Bioinformatic analysis of key pathways and genes involved in pediatric atopic dermatitis

Abstract The initiation of atopic dermatitis (AD) typically happens very early in life, but most of our understanding of AD is derived from studies on AD patients in adult. The aim of the present study was to identify gene signature speficic to pediatric AD comapred with adult AD. The gene expression profiles of four datasets (GSE32924, GSE36842, GSE58558, and GSE107361) were downloaded from the GEO database. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes pathway (KEGG) enrichment analyses were performed, and protein–protein interaction (PPI) network was constructed by Cytoscape software. Total 654 differentially expressed genes (DEGs) (394 up-regulated and 260 down-regulated) were identified in pediatric AD samples with adult AD samples as control. The up-regulated DEGs were significantly enriched in the migration and chemotaxis of granulocyte and neutrophil, while down-regulated DEGs were significantly enriched in biological adhesion. KEGG pathway analysis showed that up-regulated DEGs participated in chemokine signaling pathway while down-regulated DEGs participated in adherens junction, focal adhesion, and regulation of actin cytoskeleton. The top 10 hub genes GAPDH, EGFR, ACTB, ESR1, CDK1, CXCL8, CD44, KRAS, PTGS2, and SMC3 were involved in chemokine signaling pathway, cytokine–cytokine receptor interaction, interleukin-17 signaling pathway, and regulation of actin cytoskeleton. In conclusion, we identified DEGs and hub genes involved in pediatric AD, which might be used as therapeutic targets and diagnostic biomarkers for pediatric AD.


Introduction
Atopic dermatitis (AD) is the most common inflammatory skin disease with an estimated prevalence of around 20% in children and 7%-10% in adults [1][2][3][4]. AD is predominantly a Th2/Th22 polarized disease with Th1 polarization in the chronic phase and the impairment of Th17 pathway [5]. The initiation of AD typically happens very early in life, but most of our understanding of AD is derived from studies on AD patients in adult. Therefore, the molecular mechanism underlying pediatric AD initiation and progression is elusive, resulting in a lack of specific treatment for this disease.
Bioinformatics analysis of microarray data is increasingly valued as a promising tool in gene expression profiling in inflammatory diseases to identify differentially expressed genes (DEGs) that play important role in the diseases [6][7][8]. However, comparative analysis of the DEGs between pediatric AD and adult AD remains to be elucidated.
The aim of this study was to explore gene signature of pediatric AD and identify differentially expressed genes involved in pediatric AD comapred to adult AD. In present study, we download the original data (GSE32924, GSE36842, GSE58558, and GSE107361) from Gene Expression Omnibus and compared gene expression profiles of pediatric AD with those in adult AD. The DEGs were identified and analyzed by gene ontology (GO) and pathway enrichment analysis. From the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/), four gene expression profiles (GSE32924, GSE36842, GSE58558, and GSE107361) were selected because they were on gene expression profiling of AD samples (total 49 adult AD samples versus 19 pediatric AD samples) based on Affymetrix GPL570 platform [9][10][11][12]. The original probe-level data were converted into gene-level data using Robust multi-array average (RMA) approach for background correction and normalization. Next, limma packagein R language was used to identify the DEGs between pediatric and adult samples. Subsequently, a between-subjects t-test was performed to identify DEGs of each AD group with the cutoff criteria of log2 fold change (FC) >2 and FDR <0.01. Volcano plots were generated to visualize the distribution of DEGs between pediatric and adult samples of AD patients.

Gene Ontology and pathway enrichment analysis of DEGs
Bioinformatics analysis of the DEGs was performed as described previously [13].
Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were performed by employing an online software DAVID Database (https://david.ncifcrf.gov/). P ＜0.05 was considered statistically significant.

Integration of protein-protein interaction (PPI) network
STRING online database (http://string-db.org) was used for analyzing the protein-protein interaction (PPI) information. The cut-off criteria were a combined score of > 0.4 for a PPI network and a node degree of > 10 for screening hub genes. Cytoscape MCODE plug-in was used for searching clustered sub-networks. The default parameters were as follows: degree Downloaded from http://portlandpress.com/bioscirep/article-pdf/doi/10.1042/BSR20193517/899093/bsr-2019-3517.pdf by guest on 10 December 2020 cutoff ≥10, node score cutoff ≥0.2, K-core ≥2, and max depth = 100.

Identification of DEGs
A total of 654 genes (394 were up-regulated and 260 were down-regulated) special to pediatric AD samples were identified after the analyses in all four independent cohorts with adult AD sampels as control (Supplemental Table 1

Functional and pathway enrichment analyses
We uploaded all DEGs to the online software DAVID to identify overrepresented GO categories and KEGG pathways. GO term enrichment analysis showed that upregulated DEGs were significantly enriched in the migration and chemotaxis of granulocyte and Neutrophil, while downregulated DEGs were mainly involved in a multi-organism process, In addition, molecular function analysis showed that upregulated DEGs were mainly associated with chemokine activity, while downregulated DEGs were involved in protein binding (Table   1). Furthermore, KEGG pathway analysis showed that upregulated DEGs participated in the chemokine signaling pathway while downregulated DEGs participated in adherens junction, focal adhesion, and regulation of actin cytoskeleton ( Table 2). Based on the information in the STRING database, the top 10 hug nodes with higher degrees were screened (Table 3). Among these nodes, GAPDH showed the highest degree. A total of 594 nodes and 1,651 edges were analyzed using plug-ins MCODE. The top 3 significant modules were selected, the functional annotation of the protein involved in the modules was summarized. Enrichment analysis showed that the proteins in modules 1-3 were mainly associated with the chemokine signaling pathway, Pathway in cancer, Oxytocin signaling pathway ( Figure 3).

Discussion
Understanding of the molecular mechanism of pediatric AD might help develop approaches that can prevent atopic diathesis [14]. Previous studies have compared gene expression profiling of pediatric AD samples with adult AD sampels or normal healthy samples, respectively, but the sample size of the individual study was limited and the conclusion was controversial [9][10][11][12]. Therefore, in this study we retrieved gene expression data of 19 pediatric The analysis based on PPI networks indicated that GAPDH, EGFR, ACTB showed the highest betweenness and belonged to crucial modules of the PPI network. GAPDH is a classic glycolytic enzyme involved in membrane transport and membrane-fusion, microtubule assembly, nuclear RNA export, protein phosphotransferase/kinase reactions, and translational control of gene expression [15]. The β-actin cytoskeleton functions in cellular shape and anchorage where transmembrane glycoproteins link fibronectin in the extracellular matrix with actin microfilaments on the cytoplasmic side of the membrane [16]. While GAPDH and β-actin are regarded as housekeeping genes, accumulating evidence has suggested their mRNA levels vary with cellular proliferation [17][18][19][20][21]. Moreover, their transcription is upregulated rapidly in response to mitogenic stimuli including epidermal growth factor, transforming growth factor-β and platelet-derived growth factor [22][23][24]. We hypothesized that β-Actin and GAPDH expression levels in AD were variable and not suitable for normalizing mRNA levels. Our results were similar to some studies in asthma, which was part of the atopic march [25].
Epidermal growth factor receptor (EGFR) is a large transmembrane glycoprotein with ligand-induced tyrosine kinase activity [26]. Inhibition of EGFR signaling leads to decreased expression of cytoskeleton proteins such as actin-binding protein ACTN1 (actinin-1), increased keratinocyte adhesion, resulting in the inhibition of the migration of keratinocytes from the basal layer to the stratum corneum [27][28][29][30]. Blockade of EGFR signaling can regulate the expression of CCL26/eotaxin-3 in primary keratinocytes in AD [31,32].
In summary, we identified genes differentially expressed in pediatric AD compared to adult AD and explored their potential function and relevant pathways in the pathogenesis of pediatric AD. Moreover, our study suggested that chemokine pathway and cytoskeletal protein binding play a vital role in the molecular mechanism of pediatric AD. However, this study has limitation because it is based on bioinformatic analysis of online datasets and the differentially expressed genes in pediatric AD should be validated by real-time PCR analysis and function assay. In particular, further studies are needed to validate GAPDH, EGFR and ACTB, which can be considered as crucial genes involved in pediatric AD, with the potential to be used in the diagnosis and therapy.

Competing interests
The authors declare no conflict of interest.

Ethics statement
No ethics statement was require because this study involved no human or animals. CZ designed the study. TW, BZ, DL and XQ collected and analyzed the data. All authors read and approved the manuscript.

Data availability
All data are available upon request.

Funding
No funding was received.