Multi-omics combined analysis of turmeric volatile oil alleviating allergic rhinitis through the JNK/MAPK-Relb/NF-κB pathway

Abstract

We extracted turmeric volatile oil (TVO) from turmeric using supercritical CO₂ extraction for potential allergic rhinitis (AR) therapy. The volatile components of TVO were analyzed using gas chromatography mass spectrometry and gas chromatography-ion migration spectrometry. The pharmacological effects of TVO on ovalbumin (OVA)-induced AR in mice were evaluated through pathological markers, microbiomics, metabolomics, and proteomics. An inflammatory model using human mast cells (HMCs) was established to assess whether TVO could mitigate inflammation. TVO effectively alleviated AR symptoms in OVA-sensitized mice and reduced inflammation in lipopolysaccharide- and interferon-γ-treated HMCs. TVO suppressed the inflammatory cytokine production, restored the nasal microbiota composition, and regulated serum metabolic profiles. Proteomic analysis using four-dimensional data-independent acquisition indicated that the mitogen-activated protein kinase (MAPK) and transcription factor nuclear factor-κB (NF-κB) signaling pathways are involved in TVO’s anti-inflammatory mechanism. Accordingly, TVO inhibited the activation of MAPK and NF-κB pathways in the OVA mouse model and the inflammatory HMCs model. Jun N-terminal kinase and Relb were identified as key mediators in the therapeutic action of TVO, with Relb knockdown enhancing its anti-inflammatory effect, suggesting that TVO has potential as a therapeutic agent for alleviating AR.

Introduction

The inflammation of the nasal mucosa is known as allergic rhinitis (AR) and is mediated by immunoglobulin E (IgE)¹. Upon re-exposure to the same allergen, the antigen cross-links IgE-FcεRI complexes anchored on mast-cell and basophil membranes, triggering the immediate release of histamine (HIS), Leukotriene C4 (LTC4), Prostaglandin D2 (PGD2), and other mediators that evoke sneezing and nasal itching^2,3. Current epidemiological data suggest that this disorder is estimated to afflict 10–20% of individuals across the globe⁴. The most common triggers include environmental allergens⁵. Pollen and dust allergies can significantly impair quality of life⁶ and result in considerable indirect economic losses⁷. Existing treatments, including antihistamines and corticosteroids, offer only symptomatic relief and may cause adverse effects when used long-term⁸. Therefore, this underscores the pressing need for safer and more effective therapeutic alternatives.

The pathogenesis of AR has been linked to many signaling pathways, including the PI3K/AKT and cAMP pathways. Among them, the MAPK and NF-κB signaling pathways play a dominant part throughout the course of the disease⁹. Members of the MAPK family—including extracellular signal-regulated kinase (ERK)1/2, Jun N-terminal kinase (JNK), and p38—transmit extracellular stimuli to the nucleus¹⁰. JNK and p38, in particular, are closely related to pro-inflammatory mediators¹¹. The transcription factor NF-κB regulates the expression of genes involved in inflammation. NF-κB family members include NF-κB1, NF-κB2, RelA, RelB, and c-Rel, which can form either homo- or hetero-dimers; among these, the NF-κB1/RelA heterodimer is predominant in many cell types. The transcriptional activation of various pro-inflammatory cytokines is central to NF-κB’s role, with tumor necrosis factor (TNF)-α and interleukin (IL)-6 being two of its primary downstream targets¹². In its inactive state, NF-κB is sequestered in the cytoplasm when it binds to IκB family proteins such as IκB-α. NF-κB can move into the core and trigger pro-inflammatory gene transcription when IκB is stimulated because it is phosphorylated and degraded by the IKK complex¹³. The inflammatory mediators regulated by MAPK and NF-κB pathways are stored in mast cell granules. Upon calcium influx, these granules are mobilized to the cell membrane, where they merge and discharge their cargo into the extracellular milieu. These inflammatory mediators play key roles in triggering symptoms associated with allergic diseases¹⁴.

16S ribosomal RNA (16S rRNA) sequencing is a high-throughput method that is used for precise identification and quantification of microbial populations in environmental and biological samples. It facilitates a rigorous scientific foundation for early disease diagnosis and microbe-targeted therapies. Proteomics focuses on identifying and characterizing proteins in cells, tissues, or bodily fluids, offering insights into biological functions and potential therapeutic targets¹⁵. Metabolomics quantitatively analyzes metabolites, enabling continuous monitoring of their changes and correlations with disease states¹⁶. Together, multi-omics approaches—which include microbiomics, proteomics, and metabolomics—offer comprehensive insights into the relationships among microbial communities, metabolic pathways, and protein expression. These approaches facilitate the construction of interaction networks to identify key regulatory nodes, offering a deeper understanding of disease mechanisms that may be overlooked by single-omics analyzes¹⁷.

Turmeric volatile oil (TVO) is primarily extracted from the plant Curcuma longa L. and represents one of its main active components¹⁸. Commonly used as a food additive for its color and flavor¹⁹, TVO has demonstrated pharmacological properties such as anti-inflammatory and analgesic activities^20–22, and is considered safe and non-toxic²³. However, the specific active constituents and mechanisms underlying TVO’s potential effects in treating AR remain unclear. These preliminary findings suggest a promising direction for further research into TVO as a therapeutic agent for alleviating AR and elucidating its mechanisms of action.

Results

Composition analysis of volatile turmeric oil

The single-factor and response surface results for TVO are shown in Supplementary Fig. 1. The ideal extraction parameters were confirmed to be an extraction temperature of 50 °C, an extraction flow rate of 40 L/h, and an extraction pressure of 28 MPa. The GC-MS spectra of TVO are presented in Fig. 1A and Table 1. In total, 44 chemical components were identified (Fig. 1B), including 2 aldehydes, 7 ketones, 11 alcohols, 10 alkenes, 4 acids, 2 phenols, 1 base, 2 esters, and 5 others. The relative proportions of alcohols (25%), ketones (15.90%), and alkenes (22.70%) were relatively high. The relative arylcurcumin content was 43.36% (Fig. 1C), while the relative curcumin content was 17.70% (Fig. 1D).

Fig. 1. GC-MS and GC-IMS analysis of TVO.

A Total ion current spectrum. B Volatile component statistics of GC-MS. C Structural formula of arylcurcumin. D Structural formula of curcumin. E Three-dimensional spectrum. F Two-dimensional spectrum. G Volatile component statistics of GC-IMS. H Proportion of different types of volatile compounds.

Table 1.

Volatile components of TVO were detected using GC-MS

Number	Rt (min)	Compound name	CAS	Formula	Peak area	Relative content%
1	14.767	m-Cymene	535-77-3	C10H14	722,026	0.01
2	14.984	Eucalyptol	470-82-6	C10H18O	5,196,116	0.05
3	16.621	4-Acetyl-1-methylcyclohexene	70286-20-3	C9H14O	867,723	0.01
4	17.502	p-Cymen-8-ol	1197-01-9	C10H14O	2,860,250	0.03
5	17.68	Terpieol	10482-56-1	C10H18O	1,432,464	0.01
6	18.903	6-Ethyl-3,4-dimethylphenol	2219-78-5	C10H14O	2,379,646	0.02
7	19.008	Anethol	104-46-1	C10H12O	3,594,280	0.03
8	19.356	4,4,7a-Trimethyl-2,4,5,6,7,7a-hexahydro-1H-inden-1-one	60713-96-4	C12H18O	931,470	0.01
9	19.554	exo-2-Hydroxycineole	92999-78-5	C10H18O2	1,483,100	0.01
10	19.922	Eugenol	97-53-0	C10H12O2	862,935	0.01
11	20.283	Megastigma-4,6(E),8(Z)-triene	71186-24-8	C13H20	6,443,847	0.06
12	20.573	Isovanillin	621-59-0	C8H8O3	9,719,752	0.09
13	20.829	Bergamotol, Z-πtrans-	88034-74-6	C15H24O	3,126,451	0.03
14	20.941	Santalen	512-61-8	C15H24	31,057,253	0.29
15	21.335	Seychellene	20085-93-2	C15H24	30,919,069	0.29
16	21.644	Curcumene	644-30-4	C15H22	188,085,024	1.76
17	21.802	Zingiberene	495-60-3	C15H24	153,093,065	1.43
18	21.973	Bisabolene	495-61-4	C15H24	79,155,778	0.74
19	22.223	Sesquiphellandrene	20307-83-9	C15H24	219,942,269	2.05
20	22.401	Caryophyllene oxide	1139-30-6	C15H24O	15,181,712	0.14
21	23.038	Cubenol	21284-22-0	C15H26O	67,715,207	0.63
22	23.17	(-)-Spathulenol	77171-55-2	C15H24O	148,829,553	1.39
23	23.308	Longiborneol	465-24-7	C15H26O	48,630,032	0.45
24	23.413	(+,-)-E-Nuciferol	39599-18-3	C15H22O	140,236,444	1.31
25	23.584	Santalol, cis	19903-72-1	C15H24O	170,432,202	1.59
26	23.979	Ar-turmerone	532-65-0	C15H20O	4,640,803,874	43.36
27	24.413	Curlone	87440-60-6	C15H22O	18,943,37,804	17.70
28	24.597	1,5,5,8-Tetramethyl-12-oxabicyclo[9.1.0]dodeca-3,7-diene	19888-34-7	C15H24O	58,206,516	0.54
29	24.82	1-[(1E)-1,3-Dimethyl-1,3-butadienyl]-2,2,6-trimethyl-7-oxabicyclo[4.1.0]heptane	59744-12-6	C15H24O	298,495,706	2.79
30	25.129	6-Isopropenyl-4,8a-dimethyl-4a,5,6,7,8,8a-hexahydro-2(1H)-naphthalenone	86917-79-5	C15H22O	514,001,357	4.80
31	25.583	Dihydrocostunolide	2225-79-8	C15H22O2	79,858,558	0.75
32	25.8	1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl 3-methyl-2-butenoate	91404-82-9	C15H24O2	61,308,066	0.57
33	25.951	Velleral	50656-61-6	C15H20O2	49,479,578	0.46
34	26.155	1-(2-Isopropenyl-4,4-dimethylcyclopentyl)-3-methyl-2-buten-1-one	152481-77-1	C15H24O	61,675,232	0.58
35	27.779	Dehydrosaussurea lactone	28290-35-9	C15H20O2	34,892,012	0.33
36	28.128	Etofylline	519-37-9	C9H12N4O3	12,301,826	0.11
37	28.312	2-Methyl-E,E-3,13-octadecadien-1-ol	112-63-0	C19H36O	21,866,928	0.20
38	28.483	2-Nonadecanone	629-66-3	C19H38O	9,825,017	0.09
39	28.687	Dehydrojuvabione	16060-78-9	C16H24O3	10,397,787	0.10
40	28.825	Linoleic acid	60-33-3	C18H32O2	38,862,215	0.36
41	28.877	trans-Oleic acid	112-79-8	C18H34O2	59,787,209	0.56
42	29.101	Octadecanoic acid	57-11-4	C18H36O2	13,215,100	0.12
43	31.968	Eicosanoic acid	506-30-9	C20H40O2	736,462	0.01
44	34.466	Behenic alcohol	661-19-8	C22H46O	4,420,233	0.04

The volatile components of TVO were identified using GC-IMS. Figure 1E depicts the three-dimensional spectrum of TVO. As shown in Fig. 1F, most volatile organic compounds were distributed in a region with retention times between 500 and 3500 s and drift times of 1–2 min. The spectrum exhibited numerous signal peaks, indicating a substantial abundance of volatile aromatic compounds in the TVO sample. After labeling significant samples, the specific volatile organic compounds in the TVO samples were identified, as shown in Fig. 1G and Supplementary Table 1. A total of 48 previously unidentified and 64 known volatile substances were detected and classified into seven categories (Fig. 1H): 7 acids, 16 alcohols, 13 aldehydes, 12 ketones, 8 esters, 7 alkenes, and 1 furan. The substances with the highest concentrations included 2-methylbenzoic acid monomers and dimers, acetic acid monomers and dimers, eudinol monomers and dimers, ethanol, acetone, butyric acid, and 3-hydroxy-2-butanone monomers. Alcohols (25%) and ketones (18.75%) were the principal volatile components, consistent with the results of the GC-MS analysis.

TVO alleviates OVA-induced AR in vivo

Figure 2A shows an outline of the AR model. To examine the anti-inflammatory capacity of TVO on an OVA-induced mouse model, behavioral changes were assessed. As shown in Fig. 2B and C, nose rubbing and frequency of sneezing in the AR group were substantially greater than in the control (Ctrl) group (P < 0.0001), while TVO-H treatment significantly reduced these behaviors (P < 0.001, vs. AR group).

Fig. 2. Effects of TVO on AR mice.

A Schematic diagram of animal modeling. B Nasal rubbing. C Sneezing. D HE staining and its nasal cavity thickness vector map (scale bar = 5 mm or 50 μm). E Toluidine blue staining and vector plot of mast cell count (scale bar = 5 mm or 50 μm). F Concentrations of IL-6, IFN-γ, IgE, and HIS in serum. G Number of white blood cells, neutrophils, lymphocytes, and eosinophils in blood cells. Ctrl: Control group; AR: Allergic Rhinitis group; TVO L: AR + TVO (2.485 μg/mL); TVO H: AR + TVO (4.97 μg/mL); Fp: AR+Fluticasone propionate (50 μg/mL). ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001 VS AR, n = 6.

As AR symptoms primarily affect the nasal cavity, pathological changes in nasal tissue slices were measured using hematoxylin and eosin (HE) staining. Due to subcutaneous inflammatory cell accumulation, the nasal mucosa thickness was markedly elevated in the AR group (Fig. 2D, P < 0.0001), while treatment with TVO or fluticasone propionate (Fp) reduced this thickness (P < 0.0001). To identify the infiltrating cell types contributing to tissue thickening, mast cells were stained with toluidine blue. Relative to the Ctrl group, the number of infiltrating mast cells exhibited a pronounced increase in the AR group (Fig. 2E, P < 0.0001), while both the TVO and Fp groups showed significantly reduced mast cell infiltration (P < 0.0001).

Serum concentrations of IL-6, interferon (IFN)-γ, IgE, and HIS were evaluated (Fig. 2F), along with whole blood cell counts. The results of the experiment using the mouse model indicated that TVO markedly attenuated the inflammatory levels of these inflammatory mediators, consistent with the in vitro findings. Blood cell analysis (Fig. 2G) showed that TVO treatment resulted in lower numbers of eosinophils, neutrophils, lymphocytes, and white blood cells.

Nasal microbiota dysregulation regulated by TVO in AR mice

The effect of TVO on nasal microbiota imbalance in AR mice was investigated using 16S rRNA gene sequencing. Alpha diversity analysis showed differences among the Ctrl, AR, and TVO groups based on Shannon, Simpson, and Pielou’s evenness indices. The AR group showed a significant decrease in OTU numbers, community richness, and evenness, whereas the TVO group showed an upward trend, though not statistically significant. Goods coverage indicated no major changes compared with the AR group, but microbial diversity in the TVO group trended downward (Fig. 3A).

Fig. 3. Nasal microbiota imbalance regulated by TVO in AR mice.

A Analysis of alpha diversity. B PCoA analysis. C Phylum-level composition. D Genus-level composition. E LEfSe analysis. F Changes in bacterial abundance at phylum level. G Changes in bacterial abundance at genus level. Ctrl: Control group; AR: Allergic Rhinitis group; TVO: AR + TVO (4.97 μg/mL). *P＜0.05, n = 6.

Principal coordinate analysis revealed a clear division between the AR group and both the Ctrl and TVO groups, suggesting a distinct shift in nasal microbiota structure due to AR (Fig. 3B). Taxonomic composition at the genus and phylum levels was analyzed. The top 10 genera included Escherichia, Mycoplasmopsis_B, Ligilactobacillus, Rodentibacter_A, Enterococcus_H, Streptococcus, Muribacter, Rothia, Gemella, and Lactobacillus (Fig. 3C). At the phylum level, the most abundant taxa were Proteobacteria, Firmicutes_D, Actinobacteriota, Firmicutes_A, Bacteroidota, Gemmatimonadota, Acidobacteriota, Methylomirabilota, Desulfobacterota_B, and Chloroflexota (Fig. 3D).

Linear discriminant analysis effect size (LEfSe) analysis identified 26 taxonomic units demonstrating marked between-group discrepancies. In Fig. 3E, Ctrl had a higher abundance of Enterococcaceae, Enterococcus, Alphaproteobacteria, Staphylococcales, Gemellaceae, Gemella, and Marinilabiliaceae. The AR group was dominated by Escherichia, Mycoplasmatales, Metamycoplasmataceae, Mycoplasmopsis, Allobaculum, Pseudomonas, and Alloprevotella. The TVO group showed increased abundance of Dysgonomonadaceae, Saprospiraceae, CAG41, Fermentimonas, UBA1381, Pseudonocardiaceae, Aquabacterium, and Enterenecus. Figure 3F shows changes in abundance across four phyla, while Fig. 3G highlights significant changes in five genera. Rodentibacter was significantly elevated in the AR group, whereas Mycoplasmopsis, Muribacter muris, Bacteroides_H uniformis, and Streptococcus acidominimus were reduced. TVO treatment significantly decreased the abundance of Rodentibacter and increased the abundance of the other four genera.

Influence of TVO on serum metabolites

To further investigate the therapeutic effect of TVO on AR, non-targeted metabolomic analysis of mouse serum was performed using liquid chromatography mass spectrometry (LC-MS). In total, 880 metabolites were detected and classified into 11 categories (Fig. 4A). OPLS-DA score plots revealed a clear separation among the Ctrl, AR, and TVO groups (Fig. 4B), confirming the model’s reliability (Supplementary Table 2). Compared with the Ctrl group, the AR group exhibited 89 elevated and 707 inhibited metabolites. Compared with the AR group, the TVO group exhibited 361 elevated and 161inhibited metabolites (Fig. 4C, D).

Fig. 4. Effect of TVO on serum metabolites.

A Classification of metabolites in positive and negative ion modes. B OPLS-DA scoring chart. C Ctrl VS AR and D AR VS TVO bar chart of differences in metabolites. E Ctrl VS AR and AR VS TVO KEGG pathway. F Statistical chart of differential metabolites of L-kynourine, ceramide, his, sphingosine and prostaglandin D1. Ctrl: Control group; AR: Allergic Rhinitis group; TVO: AR + TVO (4.97 μg/mL). ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001 VS AR, n = 6.

Pathway analysis indicated that TVO significantly modulated metabolic pathways involved in AR, particularly affecting ATP synthase-binding cassette (ABC) transport, nicotine metabolism, and neuroactive ligand-receptor interactions (Fig. 4E). Differential metabolites, such as L-kynurenine, ceramide, HIS, sphingosine, and prostaglandin D₁, were elevated in the AR group but regulated back toward normal levels in the TVO group (Fig. 4F).

TVO alters the proteomics of AR mice

Four-dimensional data-independent acquisition (4D-DIA) non-label quantitative proteomics was utilized to study protein expression changes in the nasal mucosa of mice. The principal component analysis score scatter plot (Fig. 5A) clearly shows distinct separation among the Ctrl, AR, and TVO groups. Notably, PC1 indicates that the TVO group is more similar to the Ctrl group and distinct from the AR group. A total of 10,447 non-overlapping proteins were identified based on one or more unique peptides. Screening for differential proteins with FC > 1.5 or <0.67 and P < 0.05. Compared against the AR group, 711 upregulated and 548 downregulated (Differentially Expressed Proteins) DEPs were identified in the Ctrl group, while 660 upregulated and 636 downregulated DEPs were found in the TVO group (Fig. 5B).

Fig. 5. 4D-DIA proteomics results.

A PCoA analysis. B Difference volcano chart. C Venn diagram. D Heat map. E GO analysis. F KEGG analysis. G Differential Protein Statistics Chart. Ctrl: Control group; AR: Allergic Rhinitis group; TVO: AR + TVO (4.97 μg/mL). ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001 VS AR, n = 6.

The Venn diagram in Fig. 5C displays shared and unique DEPs across groups, with a total of 749 proteins overlapping between comparisons. The top 20 DEPs were visualized via heatmap (Fig. 5D), showing inflammation-related DEPs enhanced in the AR group and diminished in both the Ctrl and TVO groups.

Systematically classify the 749 DEPs from Ctrl/AR ∩ AR/TVO using the GO repository. (Fig. 5E). Fisher’s exact test (P < 0.05) illustrated that inflammation-related DEPs were most enriched. KEGG analysis of the 20 DEPs showed that TVO is closely associated with the MAPK and NF-κB signaling pathways, as shown in Fig. 5F. These findings suggest TVO’s therapeutic potential in AR by targeting key inflammatory signaling pathways.

Among the top 20 cross-DEPs, proteins such as MAPK9, Relb, MAPK14, MAPK13, Nfkb1, and MAPK3 were key components of these pathways. In Fig. 5G, these proteins were overexpressed in the AR group and significantly reduced following TVO treatment.

Validation of key DEPs in mouse nasal mucosa

Key DEPs were verified through quantitative reverse transcription polymerase chain reaction (qRT-PCR) and western blotting. The findings (Fig. 6A–D) demonstrated that OVA markedly elevated the protein expression of p-JNK, p-ERK, p-p38, cyclooxygenase-2 (COX-2), IκBα, and NF-κB, as well as the transcription of JNK, ERK, p38, and NF-κB. In contrast, TVO treatment inhibited their expression. These results confirm that TVO alleviates OVA-induced AR by suppressing MAPK and NF-κB signaling pathways. Immunohistochemical staining (Fig. 6E, F) further confirmed that TVO inhibited the OVA-induced upregulation of JNK and NF-κB expression.

Fig. 6. Verification results of the key differentially expressed proteins (DEPs) in the nasal mucosa tissue of mice.

A Immunoblot bands of ERK, p-ERK, p38, p-p38, JNK, and p-JNK in nasal tissue. B ERK, p-ERK, p38, p-p38, JNK, and p-JNK in nasal tissue were quantified by western blot and the mRNA levels of ERK, JNK, and p38. C Immunoblot bands of COX-2, IκBα, and NF-κB in nasal tissue. D Quantitative analysis of COX-2, IκBα, NF-κB in nasal tissue by western blot. E Representative immunohistochemical images and quantitative analysis of JNK in mouse noses. F Representative immunohistochemical images and quantitative analysis of NF-κB in mouse noses. Ctrl: Control group; AR: Allergic Rhinitis group; TVO L: AR + TVO (2.485 μg/mL); TVO H: AR + TVO (4.97 μg/mL); Fp: AR+Fluticasone propionate (50 μg/mL). ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001 VS AR, n = 3.

TVO relieves allergy symptoms in vitro

Co-culture of TVO with human mast cells (HMC)-1 cells demonstrated that TVO at 0.1, 0.2, and 0.4 μL/mL promoted cell proliferation (Fig. 7A). Exposure to lipopolysaccharide LPS (100 ng/mL) and IFN-γ (40 ng/mL) caused approximately 60% cell damage, making these concentrations suitable for modeling (Fig. 7B). Co-treatment with TVO improved cell proliferation post-damage (Fig. 7C).

Fig. 7. Effect of TVO on HMC-1.

A Effect of TVO on proliferation of HMC-1. B Damage of HMC-1 induced by LPS+IFN-γ. C Effect of LPS+IFN-γ and TVO co-culture on proliferation of HMC-1. D IL-1β, TNF-α, and IL-6 levels in HMC-1 cells. E TNF-α, IL-1β, IL-6 and IL-10 mRNA levels in HMC-1 cells. Ctrl: Control group; LPS+IFN-γ: Model group; TVO1: LPS+IFN-γ+TVO (0.1 μL/mL); TVO2: LPS+IFN-γ+TVO (0.2 μL/mL); TVO3: LPS+IFN-γ+TVO (0.4 μL/mL). ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001VS LPS+IFN-γ, n = 3.

To elucidate the anti-inflammatory effects, ELISA was used to measure the IL-6, TNF-α, and IL-1β levels. As shown in Fig. 7D, LPS and IFN-γ markedly increased pro-inflammatory cytokines, which were subsequently reduced by TVO treatment. Additionally, qRT-PCR results (Fig. 7E) showed that TVO significantly downregulated TNF-α, IL-1β, and IL-6 transcription, while upregulating anti-inflammatory IL-10.

TVO reduces inflammation by inhibiting JNK/MAPK and NF-κB pathways

MAPKs are activated through phosphorylation and nuclear translocation upon LPS stimulation. To examine the impact of TVO on this pathway, qRT-PCR and western blotting were conducted. TVO reduced the LPS- and IFN-γ-induced increases in p-JNK, p-p38, and p-ERK levels in a dose-dependent manner (Fig. 8A, C), and similarly decreased JNK, ERK, and p38 mRNA levels (Fig. 8E). Comparison with SCH772984 confirmed TVO’s effective inhibition of MAPK activation (Fig. 8B, D).

Fig. 8. TVO attenuates LPS and IFN-γ-induced inflammatory responses in HMC-1 cells via the MAPK signaling pathway.

A Immunoblot bands of ERK, p-ERK, p38, p-p38, JNK, and p-JNK in HMC-1 cells. B Immunoblot bands of ERK, p-ERK, p38, p-p38, JNK, and p-JNK in SCH772984-treated HMC-1 cells. C Western blot quantitation of ERK, p-ERK, p38, p-p38, JNK, and p-JNK. D Western blot quantization of ERK, p-ERK, p38, p-p38, and JNK and p-JNK in SCH772984-treated HMC-1 cells. E mRNA levels of ERK, p38, and JNK in HMC-1 cells. Ctrl: Control group; LPS+IFN-γ: Model group; TVO1: LPS+IFN-γ+TVO (0.1 μL/mL); TVO2: LPS+IFN-γ+TVO (0.2 μL/mL); TVO3: LPS+IFN-γ+TVO (0.4 μL/mL); SCH772984: LPS+IFN-γ+TVO (0.4 μL/mL)+SCH772984(10 μM). ^#P < 0.05, ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001VS LPS + IFN-γ, n = 3.

NF-κB signaling was assessed by LPS and IFN-γ induced nuclear translocation of NF-κB and increased COX-2, IκBα, and NF-κB expression and mRNA levels of NF-κB, which were inhibited by TVO treatment (Fig. 9A, B). Immunofluorescence imaging (Fig. 9C) and flow cytometry (Fig. 9D) confirmed reduced NF-κB activation and nuclear localization following TVO treatment, indicating anti-inflammatory activity via the NF-κB pathway.

Fig. 9. TVO attenuates LPS and IFN-γ-induced inflammatory responses in HMC-1 cells through the NF-κB signaling pathway.

A Immunoblot bands of COX-2, IκBα, NF-κB, and NF-κB in cytoplasm and nucleus of HMC-1 cells. B Western blotting of COX-2, IκBα, NF-κB, and NF-κB in cytoplasm and nucleus, and mRNA levels of NF-κB. C Immunofluorescence image of NF-κB in HMC-1 cells. D The expression of NF-κB in HMC-1 cells was quantitatively analyzed by flow cytometry. Ctrl: Control group; LPS+IFN-γ: Model group; TVO1: LPS+IFN-γ+TVO (0.1 μL/mL); TVO2: LPS +IFN-γ+TVO (0.2 μL/mL); TVO3: LPS+IFN-γ+TVO (0.4 μL/mL). ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001 VS Control; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001VS LPS+IFN-γ. n = 3.

Relb knockdown enhances the anti-inflammatory effect of TVO

4D-DIA analysis identified Relb as a key target of TVO in AR treatment. Relb, part of the NF-κB family, forms complexes (e.g., RelB/p50) that modulate inflammation-related gene expression²⁴. TVO significantly inhibited Relb expression in both mouse and cell models (Fig. 10A, B). For clarify Relb’s role, siRNA knockdown was conducted. qRT-PCR and western blot confirmed effective knockdown (Fig. 10C, D). Notably, Relb knockdown decreased nuclear p65 and increased cytoplasmic p65, and TVO inhibited its nuclear translocation (Fig. 10E, F). Additionally, Relb knockdown in combination with TVO suppressed the expression of inflammatory cytokines (Fig. 10G), underscoring the role of Relb in the anti-inflammatory mechanism of TVO in AR.

Fig. 10. Relb knockdown accelerates the attenuation of the anti-inflammatory effect of TVO in HMC-1 cells treated with LPS+IFN-γ.

A, B Immunoblotting and quantitative analysis of Relb in the nasal mucosa of AR mice and in HMC-1 cells treated with LPS and IFN-γ. C, D The expression of Relb mRNA and protein in HMC-1 cells transfected with Si-Relb. E, F Western blot analysis of p65 (cytoplasmic and nuclear) protein levels. G The relative expression of TNF-α, IL-6, and IL-1β in LPS+IFN-γ-treated HMC-1 cells. ^##P < 0.01, ^###P < 0.001, ^####P < 0.0001; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001, n = 3.

Correlation analysis

To explore the correlations among behavioral parameters, inflammatory markers, blood profiles with nasal microbiota, metabolites, and proteins, we performed further correlation analyzes using 5 selected nasal bacteria, 5 differential metabolites, and 7 DEPs, and constructed a correlation network (Fig. 11A). Behavior, inflammatory cytokines, and hematological indices were mostly positively correlated with other parameters. Rodentibacter showed a strong positive correlation with inflammation, while Bacteroides_H uniformis, Muribacter muris, S. acidominimus, and Mycoplasmopsis were negatively correlated with inflammatory markers. Differential proteins and metabolites were positively correlated with most inflammatory indicators.

Fig. 11. Multi-omics combined analysis of turmeric volatile oil alleviating allergic rhinitis.

A Association diagram. B Mechanism pathway diagram. Created in BioRender.

Discussion

Turmeric is commonly used to treat inflammation-related diseases, particularly in southeastern coastal and southwestern regions¹⁸. Inflammation is widely acknowledged as a pivotal pathophysiologic determinant driving AR development²⁵. The main components of turmeric are TVO. While curcumin has been extensively documented for potent inflammation-suppressive activity^26–28, the mechanism of action of TVO in treating inflammation-related diseases remains unclear. Thus, we sought to explore the anti-inflammatory capacity of TVO for AR, with the goal of identifying new therapeutic strategies for AR management.

In the present study, turmeric was sourced from Guangxi, China. Supercritical CO₂ extraction yielded an almost curcumin-free volatile oil rich in 43.36% ar-turmerone, consistent with the results of previous reports²⁹. The composition of TVO was analyzed using an integration of GC-MS and GC-IMS, revealing a high abundance of alcohols and esters—findings that align with earlier research³⁰. Studies have demonstrated that certain naturally derived alcohols (e.g., cedrol) demonstrate pronounced anti-inflammatory efficacy across both cellular and animal experimental systems³¹. Moreover, esters have been confirmed to possess anti-inflammatory properties³². Thus, it is inferred that the alcohols and esters in TVO act synergistically to exert anti-inflammatory effects.

TVO treatment rapidly lowered serum total IgE, blunting allergen-driven systemic IgE synthesis^2,33 and reducing FcεRI loading on mast cells and basophils, while peripheral eosinophil counts fell. HE and toluidine-blue staining of nasal mucosa showed diminished eosinophil infiltration and fewer mast cells, establishing an “IgE decrease–mast-cell load decrease–eosinophil chemotaxis decrease” negative chain consistent with IgE-driven pathology³⁴. Concomitant decreases in HIS and IL-6 indicated suppressed mast-cell degranulation. IFN-γ, which correlated positively with inflammation score in the OVA-AR model (r = 0.72)³³, served as an inflammatory amplifier under chronic conditions; TVO downregulated IFN-γ, cooperatively attenuating mast-cell and eosinophil infiltration. High-dose TVO significantly reduced serum IgE, HIS, IL-6, peripheral eosinophil percentage, and IFN-γ (P < 0.05) and alleviated nasal mucosal inflammatory-cell infiltration and mast-cell degranulation. At low doses, while an improvement was observed, this did not reach the level of statistical significance, suggesting that TVO’s anti-inflammatory effect is dose-dependent. Although it is weaker than the positive drug Fp overall, as a food active ingredient, TVO can moderately regulate immunity through multiple targets, is safe and non-toxic²², and is suitable for long-term conditioning and prevention. It complements the drug and provides a basis for functional food-assisted intervention in AR. In the future, the extraction process and dosage form will be optimized to improve bioavailability.

In humans, >99% of the microbiota consists of Firmicutes, Bacteroidetes, Proteobacteria, Pseudomonadota, and Actinobacteria³⁵. Recent studies have indicated that AR disrupts the structure and taxonomic richness of the nasal microbiota. Characteristic changes in AR mice include a notable boost in Proteobacteria and a decrease in Firmicutes and Actinobacteria³⁶. Proteobacteria, a major phylum containing pathogenic bacteria, release endotoxins that may activate the immune system via LPS, triggering pro-inflammatory responses in susceptible hosts³⁷. Consequently, the growth of Proteobacteria is considered to correlate with AR-related inflammation, particularly under a disrupted immune system. Our study showed that the abundance of Rodentibacter increased in AR mice, and Alloprevotella (from the Bacteroidetes phylum) may further amplify inflammation. TVO treatment increased the relative abundance of Enterenecus and S. acidominimus, both belonging to the Firmicutes phylum, which play crucial parts in maintaining microbiota balance and reducing inflammatory levels.

The impact of serum metabolites on AR may involve multiple pathways. In the present study, LC-MS analysis of serum samples revealed 519 differential metabolites between the AR and TVO groups. Pathway enrichment analysis showed significant involvement of ABC transporters, nicotine metabolism, and neuroactive ligand-receptor interaction pathways. Previous studies reported that AR affects serum metabolite profiles, involving pathways related to fatty acid and amino acid metabolism³⁸. ABC transporters regulate these pathways. Nicotine has been shown to influence allergic disease development by altering pro-inflammatory metabolites and promoting oxidative stress and immune dysregulation. It activates nAChR receptors, prompting mast cells to release HIS and leukotrienes (LTs)³⁹. Neuroactive ligand-receptor interactions contribute to allergic responses through the regulation of immune metabolites (e.g., short-chain fatty acids, LTs), neurotransmitters (e.g., HIS, substance P), and energy metabolites (e.g., ATP)⁴⁰. In our study, TVO treatment alleviated the inflammatory response and significantly reduced IL-6, TNF-α, and HIS levels in mouse serum.

Subsequently, we applied proteomics techniques. GO and KEGG enrichment analyzes of differential proteins revealed significant enrichment in the MAPK and NF-κB signaling pathways—both classical inflammatory pathways. Camacho-Barquero et al. found that curcumin inhibited p38MAPK activity in colitis without affecting JNK activation⁴¹. Recently, Meiyanto et al. reported that curcumin and its derivatives inhibit p65 expression, indicating suppression of NF-κB activity. Curcumin’s affinity for targets, such as HER2, EGFR, IKK, and ER, contributes to this effect⁴². Relb, a key NF-κB family member, activates signaling cascades upon recognition of pathogens and regulates inflammatory gene expression⁴³. Based on 4D-DIA results, our study identified Relb as a critical TVO target. Knockdown experiments showed that Relb suppression enhanced the anti-inflammatory effects of TVO in LPS and IFN-γ-treated human mast cells. This finding not only reveals a new mechanism of TVO action but also highlights Relb as a potential therapeutic target in AR. These results are consistent with prior research identifying Relb as a central regulator in NF-κB-mediated inflammation⁴⁴. Our findings deepen our current understanding by elucidating the therapeutic mechanism of TVO through Relb-mediated pathways.

Correlation analysis further indicated that Rodentibacter in the microbiome exhibited strong positive correlations with other inflammatory indicators. In the metabolome, L-kynurenine and prostaglandin D1 were highly correlated with neutrophil count, white blood cell count, HIS levels, and Nfkb2. Overall, significant changes in microbial and molecular profiles occurred in OVA-induced AR mice, reflecting TVO’s in vivo activity. Furthermore, to our knowledge, this is the first study to jointly analyze the microbiome, proteome, and metabolome, demonstrating that TVO can simultaneously reverse the imbalance of the microbial community and metabolic disorders. The analysis results are accurate and have broadened our understanding for future related research.

The present study provides compelling evidence that TVO effectively mitigates AR-induced inflammation by inhibiting the MAPK and NF-κB signaling pathways (Fig. 11B). Potential TVO targets were predicted using 4D-DIA, identifying Relb as a key therapeutic target, which was validated in vitro. The regulatory effects of TVO on microbial diversity and metabolic profiles were further elucidated using 16S rRNA and metabolomics approaches. In LPS and IFN-γ-induced HMC-1 cells and OVA-induced AR mouse models, TVO markedly attenuated levels of IgE, IL-6, TNF-α, HIS, and the expression of JNK, p-ERK, p-P38, COX-2, IκBα, and NF-κB via the MAPK and NF-κB pathways. These findings indicate that TVO has strong viability as a remedial agent for AR. Future research should aim to provide additional experimental evidence clarifying the interactions among the signaling pathways modulated by TVO.

In conclusion, our research results indicate that TVO exerts anti-inflammatory effects to alleviate Relb-mediated immune inflammation. Mechanism analysis indicates that TVO exerts its anti-inflammatory effect by inhibiting the MAPK and NF-κB type pathways. It is worth noting that Relb has been identified as a therapeutic target for TVO. This discovery enhances our insight into the molecular pathways that underlie the clinical efficacy of TVO and supports the development of TVO as a promising therapeutic strategy for AR.

Methods

Materials

Turmeric (Curcuma longa L.) was purchased from Hongjian Pharmacy (Lot No. 20220225) and authenticated by Associate Professor Zhe Wang of the Changchun University of Traditional Chinese Medicine. OVA and LPS were purchased from Sigma-Aldrich (St. Louis, MO, USA), and aluminum hydroxide from Thermo Scientific (Waltham, MA, USA). Fluticasone propionate nasal spray was procured from the Flixonase (HJ20140117). Pancreas was obtained from Shandong Jingyu Biotechnology Co., Ltd. Fetal bovine serum and culture medium were purchased from Gibco. Nitric oxide (NO), IFN-γ, IL-6, IL-1β, HIS, and TNF-α detection kits were sourced from Sinobestbio (Shanghai, China). RIPA Lysis Buffer and Cell Counting Kit-8 (CCK-8) were obtained from APExBIO Technology LLC (Houston, TX, USA). The bicinchoninic acid (BCA) protein concentration kit, TRIzol lysate, anti-fluorescence quencher, nuclear and cytoplasmic protein extraction kit, Triton, and DAPI were purchased from Beyotime (Jiangsu, China). Rabbit anti-JNK, p-JNK, ERK, p-ERK, p38, p-p38, NF-κB, COX-2, GAPDH, and RelB antibodies were purchased from Bioworld Technology (Lin, Minnesota, MN, USA). Murine anti-IκBα antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). The SCH772984 inhibitor and Poly Fast were purchased from MCE (Shanghai, China). Paraformaldehyde was purchased from Yuanyuan Leaf (Shanghai, China). Si-Relb was synthesized in the Public Protein/Plasmid Library.

Extraction process of turmeric volatile oil

A total of 1.7 kg of turmeric medicinal material was crushed (80–100 mesh). Supercritical CO₂ extraction was used to collect essential oil (in liquid form at room temperature), and the extraction yield of TVO was taken as the evaluation index.

GC-MS

Chromatographic conditions included an Agilent HP-5 capillary column, with helium carrier gas at 1 mL/min, and a 1 μL aliquot was injected. The temperature program was 60 °C, ramped to 280 °C at 10 °C/min, and maintained at 280 °C for 27 min. Mass spectrometry settings included an electrospray ionization source at 230 °C, transfer line temperature at 150 °C, a scan range of m/z 20–600, and data acquisition after a 3-min delay⁴⁵.

GC-IMS

GC-IMS (G.A.S. FlavourSpec®, Dortmund, Germany) equipped with a Restek MXT-WAX capillary column was used to analyze samples. Precisely 0.0525 g of TVO was placed in 20 mL headspace vials, in triplicate. Six C₄–C₉ ketones were identified, and retention time–retention index calibration curves were established. Retention indices were calculated and matched using the NIST 2020 GC retention index database and IMS migration time database within VOCal software⁴⁶.

Mouse OVA-induced AR model

Sixty female Balb/c mice (6 weeks old, 21 ± 2 g) were obtained from Liaoning Changsheng Animal Co., Ltd (License No. 01021698199093137) and housed under SPF conditions. All procedures were approved by the Animal Research Ethics Committee at Changchun University of Traditional Chinese Medicine (License No.: 20220115; Certificate No.: SYXK, Ji, 2022-0014).

Based on the results of the preliminary experiment (Supplementary Fig. 2), 2.485 μg/mL and 4.97 μg/mL were selected as the low-dose and high-dose. Mice were randomly allocated into the following groups (n = 12): control (phosphate-buffered saline [PBS] only), AR model (OVA with nasal PBS), TVO low-dose (2.485 μg/mL), TVO high-dose (4.97 μg/mL), and positive drug (fluticasone propionate, 50 μg/mL). On days 0, 2, 4, 6, 8, 10, and 12, all except for the control group were sensitized with 200 µL PBS containing 20 µg OVA and 2 mg aluminum hydroxide via intrathecal injection. On days 15–35, mice received intranasal administration of 100 µg OVA in 40 µL PBS (20 µL per nostril).

On days 22–35, the treatment and positive drug groups received their respective agents 30 min post OVA administration, while control and AR groups received PBS. On day 35, mice were placed in observation cages, and 6 randomly selected individuals were observed and recorded for the number of sneezes and nose rubs within 10 min. On day 36, mice were deeply anesthetized with an intraperitoneal injection of 100 mg/kg pentobarbital sodium (verify unconsciousness by loss of pedal and corneal reflexes) and then euthanized by cervical dislocation while still fully anesthetized; death was confirmed by absence of heartbeat and respiration⁴⁷.

Serum parameter analysis

On day 36, 100 µL of blood was drawn from the ocular vein into EDTA tubes. Samples were subjected to routine hematological testing (Sysmex XT-2000L, Tokyo, Japan), then stored (4 °C, 30 min), centrifuged (3000 × g, 15 min) and serum was collected. ELISA kits were used to quantify HIS, IFN-γ, IL-6, and IgE levels, following manufacturer protocols.

Histopathologic analysis of nasal tissue

Nasal tissues were fixed in paraformaldehyde for 24 h, decalcified in EDTA for 2 weeks, and paraffin-embedded. HE and toluidine blue staining were performed. Tissue morphology and mast cell infiltration were observed under a Nikon light microscope (Tokyo, Japan)⁴⁸.

16S rRNA

Non-invasive alveolar lavage fluid was collected using 1.5 mL PBS via bronchial incision, frozen at −80 °C, and subjected to 16S rRNA sequencing. DNA was extracted and verified by 0.8% agarose gel electrophoresis and Nanodrop. The V3–V4 region was amplified using primers 338F/806R. Microbiome analysis was performed via QIIME2 (v2019.4) by Shanghai Peisenol Biotechnology Co., Ltd (Shanghai, China).

Metabolomics

Mouse serum metabolic profiles were analyzed by LC-MS. Samples were treated with pre-cooled methanol:acetonitrile (1:1, v/v), vortexed, frozen, and clarified. The supernatant was concentrated, re-dissolved with methanol containing an internal standard (2-chlorophenylalanine), vortexed, and centrifuged again. After filtration, samples were analyzed, with pooled QC samples ensuring instrument stability. Analysis was conducted by Shanghai Parsonage Biotechnology Co. (Shanghai, China).

4D-DIA

Nasal mucosa was isolated and stored at −80 °C for 4D-DIA and western blot analysis. Protein concentration was quantified via the BCA method. After desalting, samples underwent LC-MS/MS. A protein spectral library was first generated using DDA, followed by DIA data collection. Quantification and statistical analysis (P < 0.05) were conducted using the spectral database. Analysis was performed by Shanghai Peisenol Biotechnology Co., Ltd.

Immunohistochemistry

Paraffin-embedded nasal tissue sections (fixed in 4% paraformaldehyde) were used for immunohistochemistry. After antigen retrieval and blocking, tissues were exposed to primary antibodies (1:200) at 4 °C overnight and secondary antibodies (1:1000) at 25 °C for 50 min. Chromogenic visualization was achieved with diaminobenzidine, followed by nuclear counterstaining with hematoxylin. Slides were dehydrated, sealed, and imaged with a Nikon microscope⁴⁹. Image-Pro Plus (v6.0) was used for analysis.

HMC-1 cell culture and treatment

HMC-1 derived from bone marrow hematopoietic stem cell were cultured in media containing 10% fetal bovine serum and 1% streptomycin/penicillin. When cell density reached >90%, cells were seeded into 96-well plates (1 × 10⁵ cells/well) for 24 h. The medium was replaced with TVO (0.1, 0.2, 0.4 μL/mL) plus LPS (100 ng/mL) and IFN-γ (40 ng/mL) for 1 day. Cell viability was assessed using CCK-8 at 490 nm after 30 min conditioning under dark conditions.

HMC-1 inflammatory factor detection

After TVO treatment, supernatants were collected, and NO, IL-6, and TNF-α levels were measured using ELISA kits per manufacturer instructions.

qRT-PCR

Extracting total RNA with TRIzol (Takara, Japan), converted to cDNA with PrimeScript RT kit, and analyzed by qRT-PCR using SYBR Green PCR Master Mix (Tiangen Biotech, China) on a Bio-Rad CFX96™ system. The 2^^-ΔΔCq method was used for GAPDH normalization. Primer sequences are listed in Table 2.

Table 2.

Primer sequences of genes used in this study

Gene		Sequence (from5’to3’)
		Sense	Antisense
HMC-1	GAPDH	ACCCAGAAGACTGTGGATGG	TCAGCTCAGGGATGACCTTG
	ERK	GGCGTGATCCTGGACTATGA	CTGTGGCGTTGTACTCACTG
	JNK	CCAGTCAGGCAAGGGATTTG	AGGGATCTTTGGTGGTGGAG
	P38	ACCACGATCCTGATGATGAA	AAGGAGGTCCCTGCTTTCAA
	NF-κB	AGCAAATAGACGAGCTCCGA	TCGGTAAAGCTGAGTTTGCG
	TNF-α	AGCACTGAAAGCATGATCCG	CCGATCACTCCAAAGTGCAG
	IL-10	GCCAAGCCTTGTCTGAGATG	GCATTCTTCACCTGCTCCAC
	IL-6	AGACAGCCACTCACCTCTTC	AGTGCCTCTTTGCTGCTTTC
	IL-1β	GCTGGAGAAAGTGGACAAGC	AGCTCGTTGTTCTTCTGGGA
	Relb	TGCATGCTTCGGTCTGGG	ACTCGTCGATGATCTCTGTGC
Mouse	ERK	TAAGGTGCCATGGAACAGGT	CAATGGGCTCATCACTTGGG
	JNK	CGCTACTACAGAGCACCAGA	GCACCCAACTGACCAAATGT
	P38	CTCTTGCGCATGCCTACTTT	GGTCAGGCTCTTCCACTCAT
	NF-κB	CAGTAGAAACAGACCGAGGAGT	CAAGAACGAAGCCATCAGCA

Western blot

HMC-1 cells were lysed in RIPA buffer, and nuclear/cytoplasmic proteins were separated using a dedicated extraction kit. Protein concentration was determined via BCA assay. Proteins were resolved by 10% SDS-PAGE, transferred to NC membranes, and probed with primary and secondary antibodies (1:1000). Bands were visualized using ECL and analyzed with ImageJ (US NIH, Bethesda, MD, USA).

Immunofluorescence

HMC-1 cells were fixed in 4% paraformaldehyde, blocked with 3% BSA (1 h, 25 °C), and incubated with primary antibodies (1:200, 4 °C, 12 h), then secondary antibodies (1:1000, 1 h, 25 °C). Nuclei were stained with DAPI (5 min), and images captured via Nikon fluorescence microscopy.

Flow cytometry

HMC-1 cells were suspended in PBS, immersed in 4% paraformaldehyde (25 min), and permeabilized by 0.1% Triton X-100 (20 min). Cells were incubated with primary antibodies (25 °C, 25 min), washed with PBST, followed by secondary antibody incubation (25 °C, 20 min). After further PBST washes, cells were blocked with 2% fetal bovine serum (200 min) before analysis.

Data analysis

Results are presented as mean ± SEM. Differences among groups were assessed via one-way analysis of variance followed by Dunnett’s post-hoc test (GraphPad Prism 9, Boston, MA, USA), with P < 0.05 taken as the threshold for statistical significance.

Ethics statement

This animal study was reviewed and approved by the Institutional Animal Care and Use Committee of Changchun University of Traditional Chinese Medicine, China (Approval No. 20220115).

Author contributions

L.S.: Writing – original draft, Visualization, Validation, Methodology. S.B.: Writing – original draft, Visualization, Validation, Methodology. X.W.: Visualization, Validation, Software, Formal analysis, Data curation, Funding acquisition. Y.X.: Writing – review & editing, Supervision, Formal analysis, Methodology, Conceptualization. M.W.: Visualization, Formal analysis, Data curation, Funding acquisition, Validation. D.Z.: Writing – review & editing, Visualization, Validation. X.B.: Writing – review & editing, Methodology, Funding acquisition, Conceptualization. W.Z.: Writing – review & editing, Methodology, Funding acquisition, Conceptualization.

Data availability

Study datasets are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41538-026-00937-x.