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

Atopic dermatitis (AD) is the most common inflammatory skin disease with an estimated prevalence of around 20% in children and 7–10% in adults [1–4]. 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 [6–8]. However, comparative analysis of the DEGs between pediatric AD and adult AD remains to be elucidated.

The aim of the present study was to explore gene signature of pediatric AD and identify differentially expressed genes involved in pediatric AD comapred with 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 [9–12]. 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.

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.

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 cutoff ≥10, node score cutoff ≥0.2, K-core ≥2, and max depth = 100.

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 (Supplementary Tables S1 and S2). Red or green dots in the volcano plots represented significantly up- or down-regulated genes, respectively (Figure 1). The top 50 up- and down-regulated genes were shown in the heat map (Figure 2).

We uploaded all DEGs to the online software DAVID to identify over-represented GO categories and KEGG pathways. GO term enrichment analysis showed that up-regulated DEGs were significantly enriched in the migration and chemotaxis of granulocyte and Neutrophil, while down-regulated DEGs were mainly involved in a multi-organism process, In addition, molecular function analysis showed that up-regulated DEGs were mainly associated with chemokine activity, while down-regulated DEGs were involved in protein binding (Table 1). Furthermore, KEGG pathway analysis showed that up-regulated DEGs participated in the chemokine signaling pathway while down-regulated 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 1651 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).

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 [9–12]. Therefore, in the present study we retrieved gene expression data of 19 pediatric AD samples and 49 adult AD samples from previous studies and identified 654 DEGs in pediatric AD samples, among which 394 were up-regulated and 260 were down-regulated. Cumulative evidence has demonstrated that the co-expressed genes normally consist of a group of genes with similar expression profiles and participate in parallel biological process. To better understand the interactions of DEGs, we further performed GO, KEGG pathway, and PPI network analysis.

GO analysis showed that DEGs mainly participated in extracellular space, anchoring junction and adherens junction, involved in granulocyte and neutrophil migration, performed functions of cytokine activity, chemokine receptor binding, chemokine activity, and cytoskeletal protein binding. Furthermore, enriched KEGG pathways of up-regulated DEGs included chemokine signaling pathway and cytokine–cytokine receptor interaction, and those of down-regulated DEGs included adherens junction, focal adhesion, and regulation of actin cytoskeleton. Therefore, all these pathways could contribute to the pathogenesis of pediatric AD.

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 [17–21]. Moreover, their transcription is up-regulated rapidly in response to mitogenic stimuli including epidermal growth factor, transforming growth factor-β, and platelet-derived growth factor [22–24]. 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 [27–30]. 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 with 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, the present 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.

All data are available upon request.

The authors declare that there are no competing interests associated with the manuscript.

The authors declare that there are no sources of funding to be acknowledged.

C.Z. designed the study. T.W., B.Z., D.L., and X.Q. collected and analyzed the data. All authors read and approved the manuscript.

No ethics statement was require because this study involved no human or animals.

AD

atopic dermatitis

DEG

differentially expressed gene

EGFR

epidermal growth factor receptor

GO

Gene ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

PPI

protein–protein interaction

RMA

Robust multi-array average