Elucidating Genes and Transcription Factors of Human Peripheral Blood Lymphocytes Involved in the Cellular Response upon Exposure to Ionizing Radiation for Biodosimetry and Triage: An In Silico Approach

Abstract

Background Gene expression–based biodosimetry is a promising method for estimating radiation dose following exposure. A panel of highly radio-responsive genes in human peripheral blood was used in the current investigation to create and evaluate a unique gene expression–based radiological biodosimetry method.

Methodology In human cellular research, we reviewed the literature on genes and proteins correlating to radiation response in vivo and in vitro. We looked at two publicly accessible independent radiation response gene expression profiles (GSE1977 and GSE1725) and identified the differentially expressed genes (DEGs).

Results The obtained data exhibited 42 genes with substantial differential expression, 25 of which were upregulated and 17 of which were downregulated in ionizing radiation exposure groups compared with control groups. The gene ontology enrichment analysis revealed that the hub genes are significantly involved in the regulation of the mitotic cell cycle phase transition, regulation of the mitotic cell cycle, and mitotic cell cycle checkpoint signaling. Out of the 42 DEGs, four top genes (CDK1, CCNB1, UBC, and UBB) were obtained through network centrality features. However, the multicomponent filtering procedure for radiation response genes resulted in cyclin-dependent kinase 1 (CDK1) as a critical gene in the dataset curated.

Conclusion Our findings suggest the possibility of discovering novel gene connections involved in the cellular response of human peripheral blood lymphocytes upon exposure to ionizing radiation.

Keywords

radiation - differentially expressed genes - CDK1 - gene ontology - peripheral blood

Introduction

Radiation is widely employed in various sectors, particularly medicine, for diagnostic and therapeutic purposes.[1] Simultaneously, there has been greater emphasis on the significance of knowing radiation's impacts on the human body, such as the processes underlying DNA double-strand breaks (DSBs) caused by radiation, radiation-induced signal transduction, and/or gene expression changes, all of which can cause cancer, and this has been well investigated.[2] Significant changes in the expression of genes involved in the cellular response to DNA damage were identified, consistent with DNA damage (cell proliferation, DNA damage repair, and apoptosis). In a complex signal transduction network, numerous genes participate in a cascade of actions in response to radiation and ensuing DNA damage.[3] In vitro irradiated cells have shown such alterations in gene expression.[4] Gene expression variations are early markers of response to radiation exposure.[5] Several gene expression studies have already been proven to be dose dependently regulated.[6] [7] Gene expression studies have been deployed for radiation biodosimetry.[5] [8] [9] Gene expression alterations in human peripheral blood cells, especially lymphocytes, have been proven helpful in high-throughput, minimally invasive biodosimetry.[5]

The available literature has shown that gene expression signatures can accurately predict radiation status and distinguish radiation exposure levels; different analytical approaches may influence gene selection, resulting in a large discrepancy in gene expression profile reproducibility.[10] [11] [12] Furthermore, the impact of other confounding factors on gene expression profiles for radiation dose estimates, such as gender, smoking, chronic irradiation, or inflammation,[13] and interindividual variation[14] has been reported. At the same time, there have been some differences in performance among laboratories utilizing different signatures and measurement procedures.[9] [15] As a result, there is still much work to produce radiation-specific gene expression-based biodosimetry. A single change in gene expression is unlikely to represent an individual response to ionizing radiation due to the activation of numerous signal transduction pathways by this radiation, making it difficult to calculate radiation dose accurately.[16]

In this study, we tried to create an objective strategy for identifying significant radiation-responsive genes with a considerable impact on radiation biodosimetry. A group of radiation-responsive genes[17] [18] was used to improve the robustness of gene expression-based biodosimetry. We used a multicomponent filtering procedure based on a system biology approach suggested in our previous study[19] to cut down this investigation's list of gene candidates. We narrowed our search to interacting networks, assuming that useful biomarkers would be embedded in significant pathways or networks containing numerous genes essential to the outcome. The estimated networks were compared with a reliable gene interaction network to determine the accuracy of the network modeling.

Methodology

Radiation-Responsive Gene Data Collection

From the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), two independent radiation response gene expression profiles (GSE1977 and GSE1725), consisting of 87 samples of ionizing radiation and 87 mock samples, were downloaded and used as discovery datasets to find differentially expressed genes (DEGs). The Affymetrix Human Genome U95 version 2 Array (HG-U95Av2) microarray platform was used to collect each of these datasets. We performed data preprocessing on the collected dataset by quintile normalization and Z-score transformation.[20] The Kruskal–Wallis U test, Student's t-test, and linear models for the limma package in R were used to perform the differential expression analysis, and Benjamini–Hochberg correction techniques were used to reduce the false discovery rate (FDR).[21] We then went on to discover the DEGs and used those DEGs to investigate biological pathways, enrichment analysis of interaction networks, and functional annotation of genes.

Gene Ontology and Molecular Pathways Analysis of DEGs

The biological significance of the radiation-responsive genes is established using gene ontology (GO) and pathways enrichment analysis.[22] GO keyword annotation of DEGs was used to highlight the biological networks of gene change with respect to three different aspects: biological processes (BPs), molecular functions (MFs), and cellular components. Using ShinyGO v0.76 (GO online program), we performed a GO enrichment analysis on the overlapped DEGs.[23] For pathway enrichment research, the Enrichr,[24] a comprehensive gene set bioinformatics web application, was employed to look into the shared molecular signaling pathways between DEGs. We employed pathway enrichment analysis with six databases, including Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, Wiki, Panther, BioCarta, and BioPlanets, to identify the biological network pathways of radiation-responsive genes. The top-mentioned pathways were chosen using the normal criteria of p-values less than 0.05.

Construction of Protein–Protein Interactions Network and Hub Gene Screening

The extensive and intricate regulatory network known as protein–protein interactions (PPIs) has been implicated in physiological and pathological processes.[25] Identification and annotation of PPIs that fully comprehend the functionality of a healthy state and the evolution of the disease from molecular etiology to biochemical cascades are heavily reliant on system biology.[26] We constructed the PPI network of upregulated and downregulated DEGs using the STRING database (version 11.5) repository, taking into account factors such as active interaction sources, including text mining, experiments, databases, co-expression, neighborhood, gene fusion, and co-occurrence as well as functional and physical protein associations, limiting the analysis to “Homo sapiens,” and medium confidence scores greater than 0.4.

Further, we used the Cytoscape tool (version 3.9.1) to visualize the PPI network. Using topological parameters, such as a degree of more than 18, we could identify the top 10 hub genes, which are highly interconnected. As a result, we could understand the molecular mechanism of the human radiation response. The hub genes were selected using the cytoHubba plugin; a Cytoscape plugin was used to determine the hub proteins or genes in the PPI network. Five methods were employed (degree, maximum neighborhood component [MNC], radiality centrality, betweenness centrality, and closeness centrality) to identify and evaluate key genes.[27]

Degree: The most elementary characteristic of a node is its degree (Deg), which represents the number of links the node has to other nodes in the network. It is given by the following equation:

Betweenness centrality (BC): Betweenness is a centrality measure of a vertex within a graph. For a graph G (V, E) with n vertices, the betweenness BC (ν) of a vertex ν is defined as

where σ_st is the number of shortest paths from s to t and σ_st (ν) is the number of shortest paths from s to t that pass through a vertex ν. A similar definition for “edge betweenness” was given by Girvan and Newman. Nodes with a higher betweenness lie on a larger number of shortest paths in a network.

Closeness centrality: CC (k) determines the rate at which information is disseminated from a node to the other nodes connected to it. It is determined as the average of the shortest path length from a node to all other nodes in the network and represents a node's closeness to all other nodes in the network. CC (k) of a node m is mathematically defined as

MNC: The neighborhood of a node ν, nodes adjacent to ν, induces a subnetwork N (ν). The score of node ν, MNC (ν), is defined to be the size of the maximum connected component of N (ν). The neighborhood N (ν) is the set of nodes adjacent to ν and does not contain node ν:

where M C (ν) is a maximum connected component of G [N (ν)] and G [N (ν)] is the induced subgraph of G by N (ν).

Radiality: The radiality of a node ν is calculated by computing the shortest path between the node ν and all other nodes in the graph. Radiality centrality is given by

where d is diameter of graph G with n vertices and d (ν, w) is the distance between vertex ν and w.

Interactom Analysis of Hub Gene

Hub genes, referred to be necessary modules with the highest degree of interconnectedness, are crucial for comprehending the paths of biological networks. Through an interactomics analysis, which depicts molecular interaction networks with physical linkages among neighbors, the functional significance of the cellular map in identifying drug targets is investigated.[28] We were able to identify the top hub gene's close neighborhood ranking network for addressing the novel function of the gene in the context of biological reactions using the Biological General Repository for Interaction Datasets (BioGRID 4.4). When selecting hub gene networks, physical interactions and degree evidence were considered (≥ 18).

Recognition of Transcriptional Factors with Connecting PPI Network

Transcription factors regulate the complex protein network of PPIs that initiates and regulates the transcription of genetic material and are essential components of numerous biological systems.[29] Using the X2K online tool[30] (regulatory networks platform), we selected the top transcriptional factors based on the hypergeometric p-value from the ChIP-seq experiments (ChEA) database. The X2K web tool creates inferred networks of transcription factors with connecting PPI, leading to upstream regulatory pathways, using the signatures of radio-responsive genes. Additionally, using the Genes2Networks (G2N) method, we identified the proteins that physically interact with these transcription factors.[31] Using PPIs or protein complexes that have been empirically verified in the disciplines of genomics and proteomics, G2N is a potent command-line and web-based tool that deciphers radio-responsive genes. To better understand cell signaling pathways, this tool aids biologists in filtering transcription factors connected to protein network complexes.

Identification of Protein Kinase Connecting with Transcription Factors and PPIs

Enzymes called protein kinases (PTKs) activate phosphorylated target proteins in the cell to dynamically control signaling proteins. PTKs were discovered using the X2K[30] (KEA module) program. The radio-responsive gene lists of mammalian proteins and the PTKs most likely to phosphorylate them can be connected with the help of the command-line program KEA.[32] In the extended subnetwork, we also constructed a regulatory kinase–substrate network containing phosphorylated transcription factors, PTKs, and PPIs. The sources of the kinase–substrate network is the Human Protein Reference Database (HPRD), PhosphoSite, phospho.ELM, NetworKIN, and Kinexus (www.kinexus.ca).

Results

Identification of Radiation-Induced DEGs

This study used microarray datasets to find therapeutic biomarkers for human radiation biodosimetry and triage using integrative bioinformatics and a machine learning approach. First, from the GEO database of the National Center for Biotechnology Information (NCBI), we obtained the ionizing radiation-induced gene expression profiles of GSE1977 and GSE1725, which comprised a total of 174 samples (87 samples exposed to ionizing radiation and 87 samples exposed to mock treatment [Mock]). Using the Benjamin–Hochberg FDR, R language was used to study the radiation-exposed gene expression profile microarray dataset and limma packages. A p-value of less than 0.05 and an absolute log-fold value of 1 were used to assess the radiation response in DEGs ([Fig. 1]). Finally, when ionizing radiation exposure groups were compared with Mock exposure groups, we detected 42 genes with substantial differential expression, 25 of which were upregulated and 17 were downregulated (absolute value of |logFC| > 1.0; p = 0.05). [Table 1] shows, for each significant p-value, the top upregulated and downregulated DEGs in relation to radiation exposure.

Fig. 1 Differentially expressed genes volcano map, screening standards. Red denotes differentially expressed genes that are upregulated with |logFC| >1 and p < 0.05. Blue indicates differentially expressed genes with |logFC| < –1 and p < 0.05 that are downregulated.

Table 1 The obtained top upregulated and downregulated genes among the differentially expressed genes

Gene symbol	Gene title	|logFC|	Regulation
ACTA2	Actin, α 2, smooth muscle, aorta	1.012	Up
ATF3	Activating transcription factor 3	1.608	Up
CD70	CD70 molecule	1.052	Up
CDKN1A	Cyclin-dependent kinase inhibitor 1A	2.129	Up
DDB2	Damage-specific DNA binding protein 2	1.25	Up
FAS	Fas cell surface death receptor	1.413	Up
FDXR	Ferredoxin reductase	1.227	Up
FOXG1	Forkhead box G1	1.247	Up
GADD45A	Growth arrest and DNA damage inducible α	2.171	Up
FOXG1	Forkhead box G1	1.247	Up
HRAS	HRas proto-oncogene, GTPase	1.43	Up
LAMP3	Lysosomal-associated membrane protein 3	1.081	Up
MDM2	MDM2 proto-oncogene	1.863	Up
MIR1204///PVT1	microRNA 1204///Pvt1 oncogene (nonprotein coding)	1.227	Up
PCNA	Proliferating cell nuclear antigen	1.146	Up
PLXNB2	Plexin B2	1.332	Up
PODXL	Podocalyxin-like	1.39	Up
PPM1D	protein phosphatase, Mg2 +/Mn2+ dependent 1D	1.296	Up
PRKAB1	Protein kinase AMP-activated noncatalytic subunit β 1	1.303	Up
PTP4A1	Protein tyrosine phosphatase type IVA, member 1	1.276	Up
SMAD5	SMAD family member 5	1.196	Up
TNFRSF10B	TNF receptor superfamily member 10b	1.219	Up
TNFSF9	Tumor necrosis factor superfamily member 9	1.401	Up
TP53I3	Tumor protein p53 inducible protein 3	1.523	Up
TP53TG1	TP53 target 1 (nonprotein coding)	1.235	Up
TRIAP1	TP53-regulated inhibitor of apoptosis 1	1.552	Up
XPC	XPC complex subunit, DNA damage recognition and repair factor	1.62	Up
ZBTB20	Zinc finger and BTB domain containing 20	1.53	Up
ARHGAP11A	Rho GTPase activating protein 11A	–1.023	Down
AURKA	Aurora kinase A	–1.525	Down
BUB1	BUB1 mitotic checkpoint serine/threonine kinase	–1.054	Down
CCNB1	Cyclin B1	–1.447	Down
CDC20	Cell division cycle 20	–1.815	Down
DLGAP5	DLG-associated protein 5	–1.104	Down
HIST1H2BJ	Histone cluster 1, H2bj	–1.045	Down
HIST1H3D	Histone cluster 1, H3d	–1.088	Down
HK2	Hexokinase 2	–1.054	Down
HMMR	Hyaluronan-mediated motility receptor	–1.185	Down
KIF14	Kinesin family member 14	–1.083	Down
KIF23	Kinesin family member 23	–1.527	Down
MYC	v-myc avian myelocytomatosis viral oncogene homolog	–1.236	Down
PLK1	Polo-like kinase 1	–1.618	Down
SLC35F6///CENPA	Solute carrier family 35 member F6///centromere protein A	–1.407	Down
UBE2C	Ubiquitin conjugating enzyme E2 C	–1.035	Down

Gene Ontology Enrichment Analysis

The goal of GO is to identify significant biological changes or gene co-expression patterns and to enable their evolution and interoperability between human genome databases by using the GO term concept to query and retrieve genes based on their common biological functions.[22] ShinyGO v0.76 was used to conduct GO enrichment studies for both upregulated and downregulated DEGs to identify the biological functions of radiation response genes. The three different ontologies were created using the GO term database as an annotation source for the GO study (BP, cellular component, and MF). At a p-value cutoff (FDR) lower than 0.05 and for a few human species, we have demonstrated GO enrichment analysis of upregulated and downregulated DEGs within three categories (BP, MF, and cellular components). We revealed that DEGs were significantly enriched in the regulation of mitotic cell cycle phase transition, regulation of mitotic cell cycle, mitotic cell cycle process, mitotic cell cycle, mitotic cell cycle process, and mitotic cell cycle checkpoint signaling from the GO enrichment results of the BP category ([Fig. 2A, E]). The centrosome, microtubule organization center, microtubule cytoskeleton, spindle, spindle midzone, and microtubule-associated complex are further DEGs ([Fig. 2B, F]). We have demonstrated that DEGs were predominantly enriched in kinase binding, PTK binding, death receptor activity, DNA insertion or deletion binding, anaphase-promoting complex binding, and damaged DNA binding for the MF category ([Fig. 2C, G]).

Fig. 2 The gene ontology (GO) analysis of (A–C) downregulated and (E–G) upregulated differentially expressed genes. The KEGG pathway analysis of (D) downregulated and (H) upregulated differentially expressed genes. Significantly enriched biological processes/cellular component/molecular function/KEGG pathways were represented by high (red) to low (green) fold enrichment.

Identification of Crucial Signaling Pathways

Pathway analysis, which attempts to comprehend high-throughput biological data and identify biological signaling pathways that are enriched by the radiation response genes in human blood cells, is one of the most significant methods employed in omics research in the life sciences. The web-based bioinformatics tool ShinyGO v0.76 is used to perform pathway enrichment analysis of upregulated and downregulated DEGs. The top signaling pathways were taken into consideration, and we used the significance of a p-value (FDR) lower than 0.05 as a standard criterion for choosing pathway analysis related with upregulated and downregulated DEGs in ionizing radiation-induced response ([Fig. 2D, H]).

Further, six databases, including KEGG, Reactome, Wiki, Panther, BioCarta, and BioPlanets, to identify the biological network pathways of radiation-responsive DEGs (upregulated and downregulated together). According to the KEGG pathway database, we identified critical biological pathways involved in radiation response, including cell cycle, oocyte meiosis, Epstein–Barr virus infection, human immunodeficiency virus 1 infection, and pathways of neurodegeneration. Human T-cell leukemia virus 1 infection, human cytomegalovirus infection, cellular senescence, MAPK signaling pathway ([Supplementary Fig. 1A], available in the online version only). Reactome pathway revealed cell cycle, mitotic, M phase, immune system, G2/M transition, mitotic G2-G2/M phases, gene expression, cyclin A/B1-associated events during G2/M transition, and mitotic prometaphase RHO GTPase effectors according to the database of the Reactome pathway ([Supplementary Fig. 1B], available in the online version only).

The data investigated from the WIKI pathway database indicating cell cycle, retinoblastoma gene in cancer VEGFA-VEGFR2 signaling pathway, B-cell receptor (BCR) signaling pathway, TGF-β signaling pathway, EGF/EGFR signaling pathway, miRNA regulation of DNA damage response, DNA damage response, genotoxicity pathway, TNF-α signaling pathway ([Supplementary Fig. 1C], available in the online version only). The pathway analysis of the panther database revealed the p53 pathway, B-cell activation, cholecystokinin receptor (CCKR) signaling map ST, Ras pathway, p53 pathway feedback loops 2, apoptosis signaling pathway, platelet-derived growth factor (PDGF) signaling pathway, p38 MAPK pathway, fibroblast growth factor (FGF) signaling pathway, EGF receptor signaling pathway associated with significantly DEG interaction in ionizing radiation-induced response ([Supplementary Fig. 1D], available in the online version only).

Through the results from the BioCarta pathway database, we identified the most significant biological function associated signaling pathways, including BCR signaling pathway, interleukin-2 (IL-2) receptor beta chain in T-cell activation, NFAT and hypertrophy of the heart, Erk1/Erk2 MAPK signaling pathway, cell cycle: G2/M checkpoint, T-cell receptor signaling pathway, angiotensin II mediated activation of JNK pathway via Pyk2-dependent signaling, MAPKinase signaling pathway, Fc epsilon receptor I signaling in mast cells, fMLP-induced chemokine gene expression in HMC-1 cells ([Supplementary Fig. 1E], available in the online version only). The data obtained from the Bioplanet pathway database indicate IL-2 signaling pathway, cell cycle, T-cell receptor regulation of apoptosis, immune system, oocyte meiosis, TGF-β signaling pathway, antigen-activated BCR generation of second messengers, mitotic G1-G1/S phases, adaptive immune system, cyclin A/B1-associated events during G2/M transition pathway interact with the number of DEGs in ionizing radiation-induced response ([Supplementary Fig. 1F], available in the online version only).

Network Construction of PPIs

The PPI networks are fundamental for comprehending both normal and pathological cell physiology. They provide crucial information on the regulation and synchronization of biological signaling pathways for almost all cellular events. We built the PPI's functional and physical network using the STRING database (version 11.5) to investigate the PPIs among the DEGs of ionizing radiation response. With the threshold set as interaction score greater than 0.9, an extended PPI network of DEGs has 702 nodes and 2,585 edges. Additionally, a core (giant) network was recovered and displayed using Cytoscape software (version 3.9.1), consisting of 645 nodes linked by 2,549 edges ([Fig. 3]).

Fig. 3 A giant network was constructed and displayed using Cytoscape software (version 3.9.1), consisting of 645 nodes linked by 2,549 edges. The nodes with large size represent the low values of degree (low value with orange and high value with blue color) and the edge size was mapped based on the co-occurrence of the node in the network.

Additionally, the hub proteins were chosen using cytoHubba's five classification techniques. [Table 2] lists the top 10 proteins, and [Fig. 4] illustrates how the top 10 proteins from each approach were overlapped to identify four key proteins. CDK1 is distinguished by a large degree value of 82, a high BC value of 0.1234, and a large CC value of 0.3457, whereas CCNB1 has a second large degree value of 72 and a seventh high BC value of 0.0595; UBC takes the third position with a degree value of 59 and BC value of 0.0861; similarly, UBB inherits a fourth large degree value of 52 and BC value of 0.0532.

Table 2 The top 10 proteins identified based on the five classification methods in cytoHubba

Sl. no.	BC	CC	Degree	MNC	Radiality
1	CDK1	CDK1	CDK1	CDK1	CDK1
2	RPS6KB1	CCNB1	CCNB1	CCNB1	CCNB1
3	EIF4E	UBC	UBC	UBC	MYC
4	UBC	UBB	CCNA2	CCNA2	UBC
5	RXRA	MYC	CCNB2	CCNB2	UBB
6	RHOA	CCNA2	UBB	UBB	CDKN1A
7	CCNB1	CCNB2	CDC20	CDC20	CDK4
8	SHC1	CDC20	TOP2A	TOP2A	RPS6KB1
9	POLR1C	CDKN1A	AURKB	AURKB	MDM2
10	UBB	CDK4	BUB1	BUB1	RXRA

Fig. 4 Betweenness centrality, closeness centrality, degree, maximum neighborhood component (MNC), and radiality were used to select hub proteins in cytoHubba tool.

Navigability

The efficacy of information transport within the network and general navigability is measured by path length. It specifies the quantity of steps/edges along the shortest pathways for all conceivable network node pairings.[33] The ionizing radiation response of the DEG PPI network is shown as a distribution of path lengths ([Supplementary Fig. S2], available in the online version only). The typical path length of the present network is 4.481. Less path lengths fall into the extreme top categories (path lengths 8 and 9) compared with the lower extremities (path lengths 1 and 2), which suggests that the majority of the proteins in the graph can only be connected by a few numbers of paths. Short pathways are typically preferred since they allow for quick information transport at a lower cost.[34] However, they may be susceptible to minor perturbations, which can swiftly spread throughout the network. Average shortest path length (ASPL), a topological feature of scale-free networks, is a commonly used metric for analyzing disease-associated genes and is used to demonstrate whether the genes are functionally cohesive. The intuition behind this is that the genes that are functionally related will be closer to each other in the interactome. Typically, a low ASPL is reported between candidate genes compared with randomly selected genes, along with an empirical p-value of statistical significance. ASPL is commonly used in gene prioritization. It has been used to study pathways in complex diseases to connect putative causal genes to target genes and to extract disease-associated network modules.[35]

Recognition of Hub Genes with Interactomics Analysis and Significance of Key Nodes in the Network

Further, we selected the nodes with BC and/or degree values larger than the mean plus standard deviation.[36] Sixty-two nodes with large degree values ([Supplementary Table S1], available in the online version only) and 60 with high BC values ([Supplementary Table S2], available in the online version only) were determined. Also, 26 nodes were defined with both large degrees and high BC values ([Table 2]).

We also employed the BioGRID to investigate interactomics based on the five categorization techniques in cytoHubba to address the unique function of the gene in the context of biological responses (degree, BC, CC, MNC, and radiality). This analysis focused on the top four hub gene interactions with closely ranking neighborhood proteins. A sophisticated protein function context can be obtained using interactomics, a key component of modern systems biology. The interactomics study only considers the expected physical network of PPIs with a score greater than 0.5. Finally, we assessed the interactome networks of hub genes, specifically CDK1, CCNB1, UBC, and UBB ([Fig. 5]) in therapeutic biomarkers for radiation biodosimetry and triage in humans. [Table 3] shows the functional significance of these identified hub genes that contribute to understanding the cellular response to ionizing radiation. [Table 3].

Fig. 5 (A–D) Based on physical interaction and degree evidence of the top four hub gene interactions, interactome network analysis for CDK1, CCNB1, UBC, and UBB using the Biological General Repository for Interaction Datasets (BioGRID).

Table 3 Functional significance of top 10 hub genes identified based on node degree distribution that contribute to understanding the cellular response to ionizing radiation

Sl. no.	Genes	Biological processes	Molecular function	Cellular component
1	CDK1	DNA repair, DNA replication, Fc-epsilon receptor signaling pathway, G1/S transition of mitotic cell cycle, G2/M transition of mitotic cell cycle	RNA polymerase II carboxy-terminal domain kinase activity, cyclin-dependent protein serine/threonine kinase activity, protein binding, protein kinase activity, protein serine/threonine kinase activity	Centrosome, cytoplasm, cytosol, extracellular vesicular exosome, membrane
3	CCNB1	G1/S transition of mitotic cell cycle, G2/M transition of mitotic cell cycle, anaphase-promoting complex-dependent proteasomal ubiquitin-dependent protein catabolic process, mitotic cell cycle, mitotic metaphase plate congression	Patched binding, protein binding, protein kinase binding	Centrosome, condensed nuclear chromosome outer kinetochore, cytoplasm, cytosol, nucleoplasm
3	UBC	DNA damage response, signal transduction by P53 class mediator resulting in cell cycle arrest, DNA repair, Fc-epsilon receptor signaling pathway, G1/S transition of mitotic cell cycle, G2/M transition of mitotic cell cycle	Poly(A) RNA binding, protease binding, protein binding	Cytosol, endocytic vesicle membrane, endosome membrane, extracellular vesicular exosome, nucleoplasm
4	CCNA2	G2/M transition of mitotic cell cycle, Ras protein signal transduction, mitotic G2 DNA damage checkpoint, mitotic cell cycle	Protein binding	Cytoplasm, nucleoplasm, nucleus
5	CCNB2	G2/M transition of mitotic cell cycle, mitotic cell cycle, mitotic nuclear envelope disassembly, regulation of cell cycle	Protein binding	Centrosome, cytosol, microtubule cytoskeleton, nucleoplasm
6	UBB	DNA damage response, signal transduction by P53 class mediator resulting in cell cycle arrest, DNA repair, Fc-epsilon receptor signaling pathway, G1/S transition of mitotic cell cycle, G2/M transition of mitotic cell cycle	Protein binding	Cytosol, endocytic vesicle membrane, endosome membrane, extracellular vesicular exosome, mitochondrion
7	CDC20	Activation of anaphase-promoting complex activity, anaphase-promoting complex-dependent proteasomal ubiquitin-dependent protein catabolic process, mitotic cell cycle	Enzyme binding, protein C-terminus binding, protein binding	Anaphase-promoting complex, cytoplasm, cytosol, nucleoplasm, nucleus, spindle
8	TOP2A	ATP catabolic process, DNA ligation, DNA topological change, DNA unwinding involved in DNA replication, apoptotic chromosome condensation	DNA binding, DNA binding, bending, DNA topoisomerase type II (ATP-hydrolyzing) activity, DNA-dependent ATPase activity, chromatin binding	DNA topoisomerase complex (ATP-hydrolyzing), centriole, nuclear chromosome, nucleoid, nucleolus
9	AURKB	Abscission, anaphase-promoting complex-dependent proteasomal ubiquitin-dependent protein catabolic process, attachment of spindle microtubules to kinetochore, cellular response to UV, cleavage furrow formation	Histone serine kinase activity, protein binding, protein serine/threonine kinase activity, protein serine/threonine/tyrosine kinase activity	Chromosome passenger complex, condensed chromosome, centromeric region, condensed nuclear chromosome, centromeric region, cytosol, midbody
10	BUB1	Mitotic cell cycle, mitotic cell cycle checkpoint, mitotic spindle assembly checkpoint, regulation of chromosome segregation, regulation of sister chromatid cohesion, spindle assembly checkpoint	Protein binding, protein kinase activity	Condensed chromosome kinetochore, cytoplasm, cytosol, kinetochore, membrane, nucleoplasm

Transcriptional Regulatory Network Analysis of DEGs

Transcription factors are essential molecules that directly maintain gene regulatory networks and regulate gene expression. The suppression or activation of transcription factors is crucial for several vital biological and cellular processes. The dysregulated transcription factors implicated in the emergence of diseases linked to ionizing radiation to control the DEGs. Using the X2K online tool from the ChIP-seq experiments (ChEA) database, we were able to identify the novel transcriptional factors regulating the expression of DEGs in cells exposed to ionizing radiation. Based on the hypergeometric p-value, we performed transcription factor enrichment analysis and determined the top 20 transcription factor candidates, including E2F transcription factor 4 (E2F-4); nuclear transcription factor Y, β (NF-YB, HAP3); upstream binding transcription factor, RNA polymerase I (UBTF); TATA box binding protein (TBP) associated factor (TAF1); CAMP responsive element binding protein 1 (CREB1); breast cancer 1, early onset (BRCA1); MYC-associated factor X (MAX); forkhead box M1 (FOXM1); nuclear transcription factor Y, α (NF-YA); SIN3 transcription regulator family member A (SIN3A); Spi-1 proto-oncogene (SPI1); E2F transcription factor 1 (E2F1); transcription factor 3 (TCF3); zinc finger protein 384 (ZNF384); activating transcription factor 2 (ATF2); chromodomain helicase DNA binding protein 1 (CHD1); zinc finger, MIZ-type containing 1 (ZMIZ1); promyelocytic leukemia (PML); runt-related transcription factor 1 (RUNX1), and signal transducer and activator of transcription 3, acute-phase response factor (STAT3), which may be shown altering gene function in the development of diseases caused on by exposure to ionizing radiation ([Fig. 6A]). Additionally, we identified proteins that physically interact with these transcription factors using the G2N algorithm to assess PPIs and transcription factor connections. The associated transcription factor regulatory network and the proteins that interact with them physically and functionally were shown based on the degree of the nodes ([Fig. 6B]). Pink nodes represented the transcription factors, while gray nodes represented the proteins connected to them.

Fig. 6 Transcription factor enrichment analysis with the protein–protein interaction (PPI) network. (A) Top transcription factors 19 regulate differentially expressed gene (DEG) regulation. (B) Enriched transcription factors physically interact with protein using Gene2Networks (G2N) algorithm. Pink nodes represent transcription factors and proteins connect 21 with them in gray.

Enrichment Analysis of the Upstream Regulatory Pathway of Kinases

PTKs control mRNA translation, cell proliferation and survival regulation, as well as the nuclear genomic response to cellular stresses, which comprise sensors and effectors from signal transduction cascades.[37] There are numerous disorders that have been associated with the aberrant deregulation of PTK activity. Since most PTKs are involved in promoting cell survival, proliferation, and migration when constitutively overexpressed or active, they are also connected to cancer development.[38] We used the Kinase Enrichment Analysis module of X2K to determine the most significant PTKs connected to radiation-induced illnesses and examine the prospective therapeutic kinase targets ([Fig. 7A]). The findings of our kinase enrichment analysis revealed that casein kinase 2, α 1 polypeptide (CSNK2A1); cyclin-dependent kinase 4 (CDK4); cyclin-dependent kinase 1 (CDK1); mitogen-activated protein kinase 14 (MAPK14); casein kinase 2 α 2 (CK2ALPHA); cyclin-dependent kinase 2 (CDC2); ATM serine/threonine kinase (ATM); mitogen-activated protein kinase 1 (MAPK1); V-akt murine thymoma viral oncogene homolog 1 (AKT1); glycogen synthase kinase 3β (GSK3B); PTK, DNA-activated, catalytic polypeptide (DNAPK); ribosomal protein S6 kinase, 90kDa, polypeptide 3 (RPS6KA3); ABL proto-oncogene 1, nonreceptor tyrosine kinase (ABL1); cyclin-dependent kinase 7 (CDK7); PTK, DNA-activated, catalytic polypeptide (PRKDC); protein kinase C, α (PRKCA); CK2; mitogen-activated protein kinase 3 (MAPK3), and checkpoint kinase 1 (CHEK1) are the top PTKs associated with intracellular signaling pathways in radiation-induced cellular abnormalities. We also used the Human Protein Reference Database (HPRD), PhosphoSite, phospho.ELM, NetworKIN, and Kinexus to construct a kinase–substrate network. An expanded subnetwork of transcription factors and intermediate proteins was discovered in our bioinformatic investigation as a regulatory kinase–substrate network, where substrates are phosphorylated by active PTKs ([Fig. 7B]).

Fig. 7 The enrichment analysis of kinase with transcription factors and protein–protein interactions network. (A) Top 20 kinase is displayed in a bar graph from kinase–substrate interaction databases, (B) Kinase mediates upstream regulatory network. Red nodes represent the top transcription factors, blue nodes represent protein kinase, green network edges represent kinase–substrate phosphorylation interactions, gray edges represent physical protein–protein interactions network, and red nodes show transcription factors.

Discussion

To measure normal tissue damage in radiation oncology and biodosimetry in nuclear incidents and accidental radiation exposures, biological markers of ionizing radiation exposure in human populations are of major interest. There is a substantial likelihood that a large population may be exposed in the event of radiation accidents or unintentional exposures. The medical management of such a mass casualty dose evaluation is of utmost importance. For instance, when transporting collected blood to distant biodosimetry facilities for dose assessment, sample processing may be delayed due to lack of preparation or insufficient capacity if samples are not collected at controlled temperatures. The hallmarks of biological reactions to ionizing radiation include DNA damage and repair. The DNA repair machinery and cell cycle checkpoints are stimulated in response to ionizing radiation-induced DNA damage, followed by downstream cellular responses, including apoptosis, a process through which damaged cells are removed. Priority candidates for radiation biodosimetry include genes involved in cellular DNA damage response and repair processes, including DNA repair, cell cycle processes, and apoptosis. It has been reported that transcript changes induced by ionizing radiation can persist over the days and identified nearly 260 radio-responsive proteins for biodosimetric applications.[39] [40] [41] The effect of confounders, such as diseases, inflammation and stress, as well as demographic factors (ethnicity, age, gender), on specific gene expression markers and related issues, such as the identification of individuals who have not been exposed to radiation, are still being discussed to evaluate radiation-responsive biomarkers and their usefulness in radiation biodosimetry in real-world human exposure scenarios. As mentioned earlier, there is still scope for improvement in the translation of ex vivo to in vivo and from animal models to humans. It is also necessary to examine the gene expression response to different radiation qualities, dose rates, and exposures to internalized radionuclides and to comprehend the impact of potential confounders. This study uses in silico methods to address these confounding problems and reduce confounder effects.

The four top genes (CDK1, CCNB1, UBC, and UBB) obtained based on network centrality features from our multicomponent filtering procedure for radiation-response genes resulted in CDK1 as a key gene among the dataset curated. The essential mechanism for managing cellular homeostasis when there is radiation exposure is the activation of cell cycle checkpoints. Several checkpoints are crucial for monitoring and controlling the course of the cell cycle as well as the stability of the genome in response to DNA damage or other stresses. CDK1 is a central component for control of the cell cycle and has been revealed to stimulate the replication of DNA and may increase chemoresistance.[42] CDK1 has emerged as the key regulator of the mammalian cell cycle. While most CDKs and cyclins are essentially replaceable, their functions in vivo cannot be substituted by other closely related CDKs.[43] [44] Numerous biological procedures, such as the G2/M transition, checkpoint activation, DNA repair, and DNA replication, are maintained by CDK1 and are essential for cell survival.[42] Hence, a recognized biomarker for radiation exposure is CDK1, a classic marker of the DNA damage response. CDK1 and related genes were primarily enriched in the mitotic nuclear division, sister chromatid segregation, G2/M transition, mitotic cell cycle checkpoint, organelle organization, regulation of cell cycle, anaphase-promoting complex-dependent catabolic process, and cell division ([Fig. 2]). The ionizing radiation dose between 0.05 and 0.5 Gy for CDKN1, FXDR, and SESN1 significantly altered the gene expression in a recent study.[45] CDKN1A is a gene that binds to and inhibits the CDK1, CDK2, and CDK4/6 complexes to control or halt cell cycle progression at the G1 and S phases. Through P53-dependent signaling pathway, this activity is regulated through ATM and CHK2 to control the cell cycle checkpoints in response to ionizing radiation. Still, it is also known to be controlled by P53-independent pathways to encourage the assembly of these complexes, especially in the G2/M phase of the cell cycle.[46] The expression of the CDKN1A gene increased dose dependently, and an in vitro dose-response curve was generated for patients undergoing interventional procedures.[8] According to findings of Qin et al,[47] CDK1 increases mitochondrial bioenergetics to satisfy the increased cellular fuel requirement for DNA repair and cell survival during genotoxic stress. Additionally, it has been discovered that cyclin B1/CDK1 can phosphorylate and activate MnSOD and p53 in mitochondria to improve cell survival.[48]

All multicellular organisms require precise control of cell proliferation for healthy development and survival. All forms of cancer share the core trait of dysregulated cell growth. Early embryonic extracts and CDK function analysis revealed that variations in CDK activity are sufficient to control cell cycle transitions without transcription. However, most eukaryotic cells undergo dynamic changes in gene expression patterns as the cell cycle progresses. The transcription of more than 70% of regularly expressed genes continues to oscillate after B-type cyclins are eliminated, according to the examination of transcriptional patterns in yeast.[49] Proteins known as transcription factors attach to specific short DNA sequences in the control regions of genes, controlling the transcription of those genes either positively or negatively. Changes in transcription factors can result in human disorders that fall into three main categories and play critical roles in many crucial cellular processes, cancer and hormonal responses that are not normal, and developmental issues. The present study's computational analysis of the transcriptional network ([Fig. 6A]) found that transcription factors are also involved in the radiation-induced altered gene expression. Regression-based approaches are just one of the recent computational techniques suggested to find transcription factor binding locations based on gene expression data. Based on the obtained data, it has been observed that CDK mainly interacts with cancer-related transcription factors like E2F, STAT3, BRCA1, FOXM1, and TAF1. The activation and inactivation of E2F are closely related to CDK activation. These links enable the temporal plans of gene expression to be connected with the periodic oscillations in the CDK cycle in healthy cells. Many genes involved in cell proliferation, especially those involved in the cell cycle's progression through the G1 and into the S phases, are regulated by E2F transcription factors. It has been demonstrated that improper S phase entry and apoptosis are induced by the deregulated expression of genes belonging to the E2F family.

Conclusion

We have shown a way to objectively identify a core network of critical interacting genes crucial to radiation response in bioinformatics. We postulated that merging a variety of datasets would increase the likelihood of finding interacting genes that are particularly crucial to radiation response. The maintenance of genomic integrity after the occurrence of DNA damage depends on the DNA damage response system. CDKs have become essential regulators of the DNA damage response, as was previously discussed. This differs from the downstream function of suppressed CDKs in checkpoint regulation and cell cycle arrest. Although our knowledge of CDK's roles in DNA repair is constantly developing, we just learned more about how they specifically respond to DNA damage. We also believe that these genes are good candidates for biomarker research.

Conflict of Interest

None declared.

• References

• 1 Vaiserman A, Koliada A, Zabuga O, Socol Y. Health impacts of low-dose ionizing radiation: current scientific debates and regulatory issues. Dose Response 2018; 16 (03) 1559325818796331
• 2 Hall EJ, Brenner DJ. Cancer risks from diagnostic radiology. Br J Radiol 2008; 81 (965) 362-378
• 3 Jen KY, Cheung VG. Transcriptional response of lymphoblastoid cells to ionizing radiation. Genome Res 2003; 13 (09) 2092-2100
• 4 Velegzhaninov IO, Shadrin DM, Pylina YI. et al. Differential molecular stress responses to low compared to high doses of ionizing radiation in normal human fibroblasts. Dose Response 2015; 13 (01) 14-058
• 5 Knops K, Boldt S, Wolkenhauer O, Kriehuber R. Gene expression in low- and high-dose-irradiated human peripheral blood lymphocytes: possible applications for biodosimetry. Radiat Res 2012; 178 (04) 304-312
• 6 Dressman HK, Muramoto GG, Chao NJ. et al. Gene expression signatures that predict radiation exposure in mice and humans. PLoS Med 2007; 4 (04) e106
• 7 Kultova G, Tichy A, Rehulkova H, Myslivcova-Fucikova A. The hunt for radiation biomarkers: current situation. Int J Radiat Biol 2020; 96 (03) 370-382
• 8 Visweswaran S, Joseph S, Dhanasekaran J. et al. Exposure of patients to low doses of X-radiation during neuro-interventional imaging and procedures: dose estimation and analysis of γ-H2AX foci and gene expression in blood lymphocytes. Mutat Res Genet Toxicol Environ Mutagen 2020; 856-857: 503237
• 9 Abend M, Badie C, Quintens R. et al. Examining radiation-induced in vivo and in vitro gene expression changes of the peripheral blood in different laboratories for biodosimetry purposes: first RENEB gene expression study. Radiat Res 2016; 185 (02) 109-123
• 10 Lacombe J, Sima C, Amundson SA, Zenhausern F. Candidate gene biodosimetry markers of exposure to external ionizing radiation in human blood: a systematic review. PLoS One 2018; 13 (06) e0198851
• 11 Lucas J, Dressman HK, Suchindran S. et al. A translatable predictor of human radiation exposure. PLoS One 2014; 9 (09) e107897
• 12 Albanese J, Martens K, Karanitsa LV, Schreyer SK, Dainiak N, Dainiak N. Multivariate analysis of low-dose radiation-associated changes in cytokine gene expression profiles using microarray technology. Exp Hematol 2007; 35 (4, Suppl 1): 47-54
• 13 Manning G, Macaeva E, Majewski M. et al. Comparable dose estimates of blinded whole blood samples are obtained independently of culture conditions and analytical approaches. Second RENEB gene expression study. Int J Radiat Biol 2017; 93 (01) 87-98
• 14 Visweswaran S, Joseph S, S VH, O A, Jose MT, Perumal V. DNA damage and gene expression changes in patients exposed to low-dose X-radiation during neuro-interventional radiology procedures. Mutat Res Genet Toxicol Environ Mutagen 2019; 844: 54-61
• 15 Ainsbury E, Badie C, Barnard S. et al. Integration of new biological and physical retrospective dosimetry methods into EU emergency response plans - joint RENEB and EURADOS inter-laboratory comparisons. Int J Radiat Biol 2017; 93 (01) 99-109
• 16 Li S, Lu X, Feng JB. et al. Developing gender-specific gene expression biodosimetry using a panel of radiation-responsive genes for determining radiation dose in human peripheral blood. Radiat Res 2019; 192 (04) 399-409
• 17 Oh JH, Deasy JO. Inference of radio-responsive gene regulatory networks using the graphical lasso algorithm. BMC Bioinformatics 2014; 15 (Suppl. 07) S5
• 18 Oh JH, Wong HP, Wang X, Deasy JO. A bioinformatics filtering strategy for identifying radiation response biomarker candidates. PLoS ONE 2012; 7 (06) e38870
• 19 Sekaran TSG, Kedilaya VR, Kumari SN, Shetty P, Gollapalli P. Exploring the differentially expressed genes in human lymphocytes upon response to ionizing radiation: a network biology approach. Radiat Oncol J 2021; 39 (01) 48-60
• 20 Zhao Y, Wong L, Goh WWB. How to do quantile normalization correctly for gene expression data analyses. Sci Rep 2020; 10 (01) 15534
• 21 Ritchie ME, Phipson B, Wu D. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015; 43 (07) e47-e47
• 22 Ashburner M, Ball CA, Blake JA. et al; The Gene Ontology Consortium. Gene ontology: tool for the unification of biology. Nat Genet 2000; 25 (01) 25-29
• 23 Ge SX, Jung D, Yao R. ShinyGO: a graphical gene-set enrichment tool for animals and plants. Bioinformatics 2020; 36 (08) 2628-2629
• 24 Chen EY, Tan CM, Kou Y. et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 2013; 14 (01) 128
• 25 Lu H, Zhou Q, He J. et al. Recent advances in the development of protein-protein interactions modulators: mechanisms and clinical trials. Signal Transduct Target Ther 2020; 5 (01) 213
• 26 Petrakis S, Andrade-Navarro MA. Editorial: protein interaction networks in health and disease. Front Genet 2016; 7: 111
• 27 Chin CH, Chen SH, Wu HH, Ho CW, Ko MT, Lin CY. cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst Biol 2014; 8 (Suppl. 04) S11
• 28 Vidal M, Cusick ME, Barabási AL. Interactome networks and human disease. Cell 2011; 144 (06) 986-998
• 29 Lambert SA, Jolma A, Campitelli LF. et al. The human transcription factors. Cell 2018; 172 (04) 650-665
• 30 Clarke DJB, Kuleshov MV, Schilder BM. et al. eXpression2Kinases (X2K) Web: linking expression signatures to upstream cell signaling networks. Nucleic Acids Res 2018; 46 (W1): W171-W179
• 31 Berger SI, Posner JM, Ma'ayan A. Genes2Networks: connecting lists of gene symbols using mammalian protein interactions databases. BMC Bioinformatics 2007; 8 (01) 372
• 32 Lachmann A, Ma'ayan A. KEA: kinase enrichment analysis. Bioinformatics 2009; 25 (05) 684-686
• 33 Xu K, Bezakova I, Bunimovich L, Yi SV. Path lengths in protein-protein interaction networks and biological complexity. Proteomics 2011; 11 (10) 1857-1867
• 34 Delprato A. Topological and functional properties of the small GTPases protein interaction network. PLoS ONE 2012; 7 (09) e44882
• 35 Embar V, Handen A, Ganapathiraju MK. Is the average shortest path length of gene set a reflection of their biological relatedness?. J Bioinform Comput Biol 2016; 14 (06) 1660002
• 36 Gollapalli P, G TS, H M, Shetty P. N SK. Network topology analysis of essential genes interactome of Helicobacter pylori to explore novel therapeutic targets. Microb Pathog 2021; 158: 105059
• 37 Bononi A, Agnoletto C, De Marchi E. et al. Protein kinases and phosphatases in the control of cell fate. Enzyme Res 2011; 2011: 329098
• 38 Bhullar KS, Lagarón NO, McGowan EM. et al. Kinase-targeted cancer therapies: progress, challenges and future directions. Mol Cancer 2018; 17 (01) 48
• 39 Fält S, Holmberg K, Lambert B, Wennborg A. Long-term global gene expression patterns in irradiated human lymphocytes. Carcinogenesis 2003; 24 (11) 1837-1845
• 40 Turtoi A, Sharan RN, Srivastava A, Schneeweiss FHA. Proteomic and genomic modulations induced by γ-irradiation of human blood lymphocytes. Int J Radiat Biol 2010; 86 (10) 888-904
• 41 Marchetti F, Coleman MA, Jones IM, Wyrobek AJ. Candidate protein biodosimeters of human exposure to ionizing radiation. Int J Radiat Biol 2006; 82 (09) 605-639
• 42 Liao H, Ji F, Ying S. CDK1: beyond cell cycle regulation. Aging (Albany NY) 2017; 9 (12) 2465-2466
• 43 Diril MK, Ratnacaram CK, Padmakumar VC. et al. Cyclin-dependent kinase 1 (Cdk1) is essential for cell division and suppression of DNA re-replication but not for liver regeneration. Proc Natl Acad Sci U S A 2012; 109 (10) 3826-3831
• 44 Satyanarayana A, Kaldis P. Mammalian cell-cycle regulation: several Cdks, numerous cyclins and diverse compensatory mechanisms. Oncogene 2009; 28 (33) 2925-2939
• 45 Howe O, White L, Cullen D. et al. A 4-gene signature of CDKN1, FDXR, SESN1 and PCNA radiation biomarkers for prediction of patient radiosensitivity. Int J Mol Sci 2021; 22 (19) 10607
• 46 El-Deiry WS. p21(WAF1) mediates cell-cycle inhibition, relevant to cancer suppression and therapy. Cancer Res 2016; 76 (18) 5189-5191
• 47 Qin L, Fan M, Candas D. et al. CDK1 enhances mitochondrial bioenergetics for radiation-induced DNA repair. Cell Rep 2015; 13 (10) 2056-2063
• 48 Candas D, Fan M, Nantajit D. et al. CyclinB1/Cdk1 phosphorylates mitochondrial antioxidant MnSOD in cell adaptive response to radiation stress. J Mol Cell Biol 2013; 5 (03) 166-175
• 49 Orlando DA, Lin CY, Bernard A. et al. Global control of cell-cycle transcription by coupled CDK and network oscillators. Nature 2008; 453 (7197) 944-947

Publication History

Article published online:
04 June 2024