Potential Novel Therapeutic Targets in Epilepsy from The View of Intestinal Comorbidities

Background: Epilepsy is a complicated neurological disorder with almost 30% refractory. Recent years, several studies showed that epilepsy is associated with its comorbidities by shared molecular mechanisms. However, the association of epilepsy and digestive comorbidities are still unclear. In this study, we aim to explore the association between inammatory bowel disease (IBD) and epilepsy, and to nd promising therapeutic targets for refractory epilepsy. Methods: Two gene expression proles (GSE134697 and GSE59071) were selected from Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) were identied by GEO2R and the DESeq2 package. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of intersection DEGs and Gene Set Enrichment Analysis (GSEA) were conducted by clusterProler package. The protein-protein interactions (PPI) network was established by using STRING and visualized by Cytoscape. Genes in the most signicant module identied by MCODE plug-in were considered as candidate hub genes. Validation of hub genes were performed by using the GSE143272 dataset. Results: CXCL8, CXCR4 and ITGAX were identied as the hub genes. Conclusions: CXCL8, CXCR4 and ITGAX were the shared molecular mechanism of epilepsy and IBD, which were potentially novel therapeutic targets in epilepsy. The three genes are signicantly upregulated in the case groups (untreated) of GSE134697 and GSE59071 datasets. no detectable expression of CD11c in control group [44]. The evidence above supported that the three hub genes (CXCL8, CXCR4 and ITGAX) found in this study were potentially the shared molecular mechanism of epilepsy and IBD. CXCL8, CXCR4 and ITGAX were discovered as potentially novel therapeutic targets in epilepsy from the view of intestinal comorbidities in this study. More details in mechanism warrant further experimental validation.


Introduction
Epilepsy is a brain disorder characterized by an enduring predisposition to generate epileptic seizures [1]. Over 70 million people worldwide are affected by epilepsy and suffering various negative consequences from epileptic seizures, including disability and mortality [2,3]. Although 70% of seizures could be controlled by using antiseizure medications (ASMs), 30% of were refractory that make epilepsy a burdensome disease [2][3][4]. Therefore, it is essential to explore novel therapy targets for epilepsy, especially for drug-resistant epilepsy [3,5].
Recent studies showed that several diseases of different systems were found more common in people with epilepsy than in the general population, including neuropsychiatric, cardiovascular, digestive, and immune system diseases [6]. This is owing to a bidirectional factor between epilepsy and its comorbidities in mechanism, including genetic, environmental, structural, and physiological factors [6,7]. It was found that epileptic seizures could be controlled by treating comorbidities [6,8,9], which potentially provided a novel treating idea for refractory epilepsy. Therefore, to explore shared molecular mechanisms of epilepsy and comorbid conditions is undoubtfully a fast and effective method to develop novel treatments for epilepsy [6]. Until now, some shared molecular mechanisms of epilepsy and neuropsychiatric comorbidities have been discovered [10][11][12], but the association of epilepsy and intestinal comorbidities are still unclear.
In ammatory bowel disease (IBD) is a common intestinal comorbid condition of epilepsy, which are characterized by a chronic and recurrent gastrointestinal condition including ulcerative colitis (UC) and Crohn's disease (CD) [6,13,14]. IBD and epilepsy were found to be involved in the shared mechanisms in in ammatory, hypercoagulable, and genetic factors [13,15]. However, the certain molecules acting in the shared mechanisms are unknown. In this study, we conducted a combined transcriptome analysis of epilepsy and IBD patients from Gene Expression Omnibus (GEO) database to identify the shared molecules mechanisms and determine the potential pathway by which develop novel therapeutic targets for epilepsy.

Data Collection
All microarray datasets were downloaded from the GEO database. We searched GEO database for microarray datasets using the keywords "epilepsy" and "IBD" {including "Crohn's disease (CD)" and "ulcerative colitis (UC)"}. Datasets were recruited by the following criteria: (1) transcriptome data; (2) samples were derived from homo sapiens; (3) the count of samples was > 15; (4) including control groups. Finally, gene expression microarray datasets GSE134697 and GSE59071 were quali ed to be selected. According to the IBD classi cation, GSE59071 was separated into two parts: GSE59071-UC and GSE59071-CD. DEGs screenings of the two datasets GSE134697 and GSE59071 were conducted by the GEO2R online analysis tool (https://www.ncbi.nlm.nih.gov/geo/geo2r/) and the R package "DESeq2" [16], respectively. P-values were adjusted by using the false discovery rate (FDR) method. Genes with adjusted P < 0.05 and |log2 fold change (FC)| >1 were considered as DEGs. We used Venn diagram web tool (http://bioinformatics.psb.ugent.be/webtools/Venn/) to nd the DEGs shared by both datasets.

Function Enrichment Analysis of DEGs
To explore the enriched biological pathways and annotations, we utilized the RStudio v4.0.5 to conduct the Gene Ontology (GO) function and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis by clusterPro ler package [17]. Adjust P-values < 0 .05 were regarded as the cut-off criteria.

Gene Set Enrichment Analysis (GSEA) of Two Expression Pro les
To further reveal the common pathway between epilepsy and IBD, GSEA was performed based on each single expression matrix. R package "clusterPro ler" [17] was utilized to perform GSEA [18]. Enrichment analyses were conducted to determine whether a series of a priori-de ned biological processes was enriched. The genes of two datasets were ranked according to the false discovery rate (FDR). P < 0.05 were considered statistically signi cant.

Protein-protein Interaction Network Construction, Signi cant Module Analysis and Candidate Hub Gene Identi cation
The protein-protein interactions (PPI) network of DEGs was established by using STRING v11.0 (https://string-db.org). A PPI network of the DEGs with combined score > 0.4 in STRING was considered as a functional link and was visualized by Cytoscape software v3.7.2. The Cytoscape's plug-in molecular complex detection technology (MCODE) was used to identify closely connected modules from the PPI network with MCODE score > 4.5. The GO enrichment analysis of the module genes was conducted by using the Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.8 (https://david.abcc.ncifcrf.gov). Genes in the most signi cant module identi ed by MCODE plug-in were selected as candidate hub genes.

Validation of Candidate Hub Genes
We used GEO2R to identify the DEGs in the GSE143272 dataset [epilepsy response to valproate (VPA) treatment vs. healthy control]. The candidate hub genes that downregulated in this dataset were identi ed as hub genes.

Identi cation of DEGs in Epilepsy and IBD
In the GSE59071 dataset, 1243 DEGs (UC tissues vs. control) and 697 DEGs (CD tissues vs. control) were identi ed. In the GSE134697 dataset, 1846 DEGs (hippocampus vs. control) were identi ed. The shared DEGs of datasets GSE134697 and GSE59071 were 78 genes, including 56 upregulated and 22 downregulated genes ( Figure 1, Table 1).

GO Enrichment Analysis of DEGs
GO enrichment analysis showed that various DEGs were upregulated in the pathways of biological processes (BP), molecular functions (MF), and cellular components (CC). In BP, upregulated DEGs were signi cantly enriched in the "regulation of vasculature development", "response to molecule of bacterial origin", "response to lipopolysaccharide", "positive regulation of cytokine production", and "positive regulation of cell-cell adhesion" (Figure 2a, Table 2). In MF, upregulated DEGs were enriched in the "glycosaminoglycan binding", "cytokine activity", "receptor ligand activity", and "signaling receptor activator activity" (Figure 2b, Table 2). In CC, upregulated DEGs were signi cantly enriched in the "secretory granule membrane" and "collagen-containing extracellular" ( Figure   2c, Table 2). There also various downregulated DEGs enriched in MF ( Figure 2d, Table 2). More detailed GO enrichment results are shown in Table 2. The number of upregulated genes involved in the different processes of BP, CC, and MF was much more than that of downregulated genes. These results suggest that epilepsy and IBD may undergo common metabolic activities involving in cytokine associated in ammatory response and other complex biological process.

KEGG Enrichment Analysis of DEGs
KEGG enrichment analysis showed that upregulated genes were enriched in 26 pathways, such as rheumatoid arthritis, graft-versus-host disease, type I diabetes mellitus, hematopoietic cell lineage, leishmaniasis, intestinal immune network for IgA production, toll-like receptor signaling pathway, in ammatory bowel disease, and cytokine-cytokine receptor interaction pathway and so on (Table 3, Figure 3). These pathways are mainly involved in in ammatory or proin ammatory response. The downregulated genes were not enriched in any pathway (Table 3).

GSEA
interaction, graft-versus-host disease, hematopoietic cell lineage, and leishmaniasis infection. As for dataset GSE59071, the expression pro les of IBD samples were enriched in cell adhesion molecules cams, chemokine signaling pathway, complement and coagulation cascades, cytokine-cytokine receptor interaction, and ECM receptor interaction. The results indicated that the cytokine-cytokine receptor interaction pathway is the communal pathway between epilepsy and IBD (Figure 4a, 4b).

PPI Network Construction, Signi cant Module Analysis and Candidate Hub Gene Identi cation
We developed a PPI network containing 51 nodes and 162 edges and visualized by using Cytoscape software (Figure 5a). Nodes were drawn in different colors, representing the node degree (up or down). There were 56 upregulated genes and 22 downregulated genes. The most signi cant module with the highest score (module score: 9.111) was shown in Figure 5b, containing 10 genes in a network with 41 edges. GO analysis revealed that these genes were involved in chemokine-mediated signaling pathway, in ammatory response, and the other biological process were shown in Table 4. Genes in the most signi cant module were selected as candidate hub genes. These genes were all upregulated genes, as follows: CXCL8, CXCR4, ITGAX, CD80, SELL, CD44, CD69, IL1B, CD38, and CXCL2.

Validation of Candidate Hub genes
We nally identi ed three hub genes CXCL8, CXCR4 and ITGAX, which validated by using GSE143272 dataset [epilepsy response to VPA treatment vs. healthy control]. In the group of epilepsy to VPA treatment, CXCL8 was signi cantly downregulated (adjusted P <0.05 and log2FC =-1.2911696), CXCR4 and ITGAX were slightly downregulated. The three genes are signi cantly upregulated in the case groups (untreated) of GSE134697 and GSE59071 datasets.

Discussion
Although previous studies pointed out that epilepsy and IBD are related, the mechanism of their association is not addressed [19][20][21]. In this study, we conducted a transcriptome analysis to explore the association of epilepsy and IBD, aiming to discover promising therapeutic targets for refractory epilepsy.
A total of 78 overlapping DEGs were identi ed based on GSE134697 and GSE59071, of which 56 genes were upregulated and 22 genes were downregulated. GO analysis revealed that DEGs were enriched in the pathways of regulation of vasculature development, response to molecule of bacterial origin, response to lipopolysaccharide, and positive regulation of cytokine production. KEGG analysis showed that the DEGs enriched in the pathways of rheumatoid arthritis, intestinal immune network for IgA production, and toll-like receptor signaling. GSEA showed that the gene expression pro le of epilepsy was enriched in cytokine-cytokine receptor interaction pathway that was same as IBD. Previous study showed that central/peripheral in ammation played an important role in triggering or sustaining epileptic activity [8,[22][23][24][25]. Therefore, IBD may induce or aggravate seizures through the communal pathway and other in ammation/proin ammatory process. We identi ed 3 hub genes of CXCL8, CXCR4, and ITGAX, which were all upregulated in both epilepsy and IBD datasets. Additionally, CXCL8, CXCR4, and ITGAX were all downregulated in epilepsy under VPA treatment (dataset GSE143272), in which CXCL8 was downregulated most obviously. These ndings suggested that the three hub genes, especially CXCL8, may be promising therapeutic targets for epilepsy.
Interleukin-8 (IL-8: CXCL8) is a chemotactic and in ammatory cytokine that involves in central nervous system (CNS) in ammation [26,27]. Chemokines are small (8-12 kDa) chemotaxis proteins which play an important physiological role in immune system, mainly being responsible for guiding leukocyte to move to in ammatory sites [28]. Increased CXCL8 expression has been widely demonstrated in colonic mucosa of UC compared with normal control subjects, and upregulated CXCL8 concentration is correlated with the degree of in ammation [28]. There also some studies showed that the levels of CXCL8 transcripts were signi cantly higher in total epileptic patients compared with healthy subjects [29]. Elevated levels for CXCL8/IL-8 were reported in several different epilepsy etiologies and media [30], such as febrile infection-related epilepsy syndrome, and epileptic temporal lobe [31]. Thus, we supposed that CXCL8 may cause epilepsy by ways of immunoregulatory.
The chemokine CXC motif receptor 4 (CXCR4) is cognate with chemokine CXCL12. It is known as stromal cell derivedfactor-1 (SDF-1), which is highly expressed in the CNS and plays a crucial role in brain architecture [32][33][34]. Several studies indicated that the chemokine receptors were upregulated expressed in epilepsy and the active stage of UC [35-37], which was consistent with this study. Additionally, AMD3100, the antagonist of CXCR4, could reverse the neurogenesis promoted by enriched environment and suppresses long-term seizure activity [38,39]. Animal experiments also showed that AMD-3100 can reduce colon injury in mice and CXCL12/CXCR4 blocking was bene cial in improving experimental colitis in rodent models [40][41][42][43].
ITGAX encodes the integrin alpha X chain protein (CD11c), which is a pan dendritic cells (DCs) marker. While normally residing in the meninges and choroid plexus, perivascular and intraparenchymal DCs can be detected in brain under several in ammatory conditions. The level of CD11c protein was higher after status epilepticus (SE), while there was no detectable expression of CD11c in control group [44].
The evidence above supported that the three hub genes (CXCL8, CXCR4 and ITGAX) found in this study were potentially the shared molecular mechanism of epilepsy and IBD. CXCL8, CXCR4 and ITGAX were discovered as potentially novel therapeutic targets in epilepsy from the view of intestinal comorbidities in this study. More details in mechanism warrant further experimental validation.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests for this research.    Figure 1 Overlapping DEGs between epilepsy and in ammatory bowel disease (IBD).
A total of 78 overlapping DEGs of the epilepsy dataset (GSE134697) and IBD (UC and CD) dataset (GSE59071) were identi ed by the Venn web tool. DEG, differentially expression genes; UC, ulcerative colitis; CD, Crohn's disease.