Abstract

Purpose

   To undercover the underlying mechanisms of luteolin against atopic
   dermatitis (AD), clinically characterized by recurrent eczematous
   lesions and intense itching, based on network pharmacology, molecular
   docking and in vivo experimental validation.

Methods

   TCMSP, STITCH and SwissTargetPrediction databases were utilized to
   screen the corresponding targets of luteolin. Targets related to AD
   were collected from DisGeNET, GeneCards and TTD databases. PPI network
   of intersection targets was constructed through STRING 11.0 database
   and Cytoscape 3.9.0 software. GO and KEGG enrichment analysis were
   performed to investigate the critical pathways of luteolin against AD.
   Further, the therapeutic effects and candidate targets/signaling
   pathways predicted from network pharmacology analysis were
   experimentally validated in a mouse model of AD induced by 2,
   4-dinitrofluorobenzene (DNFB).

Results

   A total of 31 intersection targets were obtained by matching 151
   targets of luteolin with 553 targets of AD. Among all, 20 core targets
   were identified by PPI network topology analysis, including IL-6, TNF,
   IL-10, VEGFA, IL-4, etc., and molecular docking indicated that luteolin
   binds strongly to these core targets. KEGG pathway
   enrichment analysis suggested that the intersected targets were
   significantly enriched in IL-17 signaling pathway, Th17 cell
   differentiation, Th1 and Th2 cell differentiation, JAK/STAT signaling
   pathway, etc. The in vivo experiment validated that luteolin could
   alleviate AD-like skin symptoms, as evidenced by the lower
   SCORAD score, the reduced infiltration of mast cells and the recovery
   of skin barrier function. Furthermore, luteolin restored immune balance
   by regulating the production of Th1/Th2/Th17-mediated cytokines, which
   were both the predicted core targets. Moreover, luteolin inhibited the
   phosphorylation of JAK2 and STAT3 in the lesional skin.

Conclusion

   Together, the present study systematically clarifies the ameliorative
   effects and possible molecular mechanisms of luteolin against AD
   through the combination of network pharmacology and experimental
   validation, shedding light on the future development and clinical
   application of luteolin.

   Keywords: atopic dermatitis, luteolin, network pharmacology, target
   prediction, experimental validation

Introduction

   Atopic dermatitis (AD), a chronic inflammatory and pruritic skin
   disorder, is characterized by recurrent eczematous and inflamed skin
   lesions, compromised barrier function and intense itch.[34]1,[35]2 Itch
   is burdensome symptom and can negatively influence patients’ sleep
   quality, social life and emotional well-being. The pathogenesis of AD
   is multifactorial and complex, most researchers hold the view that
   immunologic aberrations and epidermal barrier dysfunction constitute
   critical factors in the development of this disease.[36]3,[37]4
   Findings proposed that the Th1/Th2 immune imbalance caused by a
   predominant Th2 over Th1 immune response has been recognized as a
   crucial pathogenesis of cutaneous inflammation in AD. Th2-mediated
   cytokines, such as interleukin (IL)-4, IL-5, IL-13, could activate IgE
   response to allergens and disrupt skin barrier function. Th17 cells and
   Th22 cells, another two novel CD4^+ T cell subsets, have also been
   found to be increased in AD lesions and associated with AD
   severity.[38]5–7

   Current treatment options for AD are limited and mainly include topical
   emollients, corticosteroids, and immunosuppressive agents. However,
   long-term epicutaneous application of corticosteroid increases the risk
   of recurrence and results in telangiectasia, petechiae, skin atrophy
   and thinning.[39]2 Thus, researchers have moved toward the discovery of
   alternative therapeutic agents for AD from traditional Chinese
   medicine or natural products with better long-term benefits and less
   side effects.[40]8

   Luteolin, a food‐derived flavonoid, which is present in medicinal
   plants worldwide, was reported to have many noteworthy pharmacological
   properties, including anti-inflammatory, anti-oxidant and anti-allergic
   effects.[41]9–11 One previous pre-clinical study showed that luteolin
   suppressed dry skin induced spontaneous scratching behavior in mice via
   inhibiting IgE-mediated mast cell activation.[42]12 However, the role
   of luteolin on AD remains unknown. With the rapid development of
   bioinformatics, systems biology and pharmacology, network pharmacology
   has been considered as a frontier in drug discovery and development.
   Network pharmacology highlights a paradigm shift from the current “one
   molecule, one target, one disease” mode a new “network multi-targets”
   mode.[43]13,[44]14 Therefore, in the present study, we firstly
   performed network pharmacology approach to predict potential molecular
   targets and molecular mechanisms of luteolin in treating AD. In silico
   molecular docking as well as an in vivo experiment was further
   conducted to confirm the findings. An overview of this study is shown
   in [45]Figure 1.

Figure 1.

   [46]Figure 1
   [47]Open in a new tab

   Flowchart of the current research study.

Materials and Methods

Related Databases and Platforms

   TCMSP database ([48]http://lsp.nwu.edu.cn/tcmsp.php);
   SwissTargetPrediction ([49]http://www.swisstargetprediction.ch/);
   STITCH 4.0 ([50]http://stitch.embl.de/); UniProt database
   ([51]http://www.uniprot.org); DisGeNET database
   ([52]http://www.disgenet.org/), GeneCards database
   ([53]https://www.genecards.org/), Therapeutic Target Database (TTD,
   [54]https://db.idrblab.org/ttd/); Venny 2.1.0
   ([55]http://bioinfogp.cnb.csic.es/tools/venny/index.html); STRING 11.0
   ([56]https://string-db.org/), Cytoscape 3.9.0
   ([57]http://www.cytoscape.org/); g:profiler database
   ([58]https://biit.cs.ut.ee/gprofiler/convert); OmicShare cloud platform
   ([59]https://www.omicshare.com/tools/).

Prediction of the Luteolin Pharmacological Targets

   To obtain the potential targets of luteolin, we searched TCMSP, STITCH
   4.0 and SwissTargetPrediction databases, followed by filtering with the
   term “Homo sapiens”. The operation is as follows: Firstly, we directly
   enter “luteolin” in the TCMSP and STITCH 4.0 databases and then
   download the corresponding target genes. Secondly, we got the
   luteolin’s Canonical SMILES from the PubChem database, imported it to
   the SwissTargetPrediction database and then obtained the predicted
   targets of luteolin. Through the UniProt database, we analyzed and
   annotated potential pharmacodynamic targets and the candidate targets
   were both modified to the official standardized abbreviation, that is,
   Gene symbol. Duplicated non-human and non-standard target genes were
   removed.

Prediction of Target Genes Associated with AD

   Using “Dermatitis, Atopic” as the key word, the potential targets for
   the treatment of AD in the DisGeNET, GeneCards and TTD databases were
   discovered. Among them, GeneCards database was used to screen target
   genes related to AD with a ‘Relevance score ≥5’ and the DisGeNET
   database was used to screen target genes related to AD with a
   ‘Gene-Disease Score ≥0.05’. Combine the search results of the three
   databases and then remove duplicate targets.

Intersection Targets Between Luteolin and AD

   The common potential targets of luteolin and AD, which represented the
   potential targets of luteolin in the treatment of AD, were further
   obtained by the online drawing software Venny 2.1.0.

Construction of Protein–Protein Interaction (PPI) Network and Topology
Analysis

   The intersection targets for luteolin against AD were input to the
   STRING 11.0 database to construct a PPI network. “Homo sapiens” was
   chosen and a scoring value >0.7 was selected as the minimum required
   interaction score protein interactions. Finally, we imported the PPI
   information into Cytoscape 3.9.0 software to visualize the intersection
   of PPI diagram. The core targets of luteolin treating AD were mined by
   Cytohubba and MCODE plugins in Cytoscape 3.9.0 software. Top 20 nodes
   with the largest “Degree” value generated from the CytoHubba plug-in
   were selected to establish a Hub gene network. Additionally, the
   analysis parameters of MCODE were set as follows: degree cutoff = 2,
   node score cutoff = 0.2, K-score = 2 and Max depth = 100.

GO Functional Enrichment and KEGG Pathway Enrichment Analysis

   The common target genes of luteolin and AD were imported into the DAVID
   database and OmicShare cloud platform for GO function and KEGG pathway
   analysis. The channels with Q value <0.01 are filtered out and sorted
   in descending order based on the Q value (adjust P). Top 20 items are
   selected, and the screening results were visualized by an advanced
   bubble chart.

Molecular Docking

   Firstly, the 3D structure of luteolin was downloaded from the PubChem
   database. Before docking simulation, the file was imported into PyMOL
   software to hydrogenate and add charge and then saved in .pdbqt format.
   Secondly, the X-ray crystal structures of the core targets were
   retrieved from the PDB database and modified using UCSF Chimera 1.14
   software including ligand and water removal and polar hydrogen atoms
   and Gasteiger charge addition. Finally, luteolin was used as the
   ligand, and the candidate protein targets were used as receptors for
   molecular docking using Chimera 1.14.

Experimental Verification

Chemicals

   DNFB (purity grade >99%) was obtained from Alfa Aesar, Heysham, UK.
   Luteolin (purity grade > 99%) was supplied by Aladdin Ltd., Shanghai,
   China. DEX (dexamethasone, purity grade > 98%) was purchased from Saen
   Chemical Technology Co. Ltd., Shanghai, China. All other general agents
   were commercially available.

Experimental Protocol and AD Model

   Experimental female KM mice aged 6 weeks were purchased from the Animal
   Center of Wuhan University (Wuhan, China). All mice were housed in the
   cages at a 12/12 h day/night cycle and provided a standard diet. The
   protocol of this experimental study was approved by the Institutional
   Animal Care and Use Committee (IACUC), Wuhan University Center for
   Animal Experiment, Wuhan, China), and all the procedures of this study
   were done in accordance with the Guide for the Care and Use of
   Laboratory Animals (Eighth Edition, 2011, published by the National
   Academies Press, Washington, USA). After 1 week of acclimation period,
   the mice were randomly divided into 4 groups: (i) Normal group,
   (ii) Model group (DNFB), (iii) 0.5% luteolin (g/mL) group (DNFB +0.5%
   luteolin), (iv) 0.05% DEX (g/mL) group (DNFB +0.05% DEX).

   To experimentally induce AD-like model, we employed cutaneous DNFB
   sensitization and challenging in mice according to the method of
   Mandlik et al[60]15 with some modifications. Briefly, 100 μL of 1.0%
   DNFB (w/v) in vehicle composed of acetone and olive oil (3:1) was
   applied onto the shaved dorsal skin (2 cm*3 cm) on day 1. Mice were
   then topically treated with 100 μL of 0.5% DNFB onto the dorsal skin on
   day 3, 7 and 11. Simultaneously, apart from Normal group and Model
   group (only treated with the same volume of vehicle), the 0.5% luteolin
   and 0.05% DEX (100 μL) were dissolved in ethanol in vehicle composed of
   ethanol and propylene glycol (3:1) and they were topically applied onto
   the shaved back of mice daily from day 1 to day 13.

Measurements of AD Severity Score, Scratching Behaviors and TEWL

   To evaluate the ameliorative effects of luteolin on DNFB-induced
   AD-like morphology, the dermatitis severity score and scratching
   behavior episodes within 10 min among different groups were examined in
   this study. The dermatitis severity score was measured macroscopically
   at day 13 based on four criteria previously described.[61]16 Briefly,
   symptom severity was scored as 0 (none), 1 (mild), 2 (moderate), and 3
   (severe) for each of the following four different manifestations:
   erythema/hemorrhage, scaling/dryness, edema/swelling and
   erosion/excoriation. Immediately after the application of DNFB,
   scratching behavior was counted for 10 min on day 13. Just scratching
   of the hind paws to the dorsal skin was recorded as the scratching
   behavior and the consecutive scratching before putting down the hind
   paws was counted as once.[62]17

   In addition, edema was further evaluated through skin swelling by
   measuring the difference in the thickness/weight among different
   groups. After sacrifice, the dorsal skin of each group was separated,
   respectively, and made into a circle slice using a puncher with the
   diameter of 8 mm. An analytical balance was used for measuring the skin
   weight and a vernier calliper was used for measuring the skin
   thickness.

   Transepidermal water loss (TEWL) of the shaved dorsal skin was measured
   by VAPO SCAN AS-VT100RS (Asch, Japan) and the values were determined
   from three independent measurements at day 12.

Histological Assay

   After sacrifice, the dorsal lesional skin of mice among different
   groups was harvested carefully. Skin specimens were immersed in 4%
   paraformaldehyde for 24 h and then embedded into wax blocks.
   Paraffin-embedded skins were then cut into 4.5-μm thin sections and
   were stained with hematoxylin-eosin (H&E) to detect epidermal thickness
   or toluidine blue (TB) to examine mast cells, including the granulated
   and degranulated forms. Quantitative analysis of epidermal thickness
   and mast cells was measured using ImageJ software taken at 3 randomly
   selected fields per section.

Enzyme-Linked Immunosorbent Assay (ELISA)

   Whole blood was obtained from the eyeball and then centrifuged at 3500
   rpm for 15 min at 4 ℃ to collect serum samples. Skin tissues (100
   mg) were homogenized in 9-fold 0.9% saline solution, and the
   homogenates were centrifuged for 10 min at 4 ℃ at 12,000 rpm to obtain
   the supernatant. The total protein concentration of supernatants was
   determined using the bicinchoninic acid (BCA) assay (Kerui
   Biotechnology Co., Ltd, Wuhan, China). Serum and skin levels of
   Th1/Th2/Th17-related cytokines were measured using ELISA kits.

Immunohistochemical (IHC) and Immunofluorescence (IF) Analysis

   The paraffin sections of mice in each group were subjected to
   immunohistochemical analysis according to the following steps:
   deparaffinization and hydration of sections➙3% H[2]O[2] blocking
   endogenous peroxidase➙EDTA antigen microwave retrieval➙5% BSA serum
   blocking ➙ Incubation with the primary antibodies: rabbit anti-FLG,
   anti-p-JAK2 and anti-p-STAT3 (Cell Signaling Technology, Boston, USA),
   4℃, overnight➙Incubation with goat anti-rabbit IgG secondary antibody,
   37℃, 1h➙DAB or DAPI staining.

Statistical Analysis

   One-way ANOVA followed by Dunnett’s post-hoc test using GraphPad Prism
   5.0 was used for statistical analysis of data. All the experimental
   data were presented as the mean ± standard deviation (S.D.) and p-value
   <0.05 or 0.01was considered significantly.

Results

Identification of the Targets Related to Luteolin Against AD

   This study got 56, 10 and 104 candidate targets of luteolin from TCMSP,
   STITCH and SwissTargetPrediction databases, respectively, and we
   finally received 151 targets after removing duplicates ([63]Figure 2A).
   By retrieving the DisGeNET, GeneCards and TTD databases and deleting
   repeated targets, a total of 553 target genes associated with AD were
   obtained ([64]Figure 2A). The Venn diagram produced by Venny 2.1.0
   showed that 31 intersection targets between luteolin and AD were
   obtained ([65]Figure 2B).

Figure 2.

   [66]Figure 2
   [67]Open in a new tab

   (A) Potential targets of luteolin and AD-related genes. (B)
   Intersecting targets between luteolin and AD-related genes determined
   by Venny 2.1.

Construction of PPI Network and Topology Analysis

   Thirty-one anti-AD targets of luteolin were imported into STRING 11.0
   databases to obtain the PPI network ([68]Figure 3A). Cytoscape 3.9.0
   was used for network visualization. PPI network stands for a global
   view of intersection target association, and the PPI network contained
   31 nodes and 259 edges. The circles represented the targets of luteolin
   against AD and the edges represented the interaction between targets.
   “Degree” value, which is calculated from CytoHubba plug-in, is also an
   important topological property that can explain the most important
   nodes in the network. The size and color trend of the nodes varies from
   the “Degree” value. The first 20 nodes with the highest ‘Degree’ value
   were extracted to establish a Hub gene network, including MMP9, IL6,
   TNF, IL10, VEGFA, IL4, ICAM1, MPO, MAPK1, IL2, MAPK8, JUN, EGFR, PPARG,
   IFNG, MMP1, MMP2, CD40LG, NFKBIA, AHR ([69]Figure 3B). The results of
   MCODE obtained the most significant cluster with scores of 18.5.
   Convincingly, cluster 1 comprised 21 nodes and 185 edges, which further
   confirmed the importance of 20 core genes obtained from CytoHubba
   analysis ([70]Figure 3C).

Figure 3.

   [71]Figure 3
   [72]Open in a new tab

   Protein–protein interaction (PPI) network of core targets against AD.
   (A) PPI network of 31 intersection targets constructed with STRING
   11.0. (B) Top 20 hub genes were screened by CytoHubba using Cytoscape
   3.9.0. (C) The most significant module was analyzed by MCODE.

Molecular Docking Verification of Luteolin with Core Target Genes

   Molecular docking, as one of bioinformatics tools, has been
   increasingly used in the early stages of drug discovery. The luteolin
   was molecularly docked with 20 core targets to further verify the
   interaction between the compound and the core targets. In the docking
   output results, the more negative docking score is predicted to have a
   higher binding affinity between the luteolin to protein. In general,
   the binding energy (affinity) below −5.0 kcal/mol means strong binding
   activity between the receptor and the ligand.[73]18–20 The bonding
   interaction data are shown in [74]Table 1, and the binding affinities
   were found to be ranged from-5.8 to −10.1 kcal/mol.

Table 1.

   Molecular Docking Results in Terms of the Binding Energy of Luteolin
   with 20 Anti-AD Core Targets
   Gene Symbol Description PDB ID Affinity (kcal/mol)
   IL4 Interleukin-4 2B8U −7.2
   IL10 Interleukin-10 2ILK −6.9
   IL6 Interleukin-6 1ALU −7.0
   IFNG Interferon gamma 6F1E −7.8
   MAPK8 Mitogen-activated protein kinase 8 3ELJ −8.9
   NFKBIA NF-kappa-B inhibitor alpha 5TQW −7.1
   AHR Aryl hydrocarbon receptor 4M4X −8.0
   TNF Tumor necrosis factor 5MU8 −7.8
   IL2 Interleukin-2 1M49 −6.9
   MAPK1 Mitogen-activated protein kinase 1 4O6E −8.1
   ICAM1 Intercellular adhesion molecule 1 1P53 −6.3
   CD40LG CD40 ligand 3LKJ −9.8
   MMP9 Matrix metalloproteinase-9 2OW1 −10.1
   MMP1 Interstitial collagenase 1HFC −9.3
   MPO Myeloperoxidase 1MHL −7.3
   EGFR Epidermal growth factor receptor 5Y25 −8.1
   VEGFA Vascular endothelial growth factor A 3QTK −7.8
   PPARG Peroxisome proliferator-activated receptor gamma 2ZNO −8.3
   JUN Transcription factor AP-1 1JNM −5.8
   MMP2 72 kDa type IV collagenase 1QIB −9.2
   [75]Open in a new tab

GO Function Analysis

   As illustrated in [76]Figure 4, the input of the 31 overlapping target
   genes derived the top 20 biological process (BP) enrichment items
   including response to cytokines, cellular response to cytokine
   stimulus, cellular response to chemical stimulus and cytokine-mediated
   signaling pathway. Cellular component (CC) enrichments involve
   extracellular region, membrane raft and cell surface. Molecular
   function (MF), enrichments involve cytokine receptor binding, signaling
   receptor binding, and MAP kinase activity.

Figure 4.

   [77]Figure 4
   [78]Open in a new tab

   Bubble diagram of GO enrichment analysis (Top 20). (A) Biological
   process (BP). (B) Cellular component (CC). (C) Molecular function (MF).

KEGG Pathway Analysis and Luteolin-Target-Pathway Network Construction

   The KEGG pathway enrichment analysis revealed that 31 intersection
   targets were mainly enriched in 71 signaling pathways (adjust P <
   0.01). To further uncover the possible mechanisms of luteolin in the
   prevention of AD, top 20 (based on adjust P value) AD-related KEGG
   analysis results were visualized by the OmicShare tool. As shown in
   [79]Figure 5A and [80]Table 2, key mechanisms of luteolin against AD
   might include IL-17 signaling pathway, T cell receptor signaling
   pathway, Th17 cell differentiation, TNF signaling pathway, Th1 and Th2
   cell differentiation, JAK/STAT signaling pathway, etc. The
   luteolin-target-pathway network visualized by Cytoscape 3.9.0 software
   comprised 20 signaling pathways and enriched gene targets is shown in
   [81]Figure 5B.

Figure 5.

   [82]Figure 5
   [83]Open in a new tab

   Bubble diagram of KEGG enrichment analysis (Top 20). (A) KEGG
   enrichment analysis. Size and color of the bubble represent the amount
   of enriched genes and enrichment significance, respectively. (B)
   “Component-target-pathway” network diagram of luteolin against AD. The
   outer green triangle is the relevant pathway and the inner purple
   rhombus is the relevant target in the pathway.

Table 2.

   KEGG Pathway Enrichment Analysis of Luteolin Against AD
   Pathway ID KEGG_B_Class Pathway Q (Adjust P) Gene
   ko04657 Immune system IL-17 signaling pathway 2.61E-10 MAPK1, NFKBIA,
   MMP9, MAPK8, IFNG, IL4, IL6, JUN, MMP1, TNF
   ko04660 Immune system T cell receptor signaling pathway 4.90E-09 MAPK1,
   NFKBIA, CD40LG, IL2, IFNG, IL4, IL10, JUN, TNF
   ko04625 Immune system C-type lectin receptor signaling pathway 4.90E-09
   MAPK1, NFKBIA, MAPK8, IL2, IL6, IL10, SYK, JUN, TNF
   ko04659 Immune system Th17 cell differentiation 1.66E-07 MAPK1, NFKBIA,
   AHR, MAPK8, IL2, IFNG, IL4, IL6, JUN
   ko04668 Signal transduction TNF signaling pathway 1.76E-07 ICAM1,
   MAPK1, NFKBIA, MMP9, MAPK8, IL6, JUN, TNF
   ko04664 Immune system Fc epsilon RI signaling pathway 3.42E-06 ALOX5,
   MAPK1, MAPK8, IL4, SYK, TNF
   ko05321 Immune diseases Inflammatory bowel disease (IBD) 5.39E-06 IL2,
   IFNG, IL4, IL6, IL10, JUN, TNF
   ko04658 Immune system Th1 and Th2 cell differentiation 1.14E-05 MAPK1,
   NFKBIA, MAPK8, IL2, IFNG, IL4, JUN
   ko05323 Immune diseases Rheumatoid arthritis 2.23E-05 ICAM1, IFNG,
   VEGFA, IL6, JUN, MMP1, TNF
   ko04620 Immune system Toll-like receptor signaling pathway 2.43E-05
   MAPK1, NFKBIA, MAPK8, IL6, JUN, TNF
   ko04060 Signaling molecules and interaction Cytokine-cytokine receptor
   interaction 8.36E-05 CD40LG, IL2, IFNG, IL4, IL6, IL10, CXCR1, TNF
   ko04630 Signal transduction Jak-STAT signaling pathway 1.20E-04 IL2,
   IFNG, IL4, IL6, IL10, EGFR
   ko04066 Signal transduction HIF-1 signaling pathway 1.32E-04 MAPK1,
   IFNG, VEGFA, IL6, EGFR
   ko04621 Immune system NOD-like receptor signaling pathway 2.35E-04
   MAPK1, NFKBIA, MAPK8, IL6, JUN, TNF
   ko04064 Signal transduction NF-kappa B signaling pathway 3.40E-04
   ICAM1, NFKBIA, CD40LG, SYK, TNF
   ko04662 Immune system B cell receptor signaling pathway 5.20E-04 MAPK1,
   NFKBIA, SYK, JUN
   ko04012 Signal transduction ErbB signaling pathway 8.55E-04 MAPK1,
   MAPK8, EGFR, JUN
   ko04151 Signal transduction PI3K-Akt signaling pathway 9.11E-04 MAPK1,
   IL2, VEGFA, IL4, IL6, EGFR, SYK
   ko05310 Immune diseases Asthma 1.87E-03 CD40LG, IL4, IL10, TNF
   ko04010 Signal transduction MAPK signaling pathway 2.77E-03 MAPK1,
   MAPK8, VEGFA, EGFR, JUN, TNF
   [84]Open in a new tab

Experimental Verification

Luteolin Alleviates Dorsal Skin Lesions in DNFB-Induced AD Mice

   At the end of the experiment, the AD model mice successfully developed
   severe by repeated application of DNFB ([85]Figure 6A). Compared to the
   Normal group, the lesional score in the AD model group was
   significantly increased about 6.75-fold ([86]Figure 6B). In line with
   the development of skin lesions, repeated treatment with DNFB triggered
   a significant increase in the scratching episodes within 10 min
   ([87]Figure 6C). Treatment with 0.5% luteolin markedly decreased
   DNFB-induced increases in dermatitis score and scratching bouts
   ([88]Figures 6B and [89]C).

Figure 6.

   [90]Figure 6
   [91]Open in a new tab

   Luteolin alleviates DNFB-induced AD-like lesions in AD mice. (A)
   Representative pictures of dorsal skin among different groups at day
   13. (B) The severity of AD was macroscopically evaluated by SCORAD at
   day 13. (C) The number of scratching bouts within 10 min was counted at
   day 12. (D) Skin swelling by thickness (mm) in each group. (E) Skin
   swelling by weight (mg) in each group. Values are expressed as mean ±
   SD (n = 5~6). # Indicates p < 0.05 and ## indicates p < 0.01 as
   compared to Normal group; *Indicates p < 0.05 and **indicates p < 0.01
   as compared to Model group.

   Skin swelling, typically as a result of inflammation or fluid retention
   in the skin, is an indicator of skin edema.[92]21,[93]22 In this study,
   skin edema was evaluated by comparing the difference in the
   thickness/weight among different groups. As shown in [94]Figures 6D and
   [95]E, the dorsal skin thickness and skin weight were gradually
   elevated in DNFB-treated AD mice. In contrast, compared to the model
   group, 0.5% luteolin substantially reduced dorsal skin thickness from
   (1.19 ± 0.28) to (0.72 ± 0.04) mm (p < 0.01) and alleviated the skin
   swelling by weight from (43.02 ± 8.71) to (25.38 ± 11.72) mg (p <
   0.05).

Luteolin Alleviates Histopathological Changes in DNFB-Induced AD Mice

   H&E staining was performed to examine the histopathological changes,
   including the epidermal hyperplasia and inflammatory cell infiltration,
   which are the major features of AD. In DNFB-induced AD mice, marked
   increases in epidermal thickness and infiltration of inflammatory cells
   in the dermis were observed ([96]Figure 7A). TB staining was further
   carried out to observe the infiltration of granulated and degranulated
   mast cells into inflamed tissues ([97]Figure 7B). Compared to normal
   mice, AD mice significantly increased the infiltration of mast cells,
   including granulated and degranulated forms (p < 0.05 or p < 0.01,
   [98]Figure 7C–E). However, these above mentioned pathological changes
   were both remarkedly reversed by topical treatment of 0.5% luteolin (p
   < 0.05 or p < 0.01, [99]Figures 7C–E).

Figure 7.

   [100]Figure 7
   [101]Open in a new tab

   Effects of luteolin on the histopathological changes of skin lesions in
   AD mice. (A) HE staining (magnification 200×; scale bar= 100 μm),
   straight line indicates thickness of the epidermis. (B) Toluidine blue
   (TB) staining (magnification 100×; scale bar =200 μm). The insets are
   magnified twice to observe granulated (◀) and degranulated (◁) mast
   cells in the skin stained with TB. Granulated mast cells appear round
   or oval granules. The granules show red and purple, and the nuclei
   located in the center of cells are blue and small visualized by TB
   staining. While degranulation of mast cells releases lots of particle
   matters, such as histamine, cytokines and eicosanoid mediators when
   they are activated by an allergen through their FcɛRI receptor. (C)
   Epidermal thickness (μm) was quantified as means in randomly selected
   in 3 fields of view per section. Infiltration of the (D) granulated and
   (E) degranulated mast cells in 3 fields of view per section. Values are
   expressed as mean ± SD (n = 3). # Indicates p < 0.05 and ## indicates p
   < 0.01 as compared to Normal group; * Indicates p < 0.05 and **
   indicates p < 0.01 as compared to Model group.

Luteolin Improves the Skin Barrier Function in DNFB-Induced AD Mice

   Clinically, both the affected and unaffected skin of AD patients
   exhibit epidermal barrier dysfunction, manifested by the enhanced
   transepidermal water loss (TEWL) and capacity reduced water retention.
   TEWL is a gold indicator used for evaluating the skin barrier
   function.[102]23 As shown in [103]Figure 8A, DNFB-sensitized skin
   showed a significant increase in TEWL value until the end of modeling
   (p < 0.01, Model [46.22 ± 5.61] g/m^2/h vs Normal [8.28 ± 3.95]
   g/m^2/h). 0.5% luteolin significantly reduced the TEWL value from 46.22
   ± 5.61 to 28.00 ± 2.17 g/m^2/h. The TEWL value in the 0.05% DEX group
   was 32.65 ± 7.52 g/m^2/h, which was significantly higher than that of
   the 0.5% luteolin group.

Figure 8.

   [104]Figure 8
   [105]Open in a new tab

   Luteolin recovers the impaired skin barrier. (A) Measurement of TEWL at
   day 12. (B) Immunofluorescence staining of FLG in skin lesions
   (magnification 100×). Values are expressed as mean ± SD (n =3~5). ##
   Indicates p < 0.01 as compared to Normal group; *Indicates p < 0.05 as
   compared to Model group.

   The degradation product of filaggrin (FLG) in normal skin is the
   important source of moisturizing factors (NMF) that are essential for
   maintaining stratum corneum hydration.[106]21 We further evaluated the
   expression of FLG protein using immunofluorescence staining
   ([107]Figure 8B). As expected, the positive expression of FLG (green)
   was significantly suppressed in the AD model compared with the Normal
   group, while luteolin reversed the FLG expression.

Luteolin Alleviates Th1, Th2 and Th17-Related Cytokines

   Emerging evidences show that elevated Th1/Th2/Th17-associated
   chemokines could be implicated in AD pathogenesis.[108]24,[109]25
   Network pharmacology analysis also revealed that IL-17 signaling
   pathway, Th17 cell differentiation, TNF signaling pathway, Th1 and Th2
   cell differentiation serve as the targets for luteolin against AD.
   Herein, we estimated Th1/Th2/Th17-associated cytokine levels in the
   serum and inflamed skin of DNFB-sensitized mice. As illustrated in
   [110]Figure 9, we observed that topical applied 0.5% luteolin markedly
   decreased the serum or skin levels of the Th2-related cytokine, IL-4
   (p<0.01 or p<0.05, [111]Figure 9A and [112]B) and Th17-related
   cytokines, such as IL-6 (p< 0.05, [113]Figure 9E and [114]F), TNF-α
   (p<0.01 or p<0.05, [115]Figure 9G and [116]H) and IL-17 (p<0.05,
   [117]Figure 9I and [118]J). Th1-related cytokines, such as IFN-γ, are
   reported to inhibit the differentiation of Th2 cells and the secretion
   of cytokines, thus restoring the impaired balance of Th1/Th2 immune
   responses in AD patients.[119]26 In this study, we also found that the
   production of IFN-γ in the skin was significantly increased in the
   group of mice treated with luteolin (p < 0.05, [120]Figure 9D) and a
   similar trend can also be seen in the serum IFN-γ level, while there
   was no statistically significant difference (p > 0.05, [121]Figure 9C).

Figure 9.

   [122]Figure 9
   [123]Open in a new tab

   Effects of luteolin on the levels of Th1, Th2 and Th17-related
   cytokines in AD mice. (A) Serum level of IL-4. (B) Skin level of IL-4.
   (C) Serum level of IFN-γ. (D) Skin level of IFN-γ. (E) Serum level of
   IL-6. (F) Skin level of IL-6. (G) Serum level of TNF-α. (H) Skin level
   of TNF-α. (I) Serum level of IL-17. (J) Skin level of IL-17. Values are
   expressed as mean ± SD (n =5). # Indicates p < 0.05 and ## indicates p
   < 0.01 as compared to Normal group; *Indicates p < 0.05 and **indicates
   p < 0.01 as compared to Model group.

Luteolin Inhibits Jak2/Stat3 Signaling Pathway Activation

   The results of KEGG pathway enrichment in the network pharmacology
   study demonstrated the mechanism of luteolin in treating AD was
   significantly associated with the Jak-STAT signaling pathway. Thus, we
   examined the protein expression levels of phosphorylated (p)-JAK2 and
   p-STAT3 in dorsal skin by immunohistochemistry staining to explore
   whether the JAK2/STAT3 pathway is involved in the anti-AD effects of
   luteolin ([124]Figures 10A and [125]B). Compared with the normal group,
   the protein expression of p-JAK2 and p-STAT3 in paraffin section of
   skin biopsies was noticeably increased about 1.36-fold and 1.40-fold in
   the DNFB group (p < 0.01, [126]Figures 10C), while luteolin prominently
   reduced the increased levels of p-JAK2 and p-STAT3 (p < 0.05 or p <
   0.01, [127]Figure 10C).

Figure 10.

   [128]Figure 10
   [129]Open in a new tab

   Effect of luteolin on JAK2/STAT3 signaling pathway in skin biopsies.
   Immunohistochemistry staining images for (A) p-JAK2 and (B) p-STAT3 in
   paraffin sections of skin biopsies (magnification 200×). (C) Average
   optical density (AOD) analysis of p-JAK2 and p-STAT3. Values are
   expressed as mean ± SD (n =3). ## Indicates p < 0.01 as compared to
   Normal group; *Indicates p < 0.05 and **indicates p < 0.01 as compared
   to Model group.

Discussion

   AD is one of the most common skin diseases. Recurrent eczematous
   lesions and intense pruritus seriously affect social interactions and
   quality of life of AD patients.[130]1,[131]2 The standard therapy for
   AD in clinic mainly involves the corticosteroids or immunosuppressive
   agents. However, there is still an urgent need for more effective and
   fewer side effects anti-AD therapeutics because of their limited
   efficacy and unwanted additional adverse effects.[132]4,[133]27 To
   date, various studies have shown that luteolin possesses
   pharmacological activities such as anti-inflammatory, antioxidant and
   anti-allergic, which exhibits its great potential in the treatment of
   AD.[134]9–11 Zhang et al[135]28 previously demonstrated that luteolin
   could effectively diminish collagen production and attenuate skin
   fibrosis via the inhibition of TGF-β/Smads signaling mediated by
   binding to ALK5. Moreover, it has been reported that topical
   application of luteolin significantly inhibited the scratching behavior
   associated with allergic cutaneous anaphylaxis.[136]12 However, the
   effects and underlying mechanisms of luteolin against AD have not been
   currently defined. In the present study, we firstly conducted a
   systemic study via a combination of network pharmacology molecular
   docking and in vivo studies to explore the possible mechanism of
   luteolin in treating AD.

   31 overlapping targets of luteolin against AD were obtained and PPI
   network result indicated luteolin may need to simultaneously regulate
   multiple anti-inflammatory or immune-regulatory targets to attenuate
   AD-like symptoms rather than single target. Twenty core gene targets
   were screened out by topology analysis, and they were MMP9, IL6, TNF,
   IL10, VEGFA, IL4, ICAM1, NFKBIA, IFNG, etc. Among them, IL6, IL10, IL2,
   and IL4 belong to the interleukin family, which play critical role in
   regulating immune and inflammatory processes.[137]29,[138]30 MAPK8 and
   MAPK1, members of the MAPK family, are reported to mediate
   rapid responses to abiotic stress stimuli.[139]31,[140]32 The targets
   closely related to the chemotaxis and adhesion function of leukocytes
   are ICAM1 and MPO.[141]33,[142]34 Targets closely related to immune
   regulation are IL10, IFNG, IL-4, IL2 and CD40LG, AHR.[143]35,[144]36
   Targets closely related to angiogenesis are EGFR and
   VEGFA.[145]37,[146]38 Immune disorder, leukocyte chemotaxis and
   angiogenesis are reported to play critical roles in the acute and
   chronic phases of AD.[147]39,[148]40 Altogether, the above results
   showed that luteolin could effectively improve AD-like symptoms through
   multiple targets and pathological process.

   To further reveal the biological functions of luteolin in AD treatment,
   we performed GO and KEGG enrichment analysis of the common targets. GO
   analysis results showed revealed that these targets were correlated
   with the regulation of cytokines, cytokine-mediated signaling pathway,
   inflammatory response, and other biological processes. Additionally,
   KEGG pathway enrichment analysis suggested that common targets were
   enriched mainly in IL-17 signaling pathway, T cell receptor signaling
   pathway, Th17 cell differentiation, TNF signaling pathway, JAK/STAT
   signaling pathway, Th1 and Th2 cell differentiation, etc., which were
   the critical mechanism of luteolin against AD. Studies have confirmed
   that AD is a typical chronic inflammatory skin disease, largely
   mediated by T cells. Various T-cell subsets, such as Th1, Th2, Th17,
   and CD4^+ T cells are critical in immune pathogenesis and chronic
   inflammatory responses in AD.[149]41,[150]42 The dominance
   of Th2 over Th1 type response is recognized as a main hallmark of AD. A
   biased Th2-mediated immune response can lead to Th1/Th2 immune
   imbalance and skin barrier disruption, which triggers the occurrence
   and development of AD.[151]41 Besides, Th17 cells and IL-17 signaling
   have also been proved to be implicated in both the initiation and
   maintenance of AD.[152]24,[153]41 Th17-related cytokines, IL-17, IL-6
   and TNF-α, can activate Th2 immune response and impair skin barrier
   function.[154]43 Consistent with the results predicted from network
   pharmacology analysis, our results demonstrated that topical luteolin
   treatment effectively reduced the secretion of Th2-related cytokine
   (IL-4) and Th17-related cytokines (IL-17, TNF-α and IL-6) and elevated
   the level of Th1-type cytokine (IFN-γ) in dorsal skin, thus restoring
   the Th1/Th2 immune balance in AD.

   The JAK/STAT signaling pathway, a widely expressed intracellular signal
   transduction pathway, is reported to involve in many immune and
   inflammatory diseases, among which the JAK2/STAT3 pathway has been
   proposed as a well-accepted therapeutic target against
   AD.[155]44,[156]45 Previous studies reported that activated JAK2/STAT3
   signaling could upregulate the Th2 immune response and induce the
   secretion of excessive inflammatory cytokines and IgE, thus
   exacerbating skin inflammation.[157]46,[158]47 In the present research,
   KEGG pathway analysis predicted that JAK/STAT signaling could be the
   crucial mechanism of luteolin against AD and the core targets, such as
   IFNG, IL4, IL6 etc., were significantly enriched in the JAK/STAT
   signaling pathway. Taking the JAK2/STAT3 signaling as the breakthrough,
   the protein levels of p-JAK2 and p-STAT3 in skin tissues were further
   examined. In this study, we found that the expression of p-JAK2 and
   p-STAT3 were significantly increased in the dorsal skin of AD mice,
   exhibiting the activation of JAK2/STAT3 signaling pathway, while this
   phenomenon was reversed by topical application of luteolin. We here
   concluded that luteolin exhibits protective effects on AD-like skin
   injury potentially by inhibiting the JAK2/STAT3 signaling, thereby
   significantly reducing expression of inflammatory cytokines and
   alleviating the inflammatory responses.

Conclusion

   To summarize, this study has identified the core targets and key
   pathways of luteolin against AD through network pharmacology combined
   with the experimental validation. Our study suggests that luteolin can
   ameliorate the AD-like lesions through the synergistic interaction
   between multiple core targets and pathways, and Th1/Th2/Th17-related
   cytokines and JAK2/STAT3 signaling pathway are the possible core
   targets. Although the in-depth mechanism is warranted, the current
   research provides a basis for the clinical application and basic
   research of luteolin in the treatment of AD.

Funding Statement

   This research was supported by the Fundamental Research Funds for the
   Central Universities, grant number 2042022kf1077.

Author Contributions

   All authors made a significant contribution to the work reported,
   whether that is in the conception, study design, execution, acquisition
   of data, analysis and interpretation, or in all these areas; took part
   in drafting, revising or critically reviewing the article; gave final
   approval of the version to be published; have agreed on the journal to
   which the article has been submitted; and agree to be accountable for
   all aspects of the work.

Disclosure

   The authors report no conflict of interest in this work.

References