at the online platform STRING. For GSEA analysis, the gseGO and fgsea module in RStudio was utilized. The pathway analysis confirmed 351 GO terms (19 CC, 18 MF, and 314 BP,), and 68 KEGG pathways. The 20 enriched GO terms and KEGG pathways are listed in Table 3. The synaptic location and functions of these genes in brain tissue were determined in the SynGO [34] tool. The genes RELA, DPYSL3, RHEB, YWHAZ, RPS6, and AXIN1 were mapped to 8 CC and 5 BP enriched terms (Table 4) within the synapse region of the brain. The CC and BP ontology terms along with genes having key roles in their respective synapse locations are shown in Fig. 5. The GSEA analysis of 50 key genes was performed using the ClusterProfiler module of R-studio and the top 30 GO terms were plotted against their p-adj. and fold change values. The dot plot of enriched terms concerning the gene ratio in each term and the expression of genes i.e., whether genes were activated or suppressed were plotted (Fig. 6A).
Table 3
The tabulation includes the top ten to twenty GO terms and KEGG pathways that exhibited a higher level of enrichment.
Source | GO ID | Term description | Log10FC | p-value | Genes |
GO: BP | GO:1901700 | Response to oxygen-containing compound | 0.69 | 1.11E-06 | CCND1;CAT;CCR7;CAD;TNFSF4;H2AFZ;NOS3;E2F1;FCGR2B; |
| | | | | AGT;ETS1;CYP19A1;PDCD1LG2;TXNRD2;RELA;TNF;HDAC2;KRT8;TXNIP;IRF3 |
GO: BP | GO:0050727 | Regulation of inflammatory response | 1.04 | 1.32E-05 | CCR7;TNFSF4;SOCS3;FCGR2B;AGT;ETS1;CYP19A1;RELA;TNF;IRF3 |
GO: BP | GO:0042981 | Regulation of apoptotic process | 0.62 | 5.14E-05 | CAT;CCR7;KNG1;NOS3;SOX2;SOCS3;PDCD1;E2F1;AGT;RPS6;ETS1; |
| | | | | YWHAZ;RELA;TNF;HDAC2;TXNIP;ITCH |
GO: BP | GO:0051960 | Regulation of nervous system development | 0.72 | 0.00012 | FKBP4;SIX3;RHEB;EGF;SOX2;STK11;DPYSL3;E2F1;AGT;YWHAZ;RELA; |
| | | | | TNF;HDAC2 |
GO: BP | GO:0014015 | Positive regulation of gliogenesis | 1.43 | 0.00024 | RHEB;E2F1;RELA;TNF;HDAC2 |
GO:BP | GO:0050767 | Regulation of neurogenesis | 0.71 | 0.00087 | FKBP4;SIX3;RHEB;SOX2;STK11;DPYSL3;E2F1;AGT;RELA;TNF;HDAC2 |
GO: BP | GO:0051962 | Positive regulation of nervous system development | 0.8 | 0.0012 | RHEB;EGF;STK11;DPYSL3;E2F1;AGT;RELA;TNF;HDAC2 |
GO: BP | GO:0042542 | Response to hydrogen peroxide | 1.21 | 0.0017 | CAT;ETS1;RELA;HDAC2;TXNIP |
GO: BP | GO:0042325 | Regulation of phosphorylation | 0.52 | 0.0039 | CCND1;CCR7;AXIN1;EGF;NUP93;SOX2;STK11;SOCS3;FCGR2B;AGT |
| | | | | ;YWHAZ;TNF;HDAC2;ITCH |
GO: BP | GO:0043408 | Regulation of MAPK cascade | 0.68 | 0.0063 | CCR7;AXIN1;EGF;SOX2;FCGR2B;AGT;YWHAZ;TNF;ITCH |
GO: BP | GO:0050768 | Negative regulation of neurogenesis | 0.9 | 0.0066 | FKBP4;SIX3;SOX2;DPYSL3;TNF;HDAC2 |
GO: BP | GO:0034599 | Positive regulation of interleukin-12 production | 0.97 | 0.0035 | CAT;NOS3;ETS1;TXNRD2;RELA;HDAC2 |
GO: BP | GO:0001934 | Regulation of oxidoreductase activity | 0.62 | 0.0041 | CCND1;CCR7;AXIN1;EGF;SOX2;STK11;SOCS3;FCGR2B;AGT;TNF;HDAC2 |
GO: BP | GO:0051172 | Negative regulation of neuron differentiation | 0.43 | 0.0059 | TBL1X;CCND1;SIX3;KNG1;TNFSF4;H2AFZ;SOX2;SOCS3;E2F1;PAX5;AGT; |
| | | | | RELA;TNF;HDAC2;TXNIP;IRF3;ITCH |
GO: BP | GO:0043065 | Regulation of canonical Wnt signaling pathway | 0.67 | 0.0141 | KNG1;PDCD1;E2F1;AGT;RPS6;YWHAZ;TNF;TXNIP |
GO: BP | GO:0030111 | Nitric oxide mediated signal transduction | 0.79 | 0.0176 | TBL1X;SIX3;AXIN1;EGF;SOX2;STK11 |
GO: BP | GO:0002676 | Positive regulation of amyloid-beta formation | 1.88 | 0.0181 | CYP19A1;TNF |
GO: BP | GO:0000228 | Nuclear chromosome | 0.66 | 0.00054 | MCM5;SIX3;H2AFZ;BUB1;SOX2;E2F1;PAX5;SPO11;IRF4;TBP;ETS1;RELA;HDAC2;IRF3;MCM3 |
GO: BP | GO:0005667 | Transcription regulator complex | 0.9 | 0.00054 | TBL1X;CCND1;SIX3;SOX2;E2F1;TBP;ETS1;RELA;HDAC2 |
GO: CC | GO:0032991 | Protein-containing complex | 0.36 | 0.00054 | FKBP4;MCM5;MYH9;TBL1X;CCND1;CAT;SIX3;RHEB;AXIN1;CAD;EGF;H2AFZ;BUB1;NUP93; |
| | | | | SOX2;SOCS3;DPYSL3;E2F1;MYO1C;RPS6;IRF4;TBP;ETS1;RELA;TNF;HDAC2;KRT8;RAD23A;ITCH;MCM3 |
GO: CC | GO:0070013 | Intracellular organelle lumen | 0.33 | 0.00054 | FKBP4;MCM5;TBL1X;CCND1;CAT;SIX3;CAD;KNG1;EGF;H2AFZ;BUB1;NUP93;SOX2;STK11;CS;E2F1;PAX5; |
| | | | | MYO1C;SPO11;RPS6;IRF4;TBP;ETS1;YWHAZ;TXNRD2;RELA;HDAC2;KRT8;TXNIP;RAD23A;IRF3;ITCH;MCM3 |
GO: CC | GO:0031981 | Nuclear lumen | 0.34 | 0.0023 | FKBP4;MCM5;TBL1X;CCND1;SIX3;CAD;H2AFZ;BUB1;NUP93;SOX2;STK11;E2F1;PAX5;MYO1C;SPO11;RPS6; |
| | | | | IRF4;TBP;ETS1;YWHAZ;RELA;HDAC2;KRT8;RAD23A;IRF3;ITCH;MCM3 |
GO: CC | GO:0043231 | Intracellular membrane-bounded organelle | 0.19 | 0.0023 | FKBP4;MCM5;MYH9;TBL1X;CCND1;CAT;CCR7;SIX3;RHEB;AXIN1;CAD;KNG1;EGF;H2AFZ;NOS3;BUB1;NUP93; |
| | | | | AOC3;SOX2;STK11;CS;E2F1;PAX5;FUT4;MYO1C;FABP7;SPO11;RPGR;RPS6;IRF4;TBP;ETS1;YWHAZ;CYP19A1; |
| | | | | TXNRD2;RELA;HDAC2;KRT8;TXNIP;RAD23A;IRF3;ITCH;MCM3 |
GO: CC | GO:0005829 | Cytosol | 0.32 | 0.003 | FKBP4;MCM5;MYH9;CCND1;CAT;RHEB;AXIN1;CAD;NOS3;BUB1;SOX2;STK11;SOCS3;DPYSL3;PAX5;MYO1C; |
| | | | | AGT;FABP7;RPS6;IRF4;YWHAZ;TXNRD2;RELA;KRT8;TXNIP;RAD23A;IRF3;ITCH |
GO: CC | GO:0043227 | Membrane-bounded organelle | 0.15 | 0.003 | FKBP4;MCM5;MYH9;TBL1X;CCND1;CAT;CCR7;SIX3;RHEB;AXIN1;CAD;KNG1;EGF;H2AFZ;NOS3;BUB1;NUP93; |
| | | | | AOC3;SOX2;STK11;CS;DPYSL3;E2F1;PAX5;FUT4;MYO1C;AGT;FABP7;SPO11;RPGR;RPS6;IRF4;TBP;ETS1;YWHAZ; |
| | | | | CYP19A1;TXNRD2;RELA;TNF;HDAC2;KRT8;TXNIP;RAD23A;IRF3;ITCH;MCM3 |
GO: CC | GO:0005634 | Nucleus | 0.25 | 0.004 | FKBP4;MCM5;MYH9;TBL1X;CCND1;SIX3;RHEB;AXIN1;CAD;H2AFZ;NOS3;BUB1;NUP93;SOX2;STK11;CS;E2F1; |
| | | | | PAX5;MYO1C;FABP7;SPO11;RPS6;IRF4;TBP;ETS1;YWHAZ;RELA;HDAC2;KRT8;TXNIP;RAD23A;IRF3;ITCH;MCM3 |
GO: CC | GO:0005654 | Nucleoplasm | 0.35 | 0.0077 | FKBP4;MCM5;TBL1X;CCND1;CAD;BUB1;SOX2;STK11;E2F1;PAX5;MYO1C;RPS6;IRF4;TBP;ETS1;YWHAZ;RELA; |
| | | | | HDAC2;KRT8;RAD23A;IRF3;ITCH;MCM3 |
GO: CC | GO:0000790 | Nuclear chromatin | 0.6 | 0.0085 | SIX3;H2AFZ;SOX2;E2F1;PAX5;IRF4;TBP;ETS1;RELA;HDAC2;IRF3 |
GO: CC | GO:0043229 | Intracellular organelle | 0.14 | 0.0112 | FKBP4;MCM5;MYH9;TBL1X;CCND1;CAT;CCR7;SIX3;RHEB;AXIN1;CAD;KNG1;EGF;H2AFZ;NOS3;BUB1;NUP93; |
| | | | | AOC3;SOX2;STK11;CS;DPYSL3;E2F1;PAX5;FUT4;MYO1C;FABP7;SPO11;RPGR;RPS6;IRF4;TBP;ETS1;YWHAZ; |
GO: MF | GO:0008134 | Transcription factor binding | 0.8 | 0.0011 | FKBP4;TBL1X;CCND1;SIX3;E2F1;IRF4;TBP;ETS1;YWHAZ;RELA;HDAC2 |
GO: MF | GO:0019899 | Enzyme binding | 0.53 | 0.0011 | CCND1;CAT;SIX3;RHEB;AXIN1;CAD;EGF;E2F1;MYO1C;RPGR;RPS6;TBP;ETS1; |
| | | | | YWHAZ;RELA;TNF;HDAC2;TXNIP;RAD23A;ITCH |
GO: MF | GO:0043565 | Sequence-specific DNA binding | 0.64 | 0.0011 | MCM5;TBL1X;SIX3;H2AFZ;SOX2;E2F1;PAX5;IRF4;TBP;ETS1;RELA;TNF;HDAC2;IRF3; |
| | | | | MCM3 |
GO: MF | GO:0097159 | Organic cyclic compound binding | 0.32 | 0.0011 | FKBP4;MCM5;MYH9;TBL1X;CAT;SIX3;RHEB;CAD;H2AFZ;NOS3;BUB1;SOX2;STK11; |
| | | | | CS;E2F1;PAX5;MYO1C;SPO11;RPGR;RPS6;IRF4;TBP;ETS1;YWHAZ;CYP19A1;TXNRD2; |
| | | | | RELA;TNF;HDAC2;RAD23A;IRF3;MCM3 |
GO: MF | GO:1901363 | Heterocyclic compound binding | 0.32 | 0.0011 | FKBP4;MCM5;MYH9;TBL1X;CAT;SIX3;RHEB;CAD;H2AFZ;NOS3;BUB1;SOX2;STK11; |
| | | | | CS;E2F1;PAX5;MYO1C;SPO11;RPGR;RPS6;IRF4;TBP;ETS1;YWHAZ;CYP19A1;TXNRD2; |
| | | | | RELA;TNF;HDAC2;RAD23A;IRF3;MCM3 |
GO: MF | GO:0000976 | Transcription regulatory region sequence-specific DNA binding | 0.65 | 0.0035 | TBL1X;SIX3;H2AFZ;SOX2;E2F1;PAX5;IRF4;TBP;ETS1;RELA;TNF;IRF3 |
GO: MF | GO:0001216 | DNA-binding transcription activator activity | 0.83 | 0.0061 | SIX3;SOX2;E2F1;PAX5;IRF4;ETS1;RELA;IRF3 |
GO: MF | GO:0098772 | Molecular function regulator | 0.35 | 0.0379 | TBL1X;CCND1;SIX3;KNG1;EGF;TNFSF4;SOX2;STK11;SOCS3;E2F1;PAX5;AGT;RPGR;IRF4;TBP; |
GO: MF | GO:0042826 | Histone deacetylase binding | 1.13 | 0.0426 | CCND1;SIX3;RELA;HDAC2 |
GO: MF | GO:0016810 | Hydrolase activity; acting on carbon-nitrogen (but not peptide) bonds | 1.11 | 0.0459 | CAT;CAD;DPYSL3;HDAC2 |
KEGG | hsa04151 | PI3K-Akt signaling pathway | 0.94 | 0.00025 | CCND1;RHEB;EGF;NOS3;STK11;RPS6;YWHAZ;RELA |
KEGG | hsa04933 | AGE-RAGE signaling pathway in diabetic complications | 1.29 | 0.00039 | CCND1;NOS3;AGT;RELA;TNF |
KEGG | hsa04931 | Insulin resistance | 1.25 | 0.00044 | NOS3;SOCS3;AGT;RELA;TNF |
KEGG | hsa04920 | Adipocytokine signaling pathway | 1.35 | 0.0011 | STK11;SOCS3;RELA;TNF |
KEGG | hsa04218 | Cellular senescence | 1.11 | 0.0012 | CCND1;RHEB;E2F1;ETS1;RELA |
KEGG | hsa04211 | Longevity regulating pathway | 1.25 | 0.0018 | CAT;RHEB;STK11;RELA |
KEGG | hsa04066 | HIF-1 signaling pathway | 1.16 | 0.0034 | EGF;NOS3;RPS6;RELA |
KEGG | hsa04668 | TNF signaling pathway | 1.14 | 0.0037 | SOCS3;RELA;TNF;ITCH |
KEGG | hsa04071 | Sphingolipid signaling pathway | 1.12 | 0.0041 | KNG1;NOS3;RELA;TNF |
KEGG | hsa04068 | FoxO signaling pathway | 1.08 | 0.0052 | CCND1;CAT;EGF;STK11 |
KEGG | hsa04150 | mTOR signaling pathway | 1.01 | 0.0079 | RHEB;STK11;RPS6;TNF |
KEGG | hsa04622 | RIG-I-like receptor signaling pathway | 1.22 | 0.0086 | RELA;TNF;IRF3 |
KEGG | hsa04390 | Hippo signaling pathway | 1 | 0.008 | CCND1;AXIN1;SOX2;YWHAZ |
KEGG | hsa05214 | Glioma | 1.2 | 0.0088 | CCND1;EGF;E2F1 |
KEGG | hsa04660 | T cell receptor signaling pathway | 1.06 | 0.0175 | PDCD1;RELA;TNF |
KEGG | hsa04152 | AMPK signaling pathway | 0.98 | 0.0255 | CCND1;RHEB;STK11 |
KEGG | hsa04371 | Apelin signaling pathway | 0.94 | 0.0296 | CCND1;NOS3;RPS6 |
KEGG | hsa04072 | Phospholipase D signaling pathway | 0.89 | 0.0388 | RHEB;EGF;AGT |
KEGG | hsa04310 | Wnt signaling pathway | 0.87 | 0.0429 | TBL1X;CCND1;AXIN1 |
KEGG | hsa04630 | JAK-STAT signaling pathway | 0.86 | 0.0453 | CCND1;EGF;SOCS3 |
Table 4
Ontology terms and genes involved in brain tissue as retrieved from SynGO tool.
Cellular Component annotations |
Annotation ID | Genes | Ontology term |
1015 | RELA | synapse (GO:0045202) |
1658 | DPYSL3 | synapse (GO:0045202) |
1895 | RHEB | postsynaptic density (GO:0014069) |
2839 | YWHAZ | postsynaptic specialization (GO:0099572) |
2940 | RPS6 | presynaptic ribosome |
2951 | AXIN1 | synapse (GO:0045202) |
3241 | RPS6 | postsynaptic ribosome |
3911 | RHEB | synapse (GO:0045202) |
Biological process annotations |
1896 | RHEB | regulation of postsynapse organization (GO:0099175) |
2275 | RELA | postsynapse to nucleus signaling pathway (GO:0099527) |
2341 | YWHAZ | regulation of synapse maturation (GO:0090128) |
2352 | YWHAZ | synaptic target recognition (GO:0008039) |
3912 | RHEB | regulation of postsynapse organization (GO:0099175) |
The network construction of these enriched terms and genes was done as a Cnet plot. The location of genes RPGR, H2AFZ, and NUP3 can be seen in the microtubule organizing center. The genes DPYCL3, AGT, CCR7, and others are involved in various pathways such as positive regulation of cell projection organization and myeloid leukocyte migration (Fig. 6B). The ema plot of the top 30 enriched terms was plotted showing their interaction with each other in a constructed network pathway. The terms structural molecule activity, centrosome, microtubule organizing center, intracellular protein transport, the establishment of localization in cell, endocytic vesicle, cytoskeleton-dependent intracellular transport, cytoplasmic vesicle membrane, and vesicle membrane are making network with each other and some of these terms are associated with AD (Fig. 6C). The top 30 enriched terms of these genes are plotted concerning their respective p-value and log2FC in a ridge plot (Fig. 6D).
The enrichment plots of 23 enriched terms (15 BP, 5 CC, and 3 MF) found by GSEA analysis using the fgsea module of R-Studio are provided in the additional (Table S1). The chromosomal locations of these 50 key genes are given in Fig. 6E. A total of 12 upregulated and 35 downregulated genes namely; EGF, FABP7, GCGR2B, H2AFZ, IRF4, MCM3, MYO1C, NOS3, NUP93, PAX5, RPGR, RPS6, AGT, AOC3, AXIN1, BUB1, CAD, CCND1, CCR7, CYP19A1, DPYSL3, ETS1, FKBP4, FUT4, HDAC2, IRF3, ITCH, KNG1, KRT8, MCM5, MYH9, PDCD1, PDCD1LG2, RAD23A, RELA, RHEB, SIX3, SOCS3, SOX2, SPO11, STK11, TBL1X, TBP, TNF, TNFSF4, TXNIP, and YWHAZ out of 50 key genes were involved in AD-related pathways (Supplementary data).
The most noticed AD-related biological processes were regulation of nervous system development, phosphorylation, and inflammatory response, positive regulation of Aβ formation, gliogenesis, neurogenesis, IL-12 production, etc. The subcellular and macromolecular complex structure locations of these genes were also studied. These genes perform activities such as histone deacetylase binding, TF binding, and enzyme binding. KEGG pathways such as cellular senescence, sphingolipid signaling pathway, and Wnt signaling pathway were analyzed in the pathway analysis. These AD-associated 47 genes (Fig. 7A) involved in these terms are further chosen for miRNA-TF regulatory network construction.
miRNA-TF-gene regulatory network and FFL construction
We retrieved gene-associated 105 miRNAs and 173 TFs from miRNet. A literature study found that 1160 miRNAs and 93 TFs were associated with AD. In addition, AD-associated miRNAs and TFs were mapped with gene-associated miRNAs, and TFs yielded 95 miRNAs (Fig. 7B) and 93 TFs (Fig. 7C), respectively.
The AD-associated 95 miRNAs, 93 TFs, and 47 genes were submitted to STRING for network construction. The native network of the miRNA-TF mediated regulatory pathway with co-expressed genes was constructed in Cytoscape 3.10.0 having 235 nodes, and 1876 edges (Fig. 7D). The subnetworks of the top 50 hub miRNAs were retrieved by using Betweenness, Closeness, and Degree methods in the cytoHubba plugin of Cytoscape (Fig. 7E-G). Further, miRNA, TFs, and genes that were common in each of these methods were retrieved and the final regulatory network was constructed (Fig. 7H). The subnetwork constructed by the final regulatory network revealed interplay of TFs EGR1, MYC, POU5F1, NANOG, STAT3, SPI1, and HNF4A with miRNAs hsa-miR-155-5p and hsa-miR-16-5p in the inhibition and regulation of 3 upregulated and 12downregulated gene; H2AFZ, MCM3, MYO1C, AXIN1, TBP, RELA, RHEB, YWHAZ, ETS1, TXNIP, SOX2, SOCS3, TBL1X, MYH9, and CCND1, respectively.
The miRNAs, TFs, and genes retrieved from the final regulatory network were submitted to FFLtool [40] and the presence of three types of FFLs i.e., TF-FFL, miR-FFFL, and composite-FFL in the context of gene regulation were examined specifically for brain region. Since, FFL consists of three objects; TFs, miRNAs, and genes the constructed motif is called a 3-node motif. The FFL analysis depicts the interplay of miRNAs hsa-miR-155-5p (miR-155-5p) and hsa-miR-16-5p (miR-15-5p) with TFs STAT3, HNF4A, MYC, SPI1, EGR1, NANOG, and POU5F1 in inhibition and regulation of genes AXIN1, TBP, RELA, RHEB, YWHAZ, ETS1, TXNIP, SOCS3, TBL1X, MYH9, MYO1C, CCND1, MCM3, and H2AFZ.
Within the miRNA-FFL, we observed that miR-155-5p was inhibiting TF HNF4A leading to the inhibition and regulation of co-expressed genes TBL1X, MYH9, SOCS3, RHEB, MYO1C, TXNIP, AXIN1, and MCM3. Whereas when HNF4A was inhibited by miR-16-5p, it resulted in the inhibition and regulation of 7 co-expressed genes CCND1, MYH9, MCM3, SOCS3, RHEB, MYO1C, and AXIN1 (Fig. 8A). The TF-FFL revealed that when SPI1 regulates miR-16-5p, it leads to the inhibition and regulation of 9 co-expressed genes TXNIP, AXIN1, YWHAZ, MYH9, CCND1, SOCS3, MCM3, MYO1C, and RHEB (Fig. 8B).
In the composite FFL, TF/miRNA have equal contributions in the regulation and inhibition of genes by activating and suppressing each other. The composite-FFL revealed that EGR1/miR-155-5p plays a role in the inhibition and regulation of co-expressed genes RELA, TXNIP, TBL1X, YWHAZ, MYH9, CCND1, RHEB, MYO1C, AXIN1, SOCS3, and H2AFZ. STAT3/miR-155-5p inhibit and regulate each other resulting in the inhibition and regulation of co-expressed genes MCM3, RHEB, MYO1C, SOCS3, H2AFZ, and AXIN1. SPI1/ miR-155-5p has a key role in the inhibition and regulation of MYO1C, MCM3, SOCS3, H2AFZ, and AXIN1. EGR1/miR-16-5p suppress and activate each other resulting in the inhibition and regulation of co-expressed genes SOCS3, AXIN1, CCND1, RHEB, and MYO1C. STAT3/miR-16-5p mediated pathway depicts inhibition and regulation of TBP, SOCS3, MYO1C, RHEB, AXIN1, MYH9, CCND1, and MCM3. Whereas MYC/miR-16-5p mediated regulatory network inhibits and regulates MYO1C, TBP, SOCS3, AXIN1, MCM3, and RHEB (Fig. 8C).
FFL-genes/known AD-drug targets PPI network construction
At last, we overlooked whether any of the miRNA-TF mediated inhibited/regulated co-expressed genes was common among FFL motifs or not. We found genes MYH9, SOCS3, RHEB, MYO1C, TXNIP, AXIN1, MCM3, and CCND1 to be common in all FFL motifs i.e., miRNA-FFL, TF-FFL, and composite-FFL. To prove the role of the gene during AD-associated pathways, we also checked their interaction with established AD-biomarkers/drug targets such as ACHE, BCHE, APOE, APLP, COX2, SOD2, CDK5, TNFA, APP, PSEN1, PSEN2, GSK3A, GSK3B, MARK1, MARK2, BACE1, TREM2, MAP2, MAPT, ALOX5, APLP1, and AEP. We observed that all the genes have network interactions with at least one known AD-biomarkers/drug target except MCM3 (Fig. 9A). Further, subnetworks of genes making first-level interaction with known AD-biomarker/drug targets were constructed. The gene CCND1 was interacting with GSK3A, TNF, CDK5, GSK3B, APOE, and BACE (Fig. 9B). AXIN1 was interacting with PSEN1, GS3KA, GSK3B, APP, and MAPT (Fig. 9C). RHEB was interacting with TNF, GSK3A, GSK3B, and PSEN1 (Fig. 9D). MYH9 was interacting with GSK3B and TNF (Fig. 9E). TXNIP was interacting with TNF and SOD (Fig. 9F). SOCS3 and MYO1C were interacting with TNF (Fig. 9G) and GSK3B (Fig. 9H), respectively.