Integrated m6A methylome and transcriptome profiling of mRNAs and lncRNAs in nasal mucosal epithelial cells of allergic rhinitis patients undergoing allergen-specific immunotherapy

Abstract

Background

Allergic rhinitis (AR) affects 10–25% of the global population, with nasal epithelial cell (NEC) dysfunction acting as a central driver of its pathogenesis. Allergen-specific immunotherapy (AIT) is the sole disease-modifying treatment for AR; however, its optimization is hindered by the lack of validated predictive biomarkers. N6-methyladenosine (m6A) is a pivotal epitranscriptomic regulator of immune and epithelial homeostasis. This study aims to delineate the m6A modification landscape in NECs during AIT to identify candidate genes associated with clinical efficacy and elucidate the underlying molecular mechanisms.

Methods

We performed methylated RNA immunoprecipitation sequencing (MeRIP-Seq) and RNA sequencing (RNA-Seq) on NECs from house dust mite (HDM)-sensitized AR patients who completed a 3-year AIT course and severity-matched non-AIT controls. Bioinformatics analyses, including Gene Set Enrichment Analysis (GSEA), were conducted to explore biological functions and signaling pathways. Key differentially methylated and expressed mRNAs and lncRNAs were validated using qRT-PCR and MeRIP-qPCR in an independent cohort (n = 12). Clinical relevance was assessed via correlation with the Total Nasal Symptom Score (TNSS).

Results

MeRIP-Seq revealed distinct m6A topographies, identifying 1,455 hypermethylated and 4,636 hypomethylated mRNAs, alongside 267 hypermethylated and 799 hypomethylated lncRNAs in the AIT group. Integrated multi-omics analysis and GSEA demonstrated that AIT significantly downregulates pro-inflammatory cascades (e.g., PI3K-Akt, MAPK, and Ras signaling) while upregulating pathways related to cell adhesion and apoptotic regulation. We identified and validated four key mRNAs (KANK2, DBI, IL18, and TNFRSF10A) exhibiting inverse correlations between m6A modification and gene expression, which also correlated with the patients’ TNSS. Furthermore, 28.7% of differentially m6A-methylated lncRNAs were chromatin-associated. Notably, within this subset, the m6A methylation level of LUCAT1 was positively correlated with the patients’ TNSS, suggesting its potential role in epigenetic regulation during AIT.

Conclusions

This study delineates the m6A epitranscriptomic landscape in NECs following AIT, revealing its critical role in orchestrating epithelial barrier restoration and immune microenvironment homeostasis. The identified candidate mRNAs (KANK2, DBI, IL18, TNFRSF10A) and chromatin-associated lncRNAs (e.g., LUCAT1) offer novel mechanistic insights into epitranscriptomic remodeling and serve as promising molecular candidates for future therapeutic or diagnostic development in precision AR management.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13148-026-02128-7.

Introduction

Allergic rhinitis (AR) affects 10–25% of the global population, imposing substantial burdens on patients’ quality of life and healthcare systems worldwide. Pathologically, AR is characterized by type 2 inflammation in the nasal mucosa [1]. Human nasal epithelial cells (NECs) serve as a critical interface between environmental allergens such as house dust mites and the host immune system in this pathological process [2]. Dysfunction of NECs, characterized by impaired barrier integrity [3] and aberrant secretion of pro-inflammatory mediators such as thymic stromal lymphopoietin (TSLP) and interleukin-25 (IL-25), facilitates aeroallergen penetration and triggers downstream type 2 immune responses, which are central to AR pathogenesis [4]. Despite house dust mites being the predominant allergen driving respiratory allergies [5], the molecular mechanisms by which environmental stimuli induce NEC dysfunction and subsequent allergic sensitization remain incompletely elucidated.

Current management strategies for AR are predominantly focused on symptomatic control with oral or intranasal antihistamines, leukotriene receptor antagonists, and intranasal corticosteroids [6]. For moderate-to-severe AR refractory to pharmacotherapy, allergen-specific immunotherapy (AIT) represents the sole disease-modifying intervention. Recent clinical guidelines recommend early initiation of AIT when clinically indicated, as it confers substantial long-term advantages, including attenuated symptom severity upon allergen exposure, diminished dependency on symptomatic medications, and sustained efficacy even following treatment cessation [7]. These therapeutic effects are partially ascribed to AIT-mediated restoration of NEC barrier integrity and normalization of epithelial cytokine secretion [8]. However, optimization of AIT is hampered by its variable response rate (60–70%) and the paucity of validated predictive biomarkers [9], which are clinically indispensable for patient stratification, treatment monitoring, and risk assessment of adverse events [10]. Notably, NECs emerge as a promising source for biomarker discovery due to their direct exposure to allergens and non-invasive sampling feasibility. Nevertheless, their epitranscriptomic profiles during AIT remain largely uncharacterized.

N6-methyladenosine (m6A), one of the most prevalent internal modifications in eukaryotic mRNAs, functions as a dynamic epitranscriptomic regulator that modulates gene expression by governing mRNA stability, splicing, and translational efficiency [11]. Beyond these post-transcriptional effects, recent studies have revealed a reciprocal crosstalk between m6A and epigenetic machinery [12]. m6A modifications on nuclear RNAs, particularly lncRNAs, can influence chromatin accessibility by recruiting histone modifiers, thereby altering local chromatin accessibility and transcription states directly at the DNA level [13]. Through these multifaceted regulatory mechanisms, m6A has emerged as a pivotal player in immune regulation and epithelial homeostasis [14]. In the context of AR, for instance, NECs exhibit dysregulation of the m6A regulatory machinery, including upregulation of the methyltransferase methyltransferase-like 3 (METTL3), which directly enhances TSLP mRNA stability, thereby amplifying mucosal inflammation [15]. Despite these findings, the role of m6A-mediated regulation in NECs during AIT remains poorly defined, particularly regarding its contribution to treatment response via modulation of epithelial barrier function or immune signaling pathways.

To address these knowledge gaps, we integrated methylated RNA immunoprecipitation sequencing (MeRIP-Seq) and RNA sequencing (RNA-Seq) to characterize the transcriptomic and epitranscriptomic landscapes of NECs derived from AR patients who completed a 3-year AIT course and severity-matched non-AIT controls. This study had three primary objectives: first, to delineate the genome-wide m6A modification landscapes associated with AIT response; second, to identify m6A-regulated genes linked to AIT efficacy; and third, to develop and validate a panel of m6A-related biomarkers for predicting AIT efficacy. By elucidating m6A-mediated epitranscriptomic regulation in NECs, this study aims to provide actionable biomarkers to optimize AIT strategies for AR patients.

Results

Transcriptome-wide m6A methylation profiles in NECs of AIT and non-AIT AR patients

To investigate the molecular mechanisms underlying AIT efficacy, we enrolled a cohort of HDM-sensitized AR patients stratified by treatment history. The study group consisted of individuals who had completed a full 3-year course of specific AIT (n = 3), while the control group comprised treatment-naive patients (n = 3). All participants were recruited following rigorous screening based on the 2016 Allergic Rhinitis and its Impact on Asthma (ARIA) guidelines [16]. The clinical characteristics of the enrolled patients are summarized in Table 1.

Table 1.

Clinicopathological characteristics of 6 patients with AR

	AIT			Non-AIT			p value
	1	2	3	4	5	6
Gender	Male	Female	Female	Male	Female	Male	0.52
Age	40	43	24	36	29	31	0.59
Total IgE (IU/ml)	168	93	117	329	252	247	0.012
TNSS	5	5	7	12	10	9	0.014
VAS	22	22	11	51	25	35	0.056

To characterize m6A methylation patterns in NECs from AR patients receiving AIT or without AIT (non-AIT), MeRIP-Seq was performed (Fig. 1). First, we identified m6A methylation peaks in both groups and observed a lower number of methylation peaks in the AIT group compared to the non-AIT group. Specifically, for mRNAs, the AIT group exhibited 32,700 m6A peaks mapping to 12,288 mRNAs, whereas the non-AIT group showed 36,387 peaks mapping to 12,801 mRNAs. Regarding non-coding RNAs (ncRNAs), 8,141 peaks mapping to 3,874 ncRNAs were identified in the AIT group, compared to 7,012 peaks mapping to 4,284 ncRNAs in the non-AIT group (Fig. 2A, B).

Fig. 1.

Schematic illustration of the overall workflow of this study. a Recruitment of AR patients and performance of nasal mucosa provocation tests (n = 6). b Collection of NECs and extraction of RNA samples. c Performance of MeRIP-seq, RNA-seq, and bioinformatics analysis

Fig. 2.

General features of m6A methylation in AR patients receiving AIT or non-AIT treatment. A, B Venn diagrams showing overlapping m6A peaks of mRNAs (A) and ncRNAs (B) between the non-AIT and AIT groups. C Density curves showing the distribution of m6A peaks of mRNA across mRNA transcripts (5’UTR, CDS, and 3’UTR). D Proportions of m6A peaks located in distinct genomic regions of mRNAs. E Proportion of different types of ncRNAs among differentially methylated ncRNAs (F) Motifs enriched in all identified m6A peaks from AIT and non-AIT AR patients

Next, we systematically analyzed transcriptome-wide m6A distribution by comparing m6A-immunoprecipitated (m6A-IP) samples with their corresponding input (non-IP) controls. m6A-IP reads were highly enriched in the 5’ untranslated region (5’UTR), coding sequence (CDS), and 3’ untranslated region (3’UTR) (Fig. 2C), with prominent enrichment near the 3’UTR start site. To precisely map m6A peaks of mRNA, each peak was assigned to one of five non-overlapping mRNA regions: 5’UTR, exon, intergenic region, intron, and 3’UTR. In the non-AIT group, 49.94% of m6A sites resided in the 3’UTR, followed by exons (31.94%). Notably, 5’UTR sites (12.00%) exceeded intronic sites (5.92%). Compared with the non-AIT group, the AIT group exhibited decreased m6A occupancy in the 3’UTR (48.41%) and introns (5.36%), accompanied by increased occupancy in exons (33.63%) and the 5’UTR (12.39%) (Fig. 2D). Furthermore, composition analysis of the differentially m6A-methylated ncRNAs revealed that the vast majority (95.7%) were identified as lncRNAs (Fig. 2E).

Since m6A methylation depends on sequence-specific binding of m6A regulatory proteins, we performed motif enrichment analysis on the identified m6A peaks. Motif logo analysis identified RRACH (R = A/G; H = A/C/U) and ACCDR (D = A/G/U; R = A/G) as the top enriched motifs for m6A-binding proteins (Fig. 2F).

Global features of m6A methylation in NECs from AIT and non-AIT groups

Based on the criteria of |log₂(fold change (FC))| ≥ 1 and Benjamini-Hochberg (BH) adjusted P < 0.05, we identified 1,455 hypermethylated and 4,636 hypomethylated mRNAs in the AIT group compared to the non-AIT group (Fig. 3A). Concurrently, analysis of lncRNAs revealed 267 hypermethylated and 799 hypomethylated transcripts (Fig. 3B). The top 20 differentially methylated mRNAs and the top 10 differentially methylated lncRNAs are listed in Tables 2 and 3, respectively. Furthermore, genome-wide analysis showed that differentially methylated sites were distributed across all chromosomes. Specifically, differentially methylated mRNAs were predominantly enriched on chromosomes 1, 19, 2, and 3 (Fig. 3C), whereas differentially methylated lncRNAs were mainly clustered on chromosomes 1 and 2 (Fig. 3D). Notably, hypomethylated sites outnumbered hypermethylated sites across all chromosomes in both mRNA and lncRNA populations (Fig. 3E, F).

Fig. 3.

Global profiling of m6A methylation patterns in mRNAs and lncRNAs. A, B Volcano plots showing differentially methylated m6A peaks in mRNAs (A) or lncRNAs (B) between the AIT and non-AIT groups. C, D Circos plots showing the global distribution of m6A peaks in mRNAs (C) or lncRNAs (D) across the human genome in the AIT and non-AIT groups. E, F Chromosomal distribution of hypermethylated and hypomethylated m6A-modified peaks in mRNAs (E) and lncRNAs (F) within the AIT group

Table 2.

Top 20 differentially m6A-modified mRNAs in AR patients after AIT compared with non-AIT controls based on adjusted P value

Gene	chr	Chrom start	Chrom end	Annotation	Modification log2FC	Adjusted p-value
MUC4	chr3	195,790,051	195,811,752	exonic	− 1.08	< 0.001
EGR1	chr5	138,465,478	138,467,616	UTR5	− 1.46	< 0.001
EGR1	chr5	138,467,641	138,468,566	UTR3	− 1.02	< 0.001
CLDN4	chr7	73,828,666	73,832,691	UTR3	− 1.44	< 0.001
MUC2	chr11	1,092,865	1,095,116	exonic	− 1.80	< 0.001
MUC2	chr11	1,095,416	1,100,797	exonic	− 1.91	< 0.001
MUC5AC	chr11	1,184,773	1,185,098	exonic	− 1.37	< 0.001
MUC5AC	chr11	1,185,423	1,185,723	exonic	− 1.36	< 0.001
MUC5AC	chr11	1,186,048	1,186,348	exonic	− 1.60	< 0.001
MUC5AC	chr11	1,186,973	1,187,098	exonic	− 1.34	< 0.001
MUC5AC	chr11	1,188,948	1,189,423	exonic	− 1.10	< 0.001
MUC5AC	chr11	1,191,073	1,192,173	exonic	− 1.03	< 0.001
CTNND1	chr11	57,816,379	57,816,904	UTR3	− 1.68	< 0.001
FOS	chr14	75,281,119	75,282,094	UTR3	− 1.08	< 0.001
ZFP36	chr19	39,407,786	39,408,986	UTR3	− 2.10	< 0.001
ZFP36L1	chr14	68,788,984	68,789,559	UTR3	− 1.18	< 0.001
MUC16	chr19	8,955,468	8,955,843	exonic	− 1.48	< 0.001
NR4A1	chr12	52,054,613	52,055,138	exonic	− 1.66	< 0.001
KLF4	chr9	107,484,851	107,487,777	UTR3	− 1.04	< 0.001
JUNB	chr19	12,792,060	12,793,310	UTR3	− 1.48	< 0.001

Table 3.

Top 10 differentially m6A-modified lncRNAs in AR patients after AIT compared with non-AIT controls based on adjusted P value

Gene	chr	Chrom start	Chrom end	Annotation	Modification log2FC	Adjusted p-value
TSIX	X	73,825,179	73,827,329	ncRNA_exonic	− 2.22	< 0.001
XIST	X	73,825,373	73,827,173	ncRNA_exonic	− 1.89	< 0.001
XIST	X	73,849,654	73,852,723	ncRNA_exonic	− 1.18	< 0.001
XIST	X	73,845,579	73,845,979	ncRNA_exonic	− 2.73	< 0.001
AC005062	chr7	19,960,936	20,007,109	ncRNA_exonic	3.75	< 0.001
AC011446	chr19	13,153,070	13,154,193	ncRNA_exonic	− 1.09	< 0.001
LINC00632	X	140,783,564	140,784,314	ncRNA_exonic	− 1.64	< 0.001
AC010300	chr19	23,323,967	23,325,217	ncRNA_exonic	1.49	< 0.001
AC025164	chr12	92,145,622	92,146,047	ncRNA_exonic	− 1.30	< 0.001
XIST	X	73,846,429	73,846,704	ncRNA_exonic	− 2.58	< 0.001

Functional characterization of m6A-methylated mRNA landscapes in NECs following AIT

To elucidate the biological implications of m6A-modified mRNAs, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses. GO analysis highlighted that hypermethylated mRNAs were enriched in Cellular Component (CC) terms such as the membrane and mitochondrial matrix. In the Biological Process (BP) category, these genes were primarily involved in the T cell receptor signaling pathway and positive regulation of microtubule nucleation, while Molecular Function (MF) annotations were dominated by small GTPase binding and phospholipase binding (Fig. 4A). Conversely, hypomethylated mRNAs showed a distinct functional profile, with enrichment in BP terms related to the inflammatory response, cell migration, and regulation of cell adhesion. Correspondingly, their MF categories were characterized by cytokine receptor activity and glycosaminoglycan binding (Fig. 4B).

Fig. 4.

GO and KEGG enrichment analyses of differentially m6A-methylated mRNAs. A, B The top enriched GO terms for hypermethylated (A) and hypomethylated (B) mRNAs. C, D The top enriched KEGG pathways for hypermethylated (C) and hypomethylated (D) mRNAs. E, F Chord diagrams displaying the specific hypermethylated (E) and hypomethylated (F) mRNAs involved in key pathways

KEGG pathway enrichment analysis revealed that both hypermethylated and hypomethylated mRNAs were significantly enriched in the Hippo, PI3K-Akt, MAPK, and Ras signaling pathways, suggesting a core epitranscriptomic regulatory network across these cascades. Notably, the JAK-STAT signaling pathway was uniquely enriched within the hypermethylated group (Fig. 4C, D). To further characterize these regulatory patterns, KEGG chord diagrams were employed to visualize the distribution of hypermethylated and hypomethylated mRNAs involved in these key pathways (Fig. 4E, F).

Differential mRNA Expression Profiles in NECs of AIT and non-AIT AR Patients

To characterize transcriptomic changes associated with AIT in AR patients, we performed RNA-Seq to identify the differentially expressed mRNA (DEGs) using the criteria of adjusted P < 0.05 and |log₂(FC)| ≥ 1. Hierarchical clustering revealed distinct transcriptomic signatures, clearly separating the AIT group from the non-AIT group (Fig. 5A). A total of 588 DEGs were identified in the AIT-treated group, comprising 415 upregulated and 173 downregulated mRNAs (Fig. 5B).

Fig. 5.

Differential expression profiling and functional enrichment analyses of mRNAs in NECs. A Hierarchical clustering heatmap showing distinct mRNA expression patterns between the AIT and non-AIT groups. B Volcano plot identifying upregulated and downregulated mRNAs. C Bar plot of the top enriched GO terms for differentially expressed mRNAs, visualized based on the Rich Factor and FDR. D Bubble plot of the top enriched KEGG pathways for differentially expressed mRNAs, visualized based on the rich factor, FDR, and gene number

Furthermore, GO enrichment analysis was conducted to elucidate the biological roles of these DEGs. In the Biological Process (BP) category, DEGs were significantly enriched in cell adhesion, the apoptotic process, cell proliferation, and extracellular matrix organization. Cellular Component (CC) annotations identified the cytoplasm, membrane, cytosol, plasma membrane, and nucleus as the primary locations of DEG activity. Molecular Function (MF) analysis further highlighted protein binding and metal ion binding as key functional categories (Fig. 5C).

KEGG pathway analysis further elucidated the therapeutic cascades modulated by AIT. The identified DEGs were predominantly enriched in the complement and coagulation cascades, PI3K-Akt signaling pathway, oxidative phosphorylation, and cytokine-cytokine receptor interaction (Fig. 5D).

Integrated analysis of m6A methylation and gene expression in NECs following AIT

To elucidate the interplay between m6A modification and gene expression in the context of AIT, we performed an integrated analysis of RNA-Seq and MeRIP-Seq datasets. A total of 634 mRNAs exhibiting both differential m6A methylation and altered expression (differentially methylated and expressed mRNAs, DMEGs) were stratified into four categories. As illustrated in the four-quadrant plot, these DMEGs consisted of 61 hypermethylated-upregulated (hyper-up), 172 hypermethylated-downregulated (hyper-down), 198 hypomethylated-upregulated (hypo-up), 203 hypomethylated-downregulated (hypo-down) mRNAs (Fig. 6A, B).

Fig. 6.

Integrated analysis of m6A modification and mRNA expression profiles. A Four-quadrant plot showing differentially methylated and expressed mRNAs (DMEGs) between the AIT and non-AIT groups. B Bar plot showing the classification of these DMEGs. C, D GO enrichment analyses of hypermethylated (C) and hypomethylated (D) DMEGs. E, F KEGG enrichment analyses of hypermethylated (E) and hypomethylated (F) DMEGs

To further characterize the functional implications of the concerted regulation between m6A methylation and gene expression in NECs of AIT-treated AR patients, we expanded our analysis to the 634 genes with concurrent differential m6A peaks and expression changes. For the hypermethylated group, GO terms were prominently enriched in apoptotic process, cellular response to mechanical stimulus, and cell redox homeostasis (Fig. 6C). Corresponding KEGG pathways highlighted oxidative phosphorylation, tight junctions, and sphingolipid metabolism (Fig. 6D).

For the hypomethylated group, significant GO annotations included immune system process, apoptotic process, cell adhesion, and cytokine-mediated signaling (Fig. 6E). Correspondingly, KEGG analysis revealed enrichment in tight junctions, complement and coagulation cascades, focal adhesion, and the PI3K-Akt, MAPK, and Ras signaling pathways (Fig. 6F). The top enriched GO terms and KEGG pathways for each subgroup are presented in Tables 4, 5, 6, and 7.

Table 4.

GO enrichment terms of differentially hypermethylated genes between AIT and non-AIT groups

GO category	GO ID	GO term	Rich factor	− log10 (FDR)	Gene number
Biological process	GO:0034109	Homotypic cell–cell adhesion	0.6667	1.768269	2
	GO:0032757	Positive regulation of interleukin-8 production	0.5	1.489968	2
	GO:0045348	Positive regulation of MHC class II biosynthetic process	0.5	1.489968	2
	GO:0045454	Cell redox homeostasis	0.3077	1.840905	4
	GO:0000387	Spliceosomal snRNP assembly	0.3	1.434848	3
	GO:0071260	Cellular response to mechanical stimulus	0.2941	2.114579	5
	GO:0000398	mRNA splicing, via spliceosome	0.1456	1.922779	15
	GO:0006397	mRNA processing	0.1429	2.307236	20
	GO:0008380	RNA splicing	0.1382	1.893562	17
	GO:0006915	Apoptotic process	0.1216	1.455329	18
Cellular Component	GO:0005758	Mitochondrial intermembrane space	0.4348	5.512831	10
	GO:0070469	Respirasome	0.2667	1.61347	4
	GO:1,990,904	Ribonucleoprotein complex	0.2222	3.181795	12
	GO:0005681	Spliceosomal complex	0.1644	2.029057	12
	GO:0005840	Ribosome	0.1636	1.62435	9
	GO:0005743	Mitochondrial inner membrane	0.1591	2.158157	14
	GO:0005739	Mitochondrion	0.1526	5.92221	49
	GO:0005615	Extracellular space	0.1302	2.00311	22
	GO:0005783	Endoplasmic reticulum	0.1187	2.131531	33
Molecular Function	GO:0005507	Copper ion binding	0.5556	3.581073	5
	GO:0008307	Structural constituent of muscle	0.3333	1.973791	4
	GO:0035591	Signaling adaptor activity	0.3	1.434848	3
	GO:0016853	Isomerase activity	0.25	2.350585	7
	GO:0001227	DNA-binding transcription repressor activity, RNA polymerase II-specific	0.2	1.992925	8
	GO:0044325	Transmembrane transporter binding	0.2	1.401683	5
	GO:0051082	Unfolded protein binding	0.1892	1.671519	7
	GO:0003735	Structural constituent of ribosome	0.1628	1.347572	7
	GO:0016491	Oxidoreductase activity	0.1538	2.029215	14

Table 5.

GO enrichment terms of differentially hypomethylated genes between AIT and non-AIT groups

GO category	GO ID	GO term	Rich factor	− log10 (FDR)	Gene number
Biological process	GO:0002224	Toll-like receptor signaling pathway	0.3333	2.623337	5
	GO:0098609	Cell–cell adhesion	0.2222	4.222429	14
	GO:0030198	Extracellular matrix organization	0.1935	3.110939	12
	GO:0001525	Angiogenesis	0.1758	3.464391	16
	GO:0019221	Cytokine-mediated signaling pathway	0.1413	2.062794	13
	GO:0006811	Monoatomic ion transport	0.1384	3.011306	22
	GO:0043312	Neutrophil degranulation	0.1302	2.661421	22
	GO:0007155	Cell adhesion	0.1287	2.977629	26
	GO:0043066	Negative regulation of apoptotic process	0.1231	2.524853	24
	GO:0002376	Immune system process	0.1218	2.056425	19
Cellular component	GO:0070160	Tight junction	0.7143	4.580873	5
	GO:0009986	Cell surface	0.163	5.325456	30
	GO:0005615	Extracellular space	0.1325	4.822236	42
	GO:0005925	Focal adhesion	0.12	2.146523	21
	GO:0005576	Extracellular region	0.1166	4.425985	54
	GO:0030054	Cell junction	0.11	2.391592	32
	GO:0005886	Plasma membrane	0.1017	7.350035	135
	GO:0042995	Cell projection	0.101	2.161792	39
	GO:0070062	Extracellular exosome	0.0983	3.129365	68
Molecular function	GO:0004867	Serine-type endopeptidase inhibitor activity	0.5882	7.63353	10
	GO:0030414	Peptidase inhibitor activity	0.44	6.643056	11
	GO:0005178	Integrin binding	0.2564	3.707142	10
	GO:0002020	Protease binding	0.2143	2.780398	9
	GO:0019900	Kinase binding	0.186	2.131203	8
	GO:0051015	Actin filament binding	0.1507	2.021939	11
	GO:0005509	Calcium ion binding	0.1436	3.728537	26
	GO:0008017	Microtubule binding	0.1348	1.788087	12
	GO:0045296	Cadherin binding	0.1136	1.483743	15

Table 6.

KEGG enrichment pathways of differentially hypermethylated genes between AIT and non-AIT groups

Pathway ID	KEGG catalogue	Pathway name	Genes	Rich factor	FDR
ko04530	Cellular processes	Tight junction	F11R, TJP2, MYL6, MYH11, MYL12A, ACTN4, MYH7B, MYH9	0.18	0.002
ko04217	Cellular processes	Necroptosis	PPID, VDAC3, TNFRSF10A, TLR4, VDAC2, CHMP4B	0.22	0.004
ko04218	Cellular processes	Cellular senescence	NFKB1, PPID, VDAC3, SIRT1, VDAC2	0.16	0.029
ko04216	Cellular processes	Ferroptosis	VDAC3, VDAC2	0.50	0.046
ko04071	Environmental information processing	Sphingolipid signaling pathway	SGPP2, NFKB1, SGMS2, SGMS1, PLCB2	0.19	0.018
ko04020	Environmental information processing	Calcium signaling pathway	VDAC3, VDAC2, TPCN1, PLCB2	0.17	0.049
ko04066	Environmental information processing	HIF-1 signaling pathway	NFKB1, TLR4, EGLN3	0.18	0.017
ko03040	Genetic information processing	Spliceosome	SYF2, SNRPG, DHX15, THOC3, HSPA1A, PPIL1, HNRNPA1, SRSF9, SNRPD1, SNRPB, U2AF1L5	0.18	< 0.001
ko04141	Genetic information processing	Protein processing in endoplasmic reticulum	ERO1B, EDEM1, SEC62, HSPA1A, PDIA4, DNAJA1, MBTPS1, PPP1R15A	0.15	0.005
ko03010	Genetic information processing	Ribosome	RPL22, RPL26L1, RPS23, RPL36AL, RPS15A	0.14	0.046
ko00190	Metabolism	Oxidative phosphorylation	NDUFS5, COX17, COX7C, UQCRQ, NDUFB8, ATP5PO, NDUFA6	0.26	0.001
ko00600	Metabolism	Sphingolipid metabolism	SGPP2, SGMS2, SGMS1	0.43	0.014
ko00330	Metabolism	Arginine and proline metabolism	AGMAT, P4HA1, PRODH	0.38	0.018
ko00650	Metabolism	Butanoate metabolism	HMGCL, HMGCS1	0.67	0.029
ko00030	Metabolism	Pentose phosphate pathway	PGD, RPE	0.50	0.046
ko04714	Organismal systems	Thermogenesis	NDUFS5, COA6, COX17, COX7C, UQCRQ, NDUFB8, ATP5PO, NDUFA6	0.19	0.002
ko04621	Organismal systems	NOD-like receptor signaling pathway	NFKB1, VDAC3, TLR4, VDAC2, IL18, PLCB2	0.19	0.007
ko04623	Organismal systems	Cytosolic DNA-sensing pathway	NFKB1, POLR3D, IL18	0.38	0.018
ko04723	Organismal systems	Retrograde endocannabinoid signaling	NDUFS5, NDUFB8, PLCB2, NDUFA6	0.17	0.046
ko04270	Organismal systems	Vascular smooth muscle contraction	CYP4A11, MYL6, PLCB2, MYH11	0.17	0.049
ko04915	Organismal systems	Estrogen signaling pathway	HSPA1A, ESR2, PLCB2, KRT19	0.17	0.049

Table 7.

KEGG enrichment pathways of differentially hypomethylated genes between AIT and non-AIT groups

Pathway ID	KEGG catalogue	Pathway name	Genes	Rich factor	FDR
ko04530	Cellular processes	Tight junction	CLDN1, MARVELD2, TJAP1, CLDN4, SCRIB, WHAMM, TJP1, ACTG1, MYL12A, PARD6B, MYH9	0.16	0.001
ko04510	Cellular processes	Focal adhesion	VEGFA, ITGB8, VCL, BIRC2, CCND2, FLT1, BCAR1, ACTG1, MYL12A, LAMA3, FLNA	0.16	0.001
ko04810	Cellular processes	Regulation of actin cytoskeleton	PIP5K1A, ITGB8, VCL, ARPC3, BDKRB2, CFL2, BCAR1, ACTG1, MYL12A, MYH9	0.13	0.004
ko04145	Cellular processes	Phagosome	DYNC1I2, DYNC1LI1, HLA-DPA1, CD36, TLR4, CLEC7A, ACTG1	0.13	0.025
ko04218	Cellular processes	Cellular senescence	RASSF5, CALM2, SERPINE1, IGFBP3, VDAC2, CCND2, NFATC3	0.11	0.039
ko04151	Environmental information processing	PI3K-Akt signaling pathway	IL6R, EFNA1, EPHA2, MCL1, OSMR, VEGFA, ITGB8, NTRK2, TLR4, CCND2, FLT1, LAMA3	0.11	0.004
ko04010	Environmental information processing	MAPK signaling pathway	EFNA1, EPHA2, HSPA1B, VEGFA, NTRK2, DUSP5, FLT1, NFATC3, CACNA1A, RPS6KA3, FLNA	0.10	0.018
ko04015	Environmental information processing	Rap1 signaling pathway	EFNA1, RASSF5, EPHA2, CALM2, VEGFA, FLT1, BCAR1, ACTG1, PARD6B	0.13	0.010
ko04014	Environmental information processing	Ras signaling pathway	EFNA1, RASSF5, EPHA2, CALM2, VEGFA, NTRK2, FLT1, ETS2	0.11	0.025
ko04066	Environmental information processing	HIF-1 signaling pathway	IL6R, VEGFA, IFNGR1, SERPINE1, TLR4, FLT1, EGLN3	0.20	0.004
ko04141	Genetic information processing	Protein processing in endoplasmic reticulum	ERO1B, SEC62, SSR3, HSPA1B, HSPH1, SELENOS, MBTPS1, PPP1R15A, XBP1	0.13	0.007
ko03410	Genetic information processing	Base excision repair	OGG1, PARP4, XRCC1	0.43	0.018
ko03060	Genetic information processing	Protein export	SEC62, SRP54, SRP14	0.30	0.036
ko00190	Metabolism	Oxidative phosphorylation	UQCRHL, NDUFS4, UQCRQ, MT-ND2, MT-ND6	0.13	0.010
ko00450	Metabolism	Selenocompound metabolism	KYAT1, PAPSS2	0.33	0.015
ko04360	Organismal systems	Axon guidance	EFNA1, EPHA2, EPHB3, SEMA6A, CFL2, NFATC3, MYL12A, PARD6B	0.11	0.027
ko04714	Organismal systems	Thermogenesis	UQCRHL, KDM3A, NDUFS4, UQCRQ, ACTG1, RPS6KA3, MT-ND2, MT-ND6	0.11	0.027
ko04670	Organismal systems	Leukocyte transendothelial migration	RASSF5, CLDN1, CLDN4, VCL, BCAR1, ACTG1, MYL12A	0.19	0.004
ko04610	Organismal systems	Complement and coagulation cascades	CD55, PROC, SERPINE1, BDKRB2, SERPINB2, PLAUR	0.35	0.001
ko04640	Organismal systems	Hematopoietic cell lineage	IL6R, CD55, HLA-DPA1, CD36, CD44, CD9	0.27	0.003

To further validate the biological significance of these enriched pathways, we performed Gene Set Enrichment Analysis (GSEA) on the transcriptome data. Consistent with our integrated findings, GSEA demonstrated that the regulation of inflammatory response, PI3K-Akt signaling, MAPK signaling, and Ras signaling pathways were significantly downregulated in the AIT group. Conversely, terms related to the regulation of cell adhesion and the regulation of apoptotic signaling pathways exhibited a significant upward trend (Fig. 7). These results collectively reinforce the pivotal role of m6A modification in orchestrating the transition from a pro-inflammatory state to mucosal homeostasis during AIT.

Fig. 7.

GSEA validation of key pathways identified from the integrated MeRIP-seq and RNA-seq analysis. A–C GSEA enrichment plots for GO biological processes, including the regulation of inflammatory response (A), negative regulation of cell adhesion (B), and negative regulation of apoptotic signaling pathway (C). D–F GSEA enrichment plots for key KEGG signaling pathways, including the PI3K-Akt (D), MAPK (E), and Ras signaling pathways (F)

Screening and validation of candidate genes associated with AIT efficacy

To validate the reliability and clinical reproducibility of our findings, we expanded our study by including an independent validation cohort of 6 AIT-treated AR patients and 6 matched control subjects. Given that m6A modifications were most abundant in the 3’UTRs in our dataset, and considering the canonical role of 3’UTR methylation in mRNA destabilization [17], we prioritized genes showing an inverse correlation between m6A methylation in 3’UTRs and mRNA expression. By integrating the transcriptomic and m6A methylation datasets, we screened for genes exhibiting such inverse patterns. This analysis yielded 25 genes in the hypermethylated/downregulated group and 10 genes in the hypomethylated/upregulated group. To further narrow down the targets, we mapped these genes to the inflammatory signaling, mucosal barrier function, and cell survival pathways identified in our enrichment analysis. This screening process highlighted 4 distinct candidate genes: KANK2, DBI, IL18, and TNFRSF10A. Subsequently, we validated the expression and methylation status of these 4 candidates in the validation cohort (n = 12) using qPCR and MeRIP-qPCR assays (Table 8). Consistent with our sequencing data, KANK2 exhibited reduced m6A methylation levels coupled with elevated mRNA expression in the AIT group compared to the non-AIT group. Conversely, DBI, IL18, and TNFRSF10A displayed increased m6A methylation levels and downregulated expression in the AIT group (Fig. 8A-D).

Table 8.

Clinicopathological characteristics of 12 patients with AR

	Patients	Gender	Age	Total IgE (IU/ml)	TNSS	VAS
AIT	A	Male	45	126	7	17
	B	Female	23	48	5	9
	C	Female	38	198	6	14
	D	Female	34	71	5	15
	E	Male	29	119	8	20
	F	Female	36	163	6	8
non-AIT	G	Male	37	307	9	28
	H	Female	41	352	12	46
	I	Female	44	224	9	44
	G	Female	25	255	7	22
	K	Male	29	287	11	33
	L	Male	33	193	12	37
p value		0.60	0.88	0.0026	0.0023	0.0005

Fig. 8.

Validation of m6A-modified differentially expressed mRNAs. A Venn diagram showing the overlap of differentially methylated and differentially expressed genes between the AIT and non-AIT groups (|log₂(FC)|≥ 1, adjusted p < 0.05). B IGV visualization of m6A peaks in the AIT and non-AIT groups. C, D qRT-PCR verification of relative expression levels (C) and MeRIP-qPCR verification of m6A modification levels (D) for TNFRSF10A, DBI, IL18, and KANK2 in the AIT group compared to the non-AIT group. E, F Linear correlations between the relative expression (E) or m6A modification levels (F) of these 4 genes and the patients’ TNSS. Data are presented as means ± SEMs (n = 6 per group); *p < 0.05, **p < 0.01

Finally, to assess the clinical relevance of these transcriptomic alterations, we performed Pearson correlation analyses between the patients’ TNSS and both the m6A modification and expression levels of these candidate genes. Notably, the m6A methylation level of KANK2 was positively correlated with the patients’ TNSS, while its mRNA expression showed a negative correlation. Conversely, DBI showed an inverse pattern, where its m6A methylation levels were negatively correlated with the patients’ TNSS, accompanied by a positive correlation in mRNA expression. Unlike DBI, while IL18 and TNFRSF10A exhibited a significant positive correlation between mRNA expression and the patients’ TNSS, their m6A methylation levels failed to show a significant correlation. These findings collectively implicate these 4 genes as promising candidates underlying the therapeutic efficacy of AIT in AR patients (Fig. 8E, F).

Correlation analysis of lncRNA m6A modification and epigenetics in AIT-treated NECs

Previous studies have indicated an interaction between RNA methylation and histone/DNA epigenetic modifications [18], in which lncRNA methylation plays a functional role. To investigate the association between lncRNA m6A modification and epigenetics in AIT-treated samples, we screened for lncRNAs exhibiting differential m6A methylation levels in the AIT treatment group and subsequently analyzed their subcellular localization using the RNALOCATE database (http://www.rnalocate.org/). Notably, among the 998 differentially m6A-modified lncRNAs identified, 285 (28.6%) were localized to chromatin, suggesting a potential link between these differentially m6A-modified lncRNAs and chromatin modification or remodeling during AIT treatment (Fig. 9A).

Fig. 9.

Validation of m6A-modified differentially expressed chromatin-associated lncRNAs. A Venn diagram illustrating the intersection between differentially m6A-methylated lncRNAs and chromatin-associated lncRNAs annotated in the RNALOCATE database. B Venn diagram showing the overlap between the identified chromatin-associated m6A-modified lncRNAs and differentially expressed lncRNAs. C IGV visualization of m6A peaks in the AIT and non-AIT groups. D MeRIP-qPCR verification of m6A modification levels of AC105219, LUCAT1, and SNHG14 in the AIT group compared to the non-AIT group. E Linear correlations between the m6A modification levels of these 3 lncRNAs and the patients’ TNSS. Data are presented as means ± SEMs (n = 6 per group); *p < 0.05, **p < 0.01

To further pinpoint lncRNAs with significant functional potential, we performed an intersection analysis between the differentially m6A-methylated lncRNAs and those exhibiting differential expression, which yielded 23 candidate lncRNAs (Fig. 9B). Given the established role of m6A-modified lncRNAs in orchestrating chromatin modification complexes and epigenetic remodeling, we prioritized candidates based on the FC of their m6A methylation. Consequently, we selected the top 3 candidates, LUCAT1, AC105219, and SNHG14, for further validation (Fig. 9C and Table 9). Corroborating the sequencing results, MeRIP-qPCR analysis revealed downregulated m6A levels for LUCAT1 and AC105219, alongside upregulated levels for SNHG14 following AIT treatment (Fig. 9D). Notably, clinical correlation analysis highlighted a positive association between LUCAT1 m6A methylation levels and the patients’ TNSS (Fig. 9E). These findings suggest that the m6A modification of LUCAT1 may serve as a critical epigenetic regulator during AIT, potentially influencing therapeutic efficacy through chromatin-mediated mechanisms.

Table 9.

Top 10 differentially m6A-modified chromatin-associated lncRNAs based on fold change

geneName	Chromosome	Start	End	Strand	AIT peak	Non-AIT peak	Log2FC	Adjusted p-value
LUCAT1	chr5	91,314,239	91,314,414	–	0.72	3.75	− 3.03	< 0.01
AC105219	chr8	143,758,152	143,771,863	–	3.02	5.22	– 2.20	< 0.01
SNHG14	chr15	25,016,431	25,016,531	+	1.24	− 0.56	1.80	< 0.01
AP001107	chr11	66,272,031	66,272,181	–	2.17	1.03	1.14	0.01
GABPB1-AS1	chr15	50,370,298	50,371,323	+	2.39	3.35	− 0.96	0.01
AL162258	chr1	153,631,437	153,632,095	–	2.18	3.09	− 0.91	0.01
ZNF451-AS1	chr6	57,116,512	57,150,687	–	2.55	3.33	− 0.78	< 0.01
SNHG5	chr6	85,678,074	85,678,824	–	2.42	3.16	− 0.74	0.01
SNHG29	chr17	16,441,673	16,441,848	+	1.00	1.61	− 0.61	0.02
MIR4458HG	chr5	8,457,630	8,460,296	+	3.14	2.56	0.57	0.04

Discussion

m6A constitutes a ubiquitous and dynamic modification in eukaryotic transcripts, serving as a master regulator of RNA fate by influencing stability, splicing, transport, and translation [19–21]. Accumulating evidence has highlighted the pivotal role of m6A methylation in the immune system, particularly in orchestrating T cell differentiation [22] and modulating inflammatory responses in allergic diseases such as asthma [23] and AR [24]. However, the specific epitranscriptomic landscape associated with AIT for AR remains largely unexplored. To address this limitation, we performed a comprehensive analysis of the m6A methylome in NECs from AR patients following AIT treatment, providing novel insights into the epitranscriptomic mechanisms underlying therapeutic efficacy.

In our study, MeRIP-seq analysis revealed that while gene-specific m6A levels varied between groups, the global topology remained conserved with a predominant enrichment in the 3′ UTR, consistent with the established role of 3′ UTR in RNA metabolism [25]. This distribution suggests that dynamic m6A modifications likely act as critical post-transcriptional switches, modulating the stability and translation of AIT-responsive genes. Beyond the transcriptomic topology, we observed a distinct chromosomal preference for differential m6A peaks, which were predominantly enriched on chromosome 1. As the largest human chromosome, Chromosome 1 harbors approximately 2,800 genes (representing 8%–10% of the genome) and contains a dense cluster of immune-related loci involved in metabolism, signal transduction, and inflammatory responses [26]. This specific enrichment suggests that m6A modification may coordinate core cellular processes, particularly inflammatory responses and immune regulation, thereby directly impacting the therapeutic efficacy of AIT.

Allergic diseases are characterized by high heterogeneity, impairing quality of life through diverse pathophysiological mechanisms [27]. Distinct from symptomatic treatments such as antihistamines and β-agonists, AIT represents the only currently available therapeutic strategy targeting the underlying etiology, thereby modifying the immune response rather than merely masking symptoms [28, 29]. However, current approaches for evaluating AIT efficacy remain largely restricted to clinical symptom scores and medication usage, lacking objective biological markers to directly monitor these immunological changes. To address this gap, we performed an integrated analysis of the m6A methylome and transcriptome in NECs from AIT-treated patients. Our analysis revealed that the differentially m6A-methylated and expressed genes were significantly enriched in biological functions such as cell adhesion and apoptosis, as well as inflammation-related pathways, including the MAPK and PI3K-Akt signaling pathways. These findings suggest that AIT induces profound remodeling of epithelial tissue homeostasis and inflammatory modulation beyond simple immune suppression.

The integrity of the nasal epithelial barrier relies on the coordinated regulation of cell adhesion and programmed cell death [30]. Chronic inflammation in AR compromises this physical defense by disrupting junctional complexes and inducing excessive epithelial apoptosis [31, 32]. In our study, we identified a consistent inverse correlation between m6A methylation and gene expression in key transcripts governing these processes. Specifically, the upregulation of adhesion molecules and survival factors, epitranscriptomically licensed by reduced m6A abundance, suggests that AIT promotes epithelial restoration. By reinforcing intercellular connections and preventing aberrant apoptosis, AIT effectively reconstructs the damaged mucosal barrier and re-establishes tissue homeostasis.

The pathogenesis of AR involves a complex signaling network where multiple cascades orchestrate the inflammatory response. Among these, the MAPK and PI3K-Akt pathways function as key drivers, promoting pro-inflammatory cytokine production and effector cell survival via NF-κB activation [33]. Notably, our enrichment analysis specifically highlighted these pathways as primary targets of m6A modification, with significant downregulation observed in the AIT-treated group. This dampening of hyperactive signaling axes aligns with recent multi-omics evidence suggesting that clinical efficacy in immunotherapy relies on a delicate homeostatic rebalancing, characterized by the dominance of regulatory T-cell (Treg) responses and the concurrent suppression of B-cell activity [34]. Furthermore, this molecular reprogramming likely shapes a permissive niche that underpins the restoration of specific innate and adaptive immune subsets. Corroborating this, recent longitudinal methylome profiling has pinpointed that successful AIT is driven by a coordinated shift in cellular composition, specifically involving the regulation of innate populations and the enrichment of adaptive effectors, which are pivotal for maintaining long-term tolerance [35]. Therefore, the m6A-mediated downregulation of MAPK and PI3K-Akt signaling likely serves as a critical epitranscriptomic switch, restricting pro-inflammatory effector responses to create a permissive microenvironment for these protective cell populations to re-establish tissue homeostasis.

While molecular biomarkers such as FeNO [36] and periostin [37] are utilized to monitor therapeutic responses in other allergic conditions like asthma, biological indicators for AIT efficacy in AR remain undefined. In our study, we uncovered a stable inverse correlation between m6A methylation and the expression of key genes governing both epithelial barrier integrity and the immune microenvironment. Specifically, we observed the epitranscriptomic modulation of KANK2, which is pivotal for maintaining epithelial barrier integrity and focal adhesion dynamics [38]. Concurrently, we identified alterations in DBI [39], IL18 [40], and TNFRSF10A [41] as critical regulators of the inflammatory microenvironment and immune cell survival. The prominent enrichment of these genes suggests that the therapeutic efficacy of AIT is likely mediated through an m6A-dependent mechanism that coordinates structural repair with the precise modulation of the inflammatory microenvironment. Consequently, KANK2, DBI, IL18, and TNFRSF10A emerge as promising candidates underlying the therapeutic effects of AIT.

Epigenetic remodeling is increasingly recognized as a fundamental driver of immune memory and allergic inflammation [42]. Mechanistically, extensive studies have characterized lncRNAs as modular scaffolds that recruit chromatin-modifying complexes, such as the Polycomb Repressive Complex 2 (PRC2) or histone methyltransferases, to specific genomic loci [43]. Through these interactions, lncRNAs guide the deposition of histone marks and modulate DNA accessibility, thereby reshaping the epigenetic landscape to control gene expression patterns [44]. Despite this established regulatory paradigm, the specific role of m6A modification in governing these lncRNA-chromatin interactions during AIT remains undefined. In our study, we observed that 28.7% of the differentially m6A-methylated lncRNAs were chromatin-associated. This significant enrichment suggests that m6A modification might act as a structural switch, influencing the retention of lncRNAs on chromatin or their affinity for epigenetic modifiers. Guided by this regulatory paradigm, we identified LUCAT1, AC105219, and SNHG14 as candidates bridging epitranscriptomic modifications and chromatin remodeling. Notably, the specific correlation between LUCAT1 m6A methylation and the patients’ TNSS further underscores its potential as a critical epigenetic regulator orchestrating the therapeutic efficacy of AIT.

Notably, as a preliminary investigation, this study has limitations inherent to its exploratory design. The MeRIP-seq analysis was conducted on a relatively small cohort, consisting of 3 biological replicates per group. Consequently, while offering a novel snapshot of the epitranscriptomic landscape, these findings necessitate further validation in larger, independent cohorts. Second, as an exploratory study based on high-throughput sequencing, our conclusions are primarily hypothesis-generating. Although integrated analysis and qPCR validation confirmed our bioinformatic predictions, the functional consequences of these modifications remain to be fully characterized. Therefore, future mechanistic studies utilizing in vitro cell lines and in vivo animal models are essential to transform these initial insights into mechanistic understanding.

Conclusions

In conclusion, this study characterizes the distinct m6A modification landscape and mRNA expression profiles in NECs from AIT-treated versus untreated AR patients. Our findings suggest that m6A modification serves as a critical epitranscriptomic regulator, coordinating epithelial barrier integrity and immune microenvironment homeostasis. Specifically, we identified KANK2, DBI, IL18, and TNFRSF10A as key downstream targets potentially mediating the therapeutic effects of AIT. Furthermore, our analysis highlights that specific chromatin-associated lncRNAs, notably LUCAT1, AC105219, and SNHG14, may orchestrate these responses through epigenetic mechanisms. Collectively, this exploratory study provides a foundational framework for future mechanistic investigations, offering novel insights that may facilitate the development of precision therapeutic strategies for AR.

Materials and methods

Study design and participant recruitment

The study protocol was approved by the Ethics Committee of The First Affiliated Hospital, Sun Yat-sen University (Guangdong, China) (Approval No. 2022-354). Written informed consent was obtained from all participants. Patients were recruited in accordance with the ARIA guidelines [16]. Eligibility was defined by a documented history of persistent house dust mite (HDM)-induced AR, characterized by symptoms lasting ≥ 4 weeks per year, a moderate-to-severe symptom burden (Total Nasal Symptom Score (TNSS) ≥ 6 and Visual Analogue Scale (VAS) ≥ 5), and confirmed sensitization via a skin prick test with HDM extract (mean wheal diameter ≥ 5 mm relative to the negative control).

Exclusion criteria included comorbid allergic or inflammatory conditions (e.g., asthma, atopic dermatitis, chronic rhinosinusitis), systemic immune disorders, neurological or psychiatric conditions, severe organ dysfunction, and pregnancy. Additionally, patients who had received leukotriene receptor antagonists, antihistamines, or corticosteroids within 1-month preceding enrollment were excluded.

Enrolled AR patients were stratified into two cohorts: those who had completed a 3-year course of specific AIT and treatment-naive individuals with no prior AIT exposure. All participants underwent nasal mucosa provocation tests, during which nasal symptoms were evaluated and quantified using the TNSS and VAS. Concomitantly, NEC samples were collected using sterile nasal brushes for subsequent experimental analyses.

Nasal brushing sample collection and RNA preparation

NECs were harvested from AIT and non-AIT AR patients. Samples from 6 patients (3 matched pairs of AIT and non-AIT patients) were utilized for comprehensive profiling of RNA m6A modification, while samples from the remaining participants (6 AIT and 6 non-AIT patients) were employed for experimental validation of key findings.

NECs were obtained from bilateral inferior turbinate under nasal endoscope using sterile Histobrush Cytology Brushes (Puritan Medical Products, Guilford, ME, USA). For each nasal cavity, the brush was gently rotated twice to ensure adequate cell collection, then immediately transferred to an Eppendorf tube containing 500 μL of RPMI medium supplemented with 1 × antibiotic/antimycotic, 1 mM sodium pyruvate, and 10% fetal bovine serum (FBS). Trizol reagent (Invitrogen, Carlsbad, CA, USA) was promptly added to the cell suspensions, which were then stored at − 80 °C until subsequent RNA extraction.

MeRIP-seq and RNA-seq library construction and sequencing

Total RNA was extracted and purified from the collected samples following the manufacturer’s standard protocol for TRIzol reagent (Invitrogen, Carlsbad, CA, USA). RNA concentration and purity were quantified using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), while RNA integrity was assessed via an Agilent Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, USA) with a minimum RNA Integrity Number (RIN) of 7.0. RNA integrity was further confirmed by denaturing agarose gel electrophoresis.

Approximately 1 μg of total RNA was fragmented into short segments using the Magnesium RNA Fragmentation Module (NEB, cat. E6150, Ipswich, MA, USA) at 94 °C for 5 min. Ribosomal RNA (rRNA)-depleted RNA fragments were incubated with m6A antibody-conjugated Dynabead complexes for immunoprecipitation. The immunoprecipitated RNA was reverse-transcribed into first-strand cDNA using SMARTScribe™ Reverse Transcriptase (CloneTech, cat. 634,414, Tokyo, Japan), followed by adapter ligation and second-strand cDNA synthesis via polymerase chain reaction (PCR) under the following conditions: initial denaturation at 94 °C for 1 min; 5 cycles of 98 °C for 15 s (denaturation), 55 °C for 15 s (annealing), and 68 °C for 30 s (extension); and a final extension at 68 °C for 2 min.

Amplified DNA-seq libraries were purified using magnetic bead-based separation. cDNA derived from rRNA reverse transcription was eliminated using ZapR v2 and R-Probes v2 (for mammals) under the following thermal conditions: 72 °C for 2 min, 4 °C for 2 min, 37 °C for 1 h, 72 °C for 10 min, and holding at 4 °C. The purified libraries were subjected to a second round of PCR amplification using the aforementioned cycling parameters, with the number of cycles adjusted to 12–16.

Sequencing was performed as 2 × 150 bp paired-end reactions on an Illumina Novaseq™ 6000 platform (LC-Bio Technology CO., Ltd., Hangzhou, China) following the manufacturer’s recommended protocols.

Sequencing data analysis

After sequencing, raw reads were filtered to remove adapter sequences, low-quality bases (Phred score < 20), and reads containing undetermined nucleotides using fastp software (https://github.com/OpenGene/fastp) with default parameters. The quality of filtered reads from both immunoprecipitated (IP) and input samples was evaluated using FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and RseQC (http://rseqc.sourceforge.net/). High-quality clean reads were aligned to the Homo sapiens reference genome (GRCh38.p13, Version: v101) via HISAT2 (http://daehwankimlab.github.io/hisat2). Peak calling and differential peak analysis were conducted using the R package exomePeak2 (https://bioconductor.org/packages/release/bioc/html/exomePeak2.html). Peak annotations were generated by intersecting peak regions with gene structural annotations using the R package ANNOVAR (http://www.openbioinformatics.org/annovar/). De novo and known motif discovery, along with motif localization relative to peak summits, were conducted using MEME (http://meme-suite.org) and HOMER (http://homer.ucsd.edu/homer/motif). Transcript and gene expression levels from input libraries were quantified as FPKM (total exon fragments / (mapped reads [millions] × exon length [kB])) using StringTie (https:// ccb.jhu.edu/software/stringtie). Differentially expressed transcripts and genes were identified using the R package edgeR (https://bioconductor.org/packages/edgeR) with thresholds set at |log2(fold change)|≥ 1 and padj < 0.05.

Integrated analysis of MeRIP-seq and RNA-seq data

Significant differential m6A peaks and their corresponding differential expression datasets were subjected to integrative analysis. Genes harboring significant differential m6A peaks were categorized into four distinct subgroups: hypermethylated-upregulated, hypermethylated-downregulated, hypomethylated-upregulated, and hypomethylated-downregulated. A four-quadrant scatter plot was constructed using the ggplot2 package (https://github.com /tidyverse/ggplot2) to visualize these subgroups. Genes were defined as exhibiting concurrent perturbations in both methylation status and transcriptional profiles between the AIT and non-AIT cohorts if they satisfied the following thresholds: |log₂(FC of differential peaks)|≥ 1, BH adjusted P-value of differential peaks < 0.05, |log₂(FC of differentially expressed genes)|≥ 1, and BH adjusted P-value of differentially expressed genes < 0.05. Candidate genes of interest were visually inspected using the Integrative Genomics Viewer (IGV; version 2.13.1).

Functional enrichment analysis

Functional enrichment analyses were conducted using genes associated with differentially regulated m6A peaks and differentially expressed genes. For GO enrichment analysis, the target gene set was annotated against the three core ontologies: MF, BP, and CC. Statistical significance of enriched GO terms was determined by a combination of BH adjusted P-value (padj < 0.05) and false discovery rate (FDR < 0.05). For each significant GO term, the number of genes linked to altered m6A peaks or enriched within the term was quantified, and detailed gene attributes were retrieved from the Gene Ontology database (http://www.geneontology.org).

For KEGG pathway enrichment analysis, the same gene set was mapped to the KEGG database (http://www.genome.jp/kegg/) to systematically explore involved biological pathways. Pathway enrichment significance was determined using both BH adjusted P-value (padj < 0.05) and FDR (FDR < 0.05) thresholds. The gene ratio (number of input genes in a pathway divided by total genes in that pathway) was calculated to evaluate enrichment intensity. Gene Set Enrichment Analysis (GSEA) and the generation of enrichment plots were performed using the OmicStudio platform (https://www.omicstudio.cn) based on R software (version 4.1.3).

qRT-PCR

To validate the transcriptomic findings, the expression levels of 9 candidate mRNAs were quantified by qRT-PCR in 12 independent patient samples. Total RNA was extracted from NECs using TRIzol Reagent (Invitrogen) according to the manufacturer’s instructions. RNA quality and the absence of genomic DNA contamination were assessed using a NanoDrop One Microvolume UV–Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Reverse transcription was performed on 500 ng of total RNA using PrimeScript RT Master Mix (Takara, Shiga, Japan) in a total volume of 20 μL, following the manufacturer’s protocol. qRT-PCR was conducted using TB Green Premix Ex Taq II (Takara) on a QuantStudio 5 platform (Applied Biosystems, Waltham, MA, USA) with the following thermal profile: initial polymerase activation at 95 °C for 30 s, followed by 40 cycles of denaturation at 95 °C for 5 s and annealing/extension at 60 °C for 34 s. The primer sequences for the target mRNAs are provided in Table S1.

MeRIP-qPCR

MeRIP assays were performed using the m6A RNA Enrichment Kit (Epigentek, USA) according to the manufacturer’s instructions. Briefly, total RNA was fragmented and divided into immunoprecipitation (IP) and Input fractions. The RNA in the IP group was immunoprecipitated using a bead-bound m6A capture antibody, followed by elution and purification. Subsequently, the m6A modification levels were quantified via qPCR analysis, and the data were normalized to the corresponding Input RNA. The primers used for the MeRIP-qPCR assays are listed in Table S2.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 8 software. In qRT-PCR and MeRIP-qPCR assays, differences between two groups were compared using a two-tailed unpaired Student’s t test. Pearson’s correlation analysis was performed to evaluate the correlation between the m6A methylation or expression levels of target genes and the patients’ TNSS. p value < 0.05 was considered statistically significant (*p < 0.05; **p < 0.01; NS, not significant).

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AR

Allergic rhinitis

NECs

Nasal epithelial cells

AIT

Allergen-specific immunotherapy

m6A

N6-methyladenosine

MeRIP-Seq

Methylated RNA immunoprecipitation sequencing

RNA-Seq

RNA sequencing

qRT–PCR

Quantitative real-time PCR

5’UTR

5’ Untranslated region

3’UTR

3’ Untranslated region

CDS

Coding sequences

FC

Fold change

DMGs

Differentially m6A-methylated genes

GO

Gene ontology

KEGG

Kyoto encyclopedia of genes and genomes

CC

Cellular component

BP

Biological process

MF

Molecular function

DEGs

Differentially expressed genes

DMGs

Differentially methylated genes

DMEGs

Differentially methylated and expressed genes

Author contributions

Shimin Lai: Conceptualization, Data curation, Project administration, Visualization and Writing original draft. Lei Yu and Lin Sun: Formal analysis and Investigation. Hongyuan Zheng: Data curation. Tong Lu and Zhengqi Li: Methodology, Resources and Validation. Changhui Chen: Validation. Yi Wei: Funding acquisition, review and editing of the manuscript. Weiping Wen: Funding acquisition, resources and Project administration. All authors read and approved the final version of the manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (NSFC) grants 82020108009 (W.P.W.), 82171766 (W.Y.), Guangdong Natural Science Foundation of China grants 2023B1111040004 (W.P.W.), Key Clinical Technique of Guangzhou 2023P-ZD06 (W.P.W.), Guangzhou Science and Technology Project of China grant 2024A04J10019 (W.Y.)

Data availability

Primary data supporting this investigation are preserved in the controlled access database of the Department of Otorhinolaryngology, The First Affiliated Hospital of Sun Yat-sen University. As the clinical datasets were provided under a formal data use agreement specific to this study, public access is restricted. Data inquiries should be directed to the corresponding author and will be addressed upon obtaining permission from the Department of Otorhinolaryngology, The First Affiliated Hospital of Sun Yat-sen University.

Declarations

Ethics approval and consent to participate

All procedures related to patients were carried out under the approval of the Internal Review and the Ethics committee of First Affiliated Hospital of Sun Yat-sen University (Guangdong, China) (Approval No. 2022–354). Written informed consent was obtained from all patients.

Consent for publication

All authors agreed to publish this manuscript.

Competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.