Identification and Verification of Hub mRNA- miRNA-lncRNA Network on Psoriasis by Integrated Bioinformatics Analysis

Authors’ Contribution Conceptualization, SSL and QHW. Methodology, BL. Software, WYG. Validation, BL and WDW. Formal analysis, SSL. Investigation, WDW. Resources, WDW. Data curation, SSL and WDW. Writing original draft preparation, SSL. Writing review and editing, QHW. Visualization, WDW. Supervision, WYG. Project administration, WYG. All authors have read and agreed to the published version of the manuscript.


INTRODUCTION
P soriasis is a chronic autoimmune inflammatory skin disease with a relatively common etiology, with a prevalence of 2% to 3% globally (Boehncke, 2015;Parisi et al., 2013). The disease is characterized by abnormal proliferation and division of keratinocytes, lymphocyte and neutrophil infiltration, including T cells and dendritic cells, as well as the release of inflammatory cytokines (Rendon and Schäkel, 2019). Although there are several treatment options available for psoriasis, the curative effect is limited by severe side effects and high rates of recurrence, making it a worldwide medical issue. For instance, methotrexate can inhibit the proliferation of keratin cells and activate T lymphocytes to achieve certain therapeutic purposes, but it may cause adverse reactions of gastrointestinal reactions, hematopoietic system issues, and liver dysfunction (Flytström et al., 2008). In addition,

O n l i n e F i r s t A r t i c l e
psoriasis is also prone to relapse after withdrawal of steroid hormones, retinoic acid, etc. Thus, it is essential to identify novel biomarkers and pathways involved in psoriasis to discover new insights into disease treatment.
Although researchers have increasingly focused on biological intervention RNA in treating psoriasis, the exact mechanisms of action remain not fully elucidated. MicroRNAs (miRNAs) and long noncoding RNAs (lncRNAs) regulate gene expression in many biological processes, including psoriasis. Recent studies have indicated that lncRNAs play crucial roles in psoriasis development by regulating key protein expression involved in chronic inflammatory processes, immune infiltration, and hyperproliferation (Zhou et al., 2019). For example, lncRNA PRINS mediates abnormal psoriatic keratinocyte proliferation by regulating G1P3 expression (Szegedi et al., 2010). The microRNA (miRNA)-31 inhibits protein phosphatase 6 to promote keratinocyte proliferation (Yan et al., 2015), and lncRNA MSX2P1 inhibits miR-6731-5p and activates S100A7 to promote keratinocyte growth and proliferation (Qiao et al., 2018). Proliferation and inflammation of HaCaT can be inhibited by lowing miR-221-3p expression, providing potential therapeutic intervention against psoriasis (Meng et al., 2021). Furthermore, lncRNA NEAT1 (nuclear enriched abundant transcript 1) stimulates inflammasome activation in macrophage, thus the activated caspase-1 promoting IL-1β production and pyroptosis, which are elevated in psoriatic samples. Thus, it is great sense to elucidate the function of lncRNA in the treatment of psoriasis. Non-coding RNAs do not exist in isolation, but appear to form complex regulatory networks implicated in disease regulation, including psoriasis. They are important competing endogenous RNAs (ceRNAs) that inhibit miRNA-mediated target repression by competing for miRNA binding sites in RNAs (Zhou et al., 2019). Therefore, elucidating non-coding RNA mechanisms, particularly lncRNA-miRNA-mRNA networks, in psoriasis can improve understanding of the pathogenesis and facilitate drug development.
In this study, GSE142582 and GSE54456 datasets from GEO database were used to screen differentially expressed microRNAs (DEmiRNAs) and mRNAs (DEmRNAs) between psoriatic lesions and normal skin. Then bioinformatics analysis was employed to integrate and construct hub lncRNA-miRNA-mRNA networks in psoriasis. Furthermore, the psoriasis rat model was established, and RT-qPCR was employed to verify the hub networks in psoriasis ( Supplementary Fig. S1). The aim was to investigate novel and significant mechanisms of psoriasis associated dysregulation of lncRNA-miRNA-mRNA networks, which may contribute to the progression of psoriasis and provide potential therapeutic targets for psoriasis treatments.

Collection of RNA-seq datasets
The GEO database (https://www.ncbi.nlm.nih.gov/ geo) was applied to get datasets for evaluating mRNA and miRNA expression in skin tissue samples from psoriasis patients and healthy individuals. Two RNA sequencing datasets were obtained. The GSE54456 dataset contains 92 psoriasis and 82 health control samples based on the GPL9052 Illumina Genome Analyzer (H. sapiens). The GSE142582 dataset contains five psoriasis and five health control samples based on GPL20301 Illumina HiSeq 4000 (H. sapiens).

Differential expression analysis of mRNA and miRNA
The R package 3.6.3 (https://cran.r-project.org/) was used for RNA-seq expression profile analysis to identify significant DEmRNAs and DEmiRNAs. The screening criteria for significant differential expression between psoriasis and health control samples were |log2 (foldchange)| > 1 and adjusted P value < 0.05. Heat maps and volcano plots about DEmRNAs and DEmiRNAs were drawn using the Omic Studio tools (https://www. omicstudio.cn/index).

GO and KEGG analysis of DEmRNAs
The database for annotation, visualization and integrated discovery (DAVID) database (https:// david.ncifcrf.gov/) was utilized for gene ontology (GO) functional enrichment analyses. After uploading DEmRNAs, enriched biological process (BP), molecular function (MF), and cellular component (CC) pathways were identified. The Kyoto encyclopedia of genes and genomes (KEGG) enrichment analyses were performed by KOBAS database (http://kobas.cbi.pku.edu.cn/index.php). A P-value of <0.05 was considered statistically significant. Bubble charts, showing GO and KEGG analyses for significant DEmRNAs in the GSE54456 dataset, were generated using ChiPlot (https://www.chiplot.online/).

Construction of protein-protein interaction (PPI) network
The screened DEmRNAs were input into the STRING database (https://cn.string-db.org/) to construct a PPI network. The regulatory relationships between genes were visualized using Cytoscape v3.7.2 (https://cytoscape.org/). Hub genes in the regulatory network were analyzed using the CytoHubba plug-in with maximal clique centrality (MCC) algorithm.

O n l i n e F i r s t A r t i c l e
Construction of miRNA-mRNA network The miRWalk 2.0 database (http://www.umm.uni-heidelberg.de/apps/zmf/mirwalk/) was utilized to predict mRNA targets of the DEmiRNAs. The intersecting mR-NAs between the upregulated DEmiRNAs corresponding to mRNAs and down regulated DEmRNAs, and the intersecting mRNAs between the down regulated DEmiRNAs corresponding to mRNAs and up regulated DEmRNAs, were identified. These intersecting mRNAs and their corresponding miRNAs were arranged using Excel software, and then visualized using Cytoscape v3.7.2.

Construction of lncRNA-miRNA network
StarBase v2.0 (http://starbase.sysu.edu.cn) is a profundity online tool for system recognition of RNA-RNA and protein-RNA interaction networks. In this study, down-and up-regulated DEmiRNAs were assembled to predict the target lncRNAs (filter criteria, "clip ExpNum" ≥2), then the lncRNA-miRNA interactions were generated.

Construction of lncRNA-miRNA-mRNA network
To better comprehend the roles of lncRNAs, miRNAs, and mRNAs in psoriasis, a lncRNA-miRNA-mRNA regulatory network was constructed using the top ten lncRNAs, DEmiRNAs, and hub genes. The resulting network was visualized using Cytoscape 3.7.2 software.

Establishment of psoriasis rat model and hub lncRNA-miRNA-mRNA regulatory network verification
Male specific pathogen free SD rats were obtained from Shanghai Shrek experimental animal Co., Ltd. (Shanghai, China), and were treated in conform to the Guide for the Care and Use of Laboratory Animals. A total of 12 rats were housed at 25 ± 2 °C temperature with a 12-h light/dark cycle. The study was approved by the Bioethics Committee of Zhejiang Academy of Traditional Chinese Medicine (Approval No. KTSB2022018). After a seven-day adaptive feeding, the rats were shaved on their back (size 3 × 3 cm 2 ), and imiquimod (62.5 mg) was gently applied to the shaved skin areas for 7 days to establish psoriasis rats model (n = 6). Normal control rats were coated with an equal amount of vaseline on their back (n = 6). All rats were given standard diet and water.
At the end of experiment, skin samples from the psoriasis and control rats were collected and fixed in 10% neutral buffer formalin for histological studies. The skin samples were stained with hematoxylin-eosin (H&E) staining, and the images were captured using a digital camera attached to a light microscopy (Olympus, Tokyo, Japan) at 40× and 400× magnification. Epidermal thicknesses were measured using ImageJ software.
The remaining skins were promptly frozen in liquid nitrogen, and total RNA was extracted utilizing the SPARKeasy tissues/cell RNA Rapid Extraction Kit (#AC0202, Shandong Sikejie Biotechnology Co., Ltd., Shandong, China). The cDNA was synthesized using the SPARK script II RT Plus kit (#AG0304-B, Shandong Sikejie Biotechnology Co., Ltd., Supplementary Table SI). All primers were synthesized by Sangon Biotech Co., Ltd. (Shanghai, China) and are presented in Table I. Real-Time PCR was performed using the 2× SYBR Green qPCR Mix (#AH0104-B, Shandong Sikejie Biotechnology Co., Ltd.), with the following conditions: 2 min at 94 °C, and then 40 cycles of 94 °C for 10 s, 60 °C for 35 s. β-actin was used as a normalization control, and the fold change for each target gene was counted by the 2 −ΔΔCt method.

Statistical analysis
Results were presented as the mean±standard deviation (SD). SPSS 15.0 software (SPSS Inc, Chicago, IL, USA) was used for statistical analysis. The differences between groups were compared by student's t-test. P < 0.05 was considered statistically significant.

Identification of DEmRNAs and DEmiRNAs
The microarray datasets of GSE54456 and GSE142582 were normalized and analyzed (Fig. 1), and 813 significant DEmRNAs were identified in the GSE54456 dataset between normal and lesion tissue, with 480 up-regulated and 333 down-regulated mRNAs (Fig.  1A, C, E). Additionally, 299 significant DEmiRNAs were identified in the GSE142582 dataset, with 140 up-regulated and 159 down-regulated miRNAs (Fig. 1B, D, F).

GO and KEGG pathway enrichment analyses for the DEmRNAs
To perform functional enrichment analyses, the identified 813 DEmRNAs were conducted GO enrichment analysis and KEGG pathway analysis using DAVID O n l i n e

F i r s t A r t i c l e
Hub mRNA-miRNA-lncRNA Network on Psoriasis 5 and KOBAS respectively. The GO analysis on BP of up-regulated DEmRNAs were most enriched in innate immune response, followed by inflammatory response, immune response, defense response to virus ( Fig. 2A). Down-regulated DEmRNAs were most enriched in lipid metabolic process, cell adhesion, and positive regulation of gene expression, muscle contraction, aging, intermediate filament organization, and fatty acid metabolic process (Fig. 2B). KEGG pathway enrichment analyses revealed that up-regulated DEmRNAs were most enriched in metabolic pathways, and followed by proteoglycans in cancer, pathways in cancer PPAR signaling pathway, cell adhesion molecules (CAMs), and tight junction (Fig.  2C). In addition, down-regulated DEmRNAs were most enriched in metabolic pathways, followed by cytokinecytokine receptor interaction, NOD-like receptor signaling pathway, influenza A, chemokine signaling pathway and pathways in cancer (Fig. 2D).
The integrated analysis of predicted 159 down-regulated DEmiRNAs target mRNA and the identified up-regulated DEmRNAs were performed, there were 281 overlapping mRNA and 16 related down-regulated DEmiRNAs ( Supplementary Fig. S2, Table II). According to the degree of miRNA, the top ten hub down-regulated DEmiRNAs were listed in Table II.

lncRNA-miRNA interaction network
The selected 24 up-regulated DEmiRNAs and 16 down-regulated DEmiRNAs were used to predict target lncRNAs via StarBase online. As filter criteria of the clip O n l i n e

lncRNA-miRNA-mRNA interaction network
To better understand the roles of lncRNAs, miRNAs, and mRNAs in psoriasis, the top ten regulated lncRNAs, corresponding DEmiRNAs, and top ten regulated hub genes were used to establish a lncRNA-miRNA-mRNA regulatory network. As depicted in Figure 4A, the network consisted of nine lncRNAs, five down-regulated miRNAs, and three up-regulated hub genes. There were multiple connections between lncRNAs and miRNAs, as well as between miRNAs and mRNAs, such as lncRNA (XIST/ NEAT1)-miRNA (hsa-miR-10a-5p, hsa-miR-125b-5p)-mRNA (KIF2C).

Validation of the hub lncRNA-miRNA-mRNA regulatory network
The histopathological study demonstrated that, psoriatic skin showed pathological features, such as red plaques and thickening of the epidermis, consistent with the clinical pathological change (Fig. 5A). In contrast, the skin structure of the normal control group was intact, and without hyperkeratosis or spinous layer hypertrophy. After imiquimod modeling, the number of keratinocytes increased significantly in the skin, including hyperkeratosis, hypokeratosis, spinous layer hypertrophy, superior parietal mastoid, and inflammatory cell infiltration in the dermis. In comparison to the normal group, the epidermal layer thickness of the skin remarkablely increased after imiquimod modeling (p <0.05) (Fig. 5B). The expression levels of the hub lncRNA-miRNA-mRNA regulatory network were investigated in the skin of psoriatic rats (Fig. 6). The results showed that the expression of lncRNA NEAT1 was significantly increased (p <0.05), while miR-125b-5p, miR-125b-2-3p, and KIF2C were significantly decreased (p <0.01), in comparison O n l i n e W-D. Wand et al. to the normal group (Fig. 6A-D). On the other hand, the expression levels of lncRNA MALAT1 and XIST were significantly decreased (p <0.01, Fig. 6E, F), as well as a decline on MYLK, ACTG2 and MYH11 mRNA expression (p <0.05, Figure 6I-K). Comfortably, the expression of miR-135b-5p was notablely up-regulated in the skin of psoriatic rats, compared to the normal group (p <0.01, Fig.  6H). These findings suggested that the lncRNA (NEAT1)-miRNA (miR-125b-5p, miR-125b-2-3p)-mRNA (KIF2C) axis and the lncRNA (MALAT1, XIST)-miRNA (miR-135b-5p)-mRNA (MYLK ACTG2 and MYH11) axis, were vital mechanisms involved in psoriasis.

DISCUSSION
Previous studies have demonstrated the crucial role of mRNA, miRNA, and lncRNA in the development and progression of psoriasis through various signaling pathways (Zhou et al., 2019). However, few studies have integrated mRNA, miRNA, and lncRNA regulatory networks from diseased and non-diseased skin samples. Therefore, a comprehensive bioinformatics analysis was conducted in this study to identify their regulatory networks related to psoriasis.
Our analysis revealed that the up-regulated DEmRNAs were notably enriched in immune and inflammatory responses, while the down-regulated DEmRNAs were high enrichment in metabolic processes and cell adhesion. KEGG pathway analysis revealed that the up-regulated DEmRNAs were involved in cytokine-cytokine receptor interactions and NOD-like receptor signaling, while the down-regulated DEmRNAs were relative to CAMs, tight junctions, and act in cytoskeleton regulation. Our analysis also demonstrated a close relationship between IL-17 and psoriasis, which aligns with previous report (Griffiths et al., 2021). Relationships between chemokines and inflammation have been previously highlighted and implicated in psoriasis patients (Zdanowska et al., 2021). Cyclin-dependent kinase 1 (CDK1), was the upregulated hub genes belonging to the cell cycle regulatory protein family, which is involved in cell cycle maintenance; it drives the cell cycle through chemical action on serine/ threonine protein, and functions cooperatively with cyclin (Liao et al., 2017). Cyclin A2 (CCNA2), a key cell cycle regulator, is implicated in both DNA replication and mitotic entry and is vital in controlling the cell cycle's G1/S (initiation) and G2/M (mitosis) transitions (Cascales et al., 2021). Cyclin B2 (CCNB2/KIF20A) mainly involved in the cell cycle, proliferation, protein transport, and other biological processes (Shubbar et al., 2013).
ceRNA network theory suggests that lncRNAs have important roles in psoriasis development and occurrence. In these networks, NEAT1, a vital par speckle component lncRNA, has an indispensable role in par speckle formation and integrity (Bu et al., 2020). NEAT1 promotes inflammasome activation in macrophages, while stabilizing mature caspase-1 to promote IL-1β production and pyroptosis (Zhang et al., 2019b). In another study,

O n l i n e F i r s t A r t i c l e
Hub mRNA-miRNA-lncRNA Network on Psoriasis 9 relative to IL-1β and caspase-1 expression in normal skin biopsy samples, their expression was approximately 2.2-4.6 folds higher in psoriatic samples (Su et al., 2018). Thus, lncRNA NEAT1 may impact psoriasis occurrence and development via pro-inflammatory mechanisms or innate immunity.
Psoriasis is manifested as abnormal epidermal proliferation (Jia et al., 2020) and is related to apoptosis suppression (El-Domyati et al., 2013). MALAT1, which promotes cell proliferation and migration in different cancers, has been reported there is an association between MALAT1 polymorphisms and psoriasis risk, but the exact effects were unclear Mungmunpuntipantip et al., 2022). Studies have shown that macrophage miR-106b-5p excretion from impaired vitamin D receptor signaling induce inflammation (Oh et al., 2020), while Vitamin D3 analogs impact psoriasis by binding nuclear vitamin D3 receptors on genes implicated in proliferation, differentiation, and inflammation (O'neill and Feldman, 2010). MALAT1 sponges miR-106b-5p to promote colorectal cancer invasion and metastasis (Zhuang et al., 2019). For miR-142-3p, it reportedly disrupts MALAT1/ miR-142-3p sponging to decrease cervical cancer cell invasion and migration . The mRNA myosin heavy chain 11 (MYH11), which is involved in cytoskeleton formation, cell movement, signal transduction, and muscle contraction, was found to be down-regulated in breast cancer tissue, but overexpression inhibited cancer cell proliferation and migration, and promoted the formation of intercellular substances (Xiao et al., 2020;Sun and Li, 2021). In our study, miR-106a-5p and miR-142-3p were significantly increased, and MYH11 was also significantly reduced in skin lesions. Therefore, the MALAT1-miR (miR-106a-5p, miR-142-3p)-mRNA (MYH11) axis may regulate cell migration and promote apoptosis and thus play a critical role in psoriasis development. Additionally, our psoriasis rats model showed significantly increased expression levels of lncRNA (MALAT1, XIST) and mRNA (MYKL, ACTG2 and MYH11), and decreased miRNA (miR-135a-5p), suggesting the lncRNA (MALAT1, XIST)-miRNA (miR-135a-5p)-mRNA (MYKL, ACTG2 and MYH11) axis may regulate abnormal proliferation in psoriasis.

CONCLUSION
In summary, in this work, we investigated a novel mRNA-miRNA-lncRNA regulatory network associated with psoriasis. Our findings suggest that this network is related to the proliferation and tight junction weakening between epidermal cells in psoriasis, which may be related to the unbalance of lncRNA (NEAT1)-miRNA (miR-125b-5p, miR-125b-2-3p)-mRNA (KIF2C) axis and lncRNA (XIST, MALAT1)-miRNA (miR-135b-5p)-mRNA (MYLK, ACTG2 and MYH11) axis. This study provides new insights into the molecular mechanisms underlying the pathogenesis of psoriasis, and may contribute to the development of lncRNA or miRNA drugs for its treatment.

Funding
The research work was partly supported by Traditional Chinese Medicine Science and technology Project of Zhejiang Province (2023ZL342).

IRB approval
Not applicable.

O n l i n e F i r s t A r t i c l e
W-D. Wand et al.

Ethical statement
This study was approved by the Bioethics Committee of Zhejiang Academy of Traditional Chinese Medicine (Approval No. KTSB2022018).

Supplementary material
There is supplementary material associated with this article. Access the material online at: https://dx.doi. org/10.17582/journal.pjz/20220310020325

Statement of conflict of interest
The authors have declared no conflict of interest.