Therapeutic effects and mechanism of action of lavender essential oil on atopic dermatitis by modulating the STAT3/RORγt pathway

2.1 Materials

Lavender essential oil (Xinjiang Ipak Khan Spices Co., Ltd., Yi Ning, China), DNCB (Chengdu Kolon Chemical Co., Chengdu, China), acetone (Tianjin Tian li Chemical Reagent Co., Tianjin, China), and olive oil (Poli Aromatic Pharmaceutical Technology Co., Shanghai, China) were purchased.

2.2 Experimental animals

Specific pathogen-free (SPF)–grade healthy male Kunming mice (18–22 g) were purchased from Chengdu Da shuo Company Limited (Shaanxi Province, China), with animal certificate number. SCXK (Chuan) 2020–030. All experimental procedures were approved by the Animal Ethics Committee of Shaanxi University of Traditional Chinese Medicine (No. SUCMDL20220805001). The mice were acclimatized for 7 days and maintained under standard laboratory temperature and humidity conditions.

2.3 Establishment of DNCB-induced AD model in mice

The mice were randomly divided into six groups: normal group, model group, positive drug group (compound dexamethasone acetate cream), 2 % LEO low-dose group, 4 % LEO medium-dose group, and 8 % LEO high-dose group. The dose groups were all diluted with acetone olive oil solution (4:1). Each group had 10 mice. After grouping, shave approximately 2 cm × 2 cm of fur from the back of the mouse using a razor In the normal group, 100 µL of acetone olive oil solution was used as the control vehicle, and in the remaining five groups, The mice were sensitized by administering a 100 µL solution of acetone olive oil (4:1) containing 7 % DNCB to their back skin (Li et al., 2023). After 6 days, the mice were topically treated with 30 μL of acetone olive oil solution on their left ear as a control, and 30 μL of acetone olive oil solution with 1 % DNCB on their right ear for stimulation, and the stimulation experiment was repeated every 2 days for the sum of three times. One hour after the first sensitization, the sensitizer was washed with a 0.9 % mass fraction of sodium chloride solution, after which the test drug was applied, and the same dose of acetone olive oil solution was utilized for the model group. Mice were dosed twice daily for 12 days. One day after the last administration, mice were allowed to bleed and then euthanized, and dorsal skin and spleen tissue samples were collected and stored in 4 % tissue fixative and in a −80 °C refrigerator.

2.4 Determination of IL-6, TNF-α using ELISA

Mice sera were tested using an enzyme-linked immunosorbent assay (ELISA) kit (Jiangsu Meimian Industrial Co.,Ltd) and cytokine levels in each group were determined according to the instructions. Standard curves were constructed and the concentrations of interleukin-6 (IL-6) and tumour necrosis factor-α (TNF-α) were calculated in each group (Xiaojing et al., 2023).

2.5 Characterization of the chemical composition of LEO

2.6 Screening of LEO for identifying targets of active ingredients

Based on the active ingredients identified in GC–MS, Swiss Target Prediction (https://www.swisstargetprediction.ch/),the PubChem (https://pubchem.ncbi.nlm.nih.gov/), Gene Cards (https://www.genecards.org/), Target Net (https://targetnet.scbdd.com/), and The Comparative Toxicogenomics Database (https://ctdbase.org/) were reviewed to determine the active ingredients of LEO and their AD disease targets. And based on the GEO database GSE32924 microarray assay data, Differential gene analysis was performed using the Limma package for R. The screening conditions were |logFC|>1 and P < 0.05 The intersection of differential genes, LEO component targets, and AD disease targets in the GSE32924 data was then screened using Venny 2.1.0 (https://bioinfogp.cnb.csic.es/tools/venny/index.html). Cross-targets are fed into the online STRING (https://string-db.org/) platform to construct PPI network. The information was then entered into the Cytoscape 3.7.2 software for analysis. and into Cytoscape 3.7.2 (Kohl et al., 2011) software to construct an “AD-active component-AD potential target” network for LEO and analyze the network using the software's built-in network analyser tool. Furthermore, characteristic parameters were used to investigate the important components present in LEO, their targets, and the relationship between them. Gene ontology–biological process (GO-BP) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed using R Language (version 4.2.1), and the results are obtained from the enrichment analysis and then re-ranked using the method based on their weighting coefficients.

2.7 Determination of weighting factors

Since the drug is administered transdermally, we introduced an important parameter of transdermal absorption of drug, logP value, to convert the transdermal absorption rate. C_o means that the drug can be completely absorbed, while C_w means the part of the drug that is not absorbed. X indicates the absorption rate of the volatile oil (Wang et al., 2021a). Thus, we related the logP value, relative content and active ingredient to obtain the following formula Eqs. (2)–(6).

In this formula, $w$ is the relative amount of each active ingredient, Y is the sum of the scores of the target's related compounds, and S is the sum of the scores of the targets included in the pathway.

2.8 Molecular docking

We obtained the 3D structures of the important targets from the PDB database (https://www.rcsb.org/) (Burley et al., 2021) and the 2D structures of the ligands from the PubChem database. Then, we used the LibDock module of Discovery Studio 4.5 software to dock the target protein with its active ingredient and positive drug.

2.9 Metabolomics studies

2.10 Histological testing

The back skin tissue samples fixed in 4 % paraformaldehyde were dehydrated using a gradient of anhydrous ethanol and treated with xylene to increase transparency. Paraffin-embedded, 4-μm-thick tissue sections were cut, then baked on slides, dewaxed, and stained with HE and toluidine blue to observe histopathology under a microscope (Yurong et al., 2023).

2.11 Immunohistochemical testing

The sections were deparaffinized by baking in xylene and then treated with 3 % hydrogen peroxide for 25 min. They were then washed three times in phosphate-buffered saline (PBS) and incubated with 10 % normal goat serum for 30 min. After removing the serum, IL-6, IL-17A, Forkhead box protein 3 (FOXP3) and Phosphorylated JAK2 (p-JAK2) were added and incubated overnight at 4 °C The sections were then washed three times with PBS. After the introduction of secondary antibodies, the sections were incubated at 37 °C for 50 min. Dropwise addition of DAB solution was done for color development, and sections were stained again with hematoxylin solution for 2 min, dehydrated, sealed, and then observed under a microscope (Xiaoqiong, 2022).

2.12 Western blot experiments

The skin tissue from each group was lysed with the prepared protein lysate, and the tissue and lysate were mixed thoroughly and positioned on ice for 5 min. The contents were spun down at 12,000 rpm for 10 min at 4 °C, and the resulting supernatant was isolated. The collected supernatant was utilized as the protein extract. and the total protein concentration was determined using the BCA method. After protein quantification, proteins were upsampled at 4 μg/mL and separated by 10 % SDS-PAGE and then transferred to a PVDF membrane. The membranes were blocked using 5 % powdered skim milk, followed by incubation using antibodies against STAT3 (ab68153; Abcam, China), Phosphorylated STAT3 (p-STAT3) (bs-1658R; Boossen, China), and RORγt (ER1916-09; HUAbio, China). The membranes were then washed with TBST and incubated with secondary antibodies for 1 h. The membranes were washed three times with TBST, and then ECL chemiluminescence solution was added to the top surface of the membranes in a dark box for 5 min protected from light. β-Actin (66009-1-Ig; Proteintech, China) was used as an internal reference using a gel imaging system (Xiaojing et al., 2023).

2.13 Analysis of the data using statistical methods

All experimental data in this study were subjected to statistical analysis using GraphPad Prism 8.0 software. and the results are presented as $\overline{x}$  ± SD. One-way analysis of variance (ANOVA) were used to analyze statistically significant differences between two or more groups. P < 0.05 was considered to indicate a statistically significant difference between two groups.

3.1 Effect of LEO on DNCB-induced mice models

DNCB induction leads to signs of bleeding, edema, vesicles, flaking, and dryness in the skin of mice (Zhang et al., 2022). Based on animal experiments in a mice dermatitis model induced by DNCB, it was shown that the symptoms of AD were treated with different concentrations of LEO, and the skin condition was observed and found to be significantly relieved (Fig. 2A).

Close

Fig. 2

(A) Lavender essential oil treatment ameliorates DNCB-induced atopic dermatitis lesions and (B-C) Effect of lavender essential oil on IL-6 and TNF-α levels in DNCB model (Data are presented as $\overline{x}$  ± SD, n = 6. ^****P < 0.0001, ^***P < 0.001.).

Export to PPT

3.2 Serum levels of TNF-α and IL-6

To further investigate the effect of LEO on AD treatment, serum concentrations of IL-6 and TNF-α were quantified. Compared with the normal group, TNF-α (^***P < 0.001, n = 6) and IL-6 (^****P < 0.0001, n = 6) levels were significantly increased in the model group, and compared to model group, the LEO dose group showed a decrease in TNF-α and IL-6 levels (Fig. 2B–C).

3.3 GC–MS composition of LEO

The chemical composition of LEO was determined by GC–MS, see Fig. 3A. 17 chemical components were identified, the main components of which were linalool (20.47 %) and linalyl acetate (14.989 %), the results are shown in Table 1.

Close

Fig. 3

(A) Total ion current map of LEO. (B) The differential genetic volcano map of AD (C) Heat map of differential genes in AD (D) Intersection map of LEO targets, eczema targets, and differentially expressed genes. (E) PPI network diagram. (F) Network diagram of LEO for AD disorders – active ingredients – core targets.

Export to PPT

3.4 Network pharmacology analysis

3.4.3

3.4.3 Results of GO-BP and KEGG enrichment analyses

GO-BP and KEGG enrichment analyses of 33 intersecting targets were performed using the R Language. The analysis resulted in the identification of 230 biological processes and 28 signaling pathways (Fig. 4A–B). GO-BP is mainly involved in the growth hormone receptor signaling pathway, cellular response to growth hormone stimulus, regulation of ubiquitin-dependent protein catabolic process, and other pathways. KEGG analysis revealed enriched pathways such as Th17 cell differentiation, PD-1/PD-L1 in Cancer and Th1 /Th2 cell differentiation, and other pathways. Fig. 4C–D shows data after recalculation and reordering of the weighting factors. Th17 cell differentiation was still ranked first, and Th17 cell differentiation was finally identified as the most important pathway.

Close

Fig. 4

(A) Plot of KEGG enrichment analysis of LEO (B) Plot of GO-BP analysis of LEO. (C) Plot of KEGG enrichment analysis of LEO after ranking with the weighting factor method. (D) GO-BP analysis of LEO after sorting by the weighting factor method.

Export to PPT

3.5 Molecular docking results

The selection of key active components of LEO for molecular docking with key targets of the Th17 cell differentiation pathway was performed in the LibDock section of Discovery Studio 4.0 software. The docking results are based on the docking score when comparing the active ingredient with the specific ligand. If an active ingredient has the highest score and is comparable to the specific ligand score, then the active ingredient works best at that target site and has a greater role in the Th17 cell differentiation pathway. The results suggest that linalool exhibits optimal binding affinity towards crucial targets (RORγt, FOXP3) and establishes stable interactions, with scores similar to those of specific ligands (celastrol, chloroquine). Similarly, linalyl acetate demonstrates optimal binding activity and forms stable interactions with key targets (STAT3, IL-6), yielding scores comparable to specific ligands (Dibromotyrosine, Rivanicline). The RORγt target proteins, leucine (LEU), alanine (ALA), and valine (VAL), engage in hydrophobic interactions with Linalool, specifically alkyl-to-alkyl and alkyl-to-unsaturated bonds. Additionally, arginine (ARG) forms a hydrogen bond with Linalool. On the other hand, the Foxp3 target proteins, leucine (LEU), valine (VAL), tyrosine (TYR), and phenylalanine (PHE), also form hydrophobic interactions with Linalool. However, methionine (MET) exhibits an unfavorable binding with Linalool, as it is capable of binding in plan but not in stereo. Among the STAT3 target proteins, ALA, VAL, isoleucine (ILE), and cysteine (CYS) engage in hydrophobic interactions with Linalyl acetate, while ARG, glutamine (GLN), and Linalyl acetate form hydrogen bonds with each other. Similarly, the IL-6 target proteins MET, LEU, and LYS exhibit hydrophobic interactions with Linalyl acetate, respectively. This suggests that linalool and linalyl acetate could be the key active ingredients in LEO for AD treatment (Fig. 5).

Close

Fig. 5

(A) Interaction diagram of IL-6 and rivanicline. (B) Interaction diagram of IL-6 and linalyl acetate. (C) Interaction diagram of STAT3 and dibromotyrosine. (D) Interaction diagram of STAT3 and linalyl acetate. (E) Interaction diagram of RORγt and selisistat. (F) Interaction diagram of RORγt and linalool. (G) Interaction diagram of FOXP3 and chloroquine. (H) Interaction diagram of FOXP3 and linalool. (I) Heat map of molecular docking scores. The interactions of key amino acids and structures are shown, with dark green representing conventional hydrogen bonding, light green representing van der Waals forces, red representing unfavourable binding, dark purple representing π-σ, and light purple representing alkyl groups. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Export to PPT

3.6 Metabolomics analysis

Analysis of mice serum samples by GC–MS with simultaneous collection of serum QC samples demonstrates instrument stability. A 2D scatter plot generated by the PCA highlighted clear distinctions between the normal group and the model and high-dose groups. (Fig. 6A). This result indicates that LEO caused changes in metabolites in vivo and that differences were observed in metabolic levels between the groups of samples (Li et al., 2021a). To further mine the data for metabolite information, we used supervised multidimensional statistical methods (i.e., OPLS-DA) and further analyzed the normal and model groups separately (R²X = 0.315, R²Y = 0.985, Q² = 0.737); the model and dosing group samples (R²X = 0.453, R²Y = 0.996, Q² = 0.923; Fig. 6B, C). Both groups showed good separation, with improved R²Y and Q² values, and yielded more coherent interpretations and cross-validated predictions about the data for the sample classification. in addition to validating the validity and adaptability of the OPLS-DA model using permutation tests (200 times; Fig. 6D, E).

Close

Fig. 6

Metabolomics analysis. (A) PCA for normal, model, dosing, and QC groups. (B) Normal vs model OPLS-DA score plots. (C) OPLS-DA score plot for model vs dosing. (D-E) Normal vs model, model vs administered OPLS-DA model random substitution test (n = 200). Green triangles are labeled R², and blue boxes are labeled Q². (F-G) Heat map of metabolites with differential expression between normal vs model and model vs dosing. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Export to PPT

In OPLS-DA and t-test analyses using SIMCA 14.1 software, with VIP > 1 and P < 0.05 as screening criteria, 12 differentially expressed metabolites were identified in the normal model group, with five metabolites having upregulated expression in the normal group compared with that in the model group, including Caryophyllene, Tetradecanoic acid, Stearic acid, Dodecanoic acid, Urea. And seven having downregulated expression in the normal group compared with that in the model group including Phytane, mio-Inositol, 1,3-Bisurea, l-Alanine, Glyceric acid, Butanedioic acid and DL-Pyroglutamic acid. (see Table 2 for details). Thirty-two differentially expressed metabolites were identified in the model-administered group, with seven metabolites having upregulated expression and 25 metabolites having downregulated expression compared with those of the model and administered groups (see Table 3 for details). We generated heat maps based on the two sets of differentially expressed metabolite changes (Fig. 6F, G). To further investigate the metabolic pathways regulated by LEO, we introduced these differentially expressed metabolites into MetaboAnalyst 5.0. Fig. 7 displays the outcomes of the analysis. Based on the pathway impact of > 0.01, six major metabolic pathways were screened in the normal model group, mainly including the phosphatidylinositol signaling system, glycerolipid metabolism, glyoxylate and dicarboxylate metabolism, and other metabolic pathways. Thirteen major metabolic pathways were identified in the screening model administration group. The main pathways include linoleic acid (LA) metabolism D-Glutamine and D-glutamate metabolism phenylalanine, tyrosine, and other metabolic pathways.

Close

Fig. 7

(A-B) Metabolic pathways showing important metabolites in normal vs model and model vs dosing.

Export to PPT

3.7 Combined network pharmacology and metabolomics analysis

In order to comprehensively understand how LEO operates as a treatment for AD, we used network pharmacology and metabolomics to construct a network of interactions (depicted in Fig. 8). Using the MetScape plugin in Cytoscape software, we integrated the screened differential genes and differential metabolites to construct a compound-response-enzyme-gene network (Li et al., 2021b). By matching differential genes to compounds, we identified two key targets: HPGD and PTGS1 (see dark blue circles in Fig. 8). In addition, we identified key metabolites such as linoleic acid and arachidonic acid (indicated by red hexagons in Fig. 8). Affected metabolic pathways include Linoleic acid metabolism, Arachidonic acid metabolism. we then confirmed the expression of these genes which may play an important role in the efficacy of LEO as a treatment for AD.

Close

Fig. 8

Compound-reaction-enzyme-gene networks for key metabolites and genes of arachidonic acid metabolism and linoleic acid metabolism.

Export to PPT

3.8 LEO inhibits DNCB-induced epidermal thickening and mast cell infiltration in the mice model

The analysis of HE staining (Fig. 10A) revealed that the epidermal tissue of showed that the epidermal tissue of the normal group had a normal structure, with no evident hyperkeratosis of the stratum corneum and no significant thickening of the skin. By contrast, the model group showed a marked thickening of the tissue epidermis and a large number of blood crusts formed by leukocyte exudation and platelet coagulation were visible. Positive drug observation shows mild abnormalities in skin tissue structure, with slight epidermal hyperplasia, tight epidermal-dermal junctions, and a small amount of inflammatory cell infiltration visible in the tissue. In the treatment group, as the dose increased, a small amount of new epidermis was visible in the tissue, and cellular edema was significantly reduced (Kim et al., 2022).

When the tissue was stained with TB (Fig. 10B) as a mast cell marker, the normal group showed a little quantity of mast cell infiltration, and the model group had massive mast cell infiltration and a significantly higher number of mast cells than the normal group. In the positive drug group, the number of mast cells decreased with an increasing dose in the low-dose, medium-dose, and high-dose LEO groups. According to the results of mast cell count, mast cells were significantly higher in the model group than in the blank group, and mast cells were significantly lower in the LEO dose group compared to the model group (Fig. 10E).

3.9 Immunohistochemical staining results

IHC staining results (Fig. 9A) showed that IL-6 and IL-17A were both pro-inflammatory factors, and the brown area of their corresponding staining decreased with increasing doses of IL-6 and IL-17A inflammatory factors when compared with that in the model group. The protein expression level of p-JAK2 was reduced in different doses of LEO compared to the model group, and its staining was lighter in brown color. FOXP3 exhibited increased levels of FOXP3 protein expression and deepened brown staining in different doses of LEO compared to the model group. In Fig. 9B, significant differences in the expression levels of IL-6 (^**P < 0.01, n = 3), IL-17A (^**P < 0.01, n = 3), p-JAK2 (^**P < 0.01, n = 3), and FOXP3 (^**P < 0.01, n = 3) can be seen between the LEO high-dose and model group.

Close

Fig. 9

(A)IHC detection of FOXP3, IL-6, IL-17A and p-JAK2 expression in dorsal skin tissues of mice (×400 magnification) (B) Qualitative outcomes of FOXP3, IL-6, IL-17A, and p-JAK2. Data are presented as $\overline{x}$  ± SD, n = 3. ^##P < 0.01, ^##P < 0.01, ^##P < 0.01, ^##P < 0.01 vs control; ^**P < 0.01, ^**P < 0.01, ^**P < 0.01, ^**P < 0.01 vs model group.

Export to PPT

3.10 Western blot analysis results

Western blot analysis was performed to determine the protein expression of STAT3 and RORγt, which serve as regulatory factors for Th17 cell differentiation. the expression of p-STAT3/STAT3 (^***P < 0.001, n = 3) and RORγt/β-actin (^****P < 0.0001, n = 3) was dramatically raised in the model group, and compared with the model group, the LEO high-dose group dramatically reduced the expression of RORγt/β-actin (^***P < 0.001, n = 3), p-STAT3/STAT3 (^**P < 0.01, n = 3) protein expression (Bai et al., 2020) (Fig. 10C–D).

Close

Fig. 10

(A) Hematoxylin-eosin staining results of mice skin tissues (×100 magnification, Note 1: normal, 2: model 3: positive drug, 4: low-dose, 5: medium-dose 6: high-dose). As shown in the figure yellow arrow shows the epidermis, red arrow shows the dermal follicle, green arrow shows large necrosis visible in the dermis of the tissue, black arrow shows infiltration of inflammatory cells and purple arrow shows slight thickening visible in the epidermis of the tissue. (B) Toluidine blue staining results, black arrows show mast cells (×100, ×400 magnification). (C) Results of RORγt and STAT3 protein expression and their phosphorylation levels. (D) Quantitative analysis of RORγt/β-actin, p-STAT3/STAT3. Data are presented as $\overline{x}$  ± SD, n = 3. ^####P < 0.0001, ^###P < 0.001 vs control; ^***P < 0.001, ^**P < 0.01 vs model group. (E) Mast cell counts shown after toluidine blue staining. Data are presented as $\overline{x}$  ± SD, n = 3. ^###P < 0.001 vs control; ^**P < 0.01 vs model group. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Export to PPT