The lactylation-immune regulatory axis: a potential therapeutic target for migraine prevention and treatment

Abstract

Background: Migraine is a debilitating neurological disorder. Emerging evidence suggests that metabolic dysregulation and immune dysfunction contribute to migraine pathogenesis while the molecular mechanisms linking these processes remain unclear. Lactylation may serve as a crucial integrator of metabolic and immune signals in migraine.

Methods: We performed a comprehensive multi-omics Mendelian randomization study integrating DNA methylation, gene expression, and protein abundance data with genome-wide association studies on migraine. Summary-data-based Mendelian randomization, Bayesian colocalization, and two-sample MR analyses were conducted to identify lactylation-related genes causally associated with migraine. We further explored immune cell mediation using genetic data from 731 immune phenotypes and validated findings through single-cell RNA sequencing of peripheral blood mononuclear cells from migraine patients and controls.

Results: EP300, SIRT1, and SLC16A1 were identified as key regulators of migraine susceptibility across methylation, expression, and protein levels. EP300 and SLC16A1 were associated with increased migraine risk, while SIRT1 conferred a protective effect. Mediation analyses revealed that genetic effects of these genes were partially transmitted through specific immune cell subsets, particularly B cells and natural killer T cells. Single-cell transcriptomic profiling further demonstrated EP300 upregulation in B cells and T cells of migraine patients. These findings support a novel “lactylation–immune mediation–migraine axis” linking metabolic and immune dysregulation to migraine pathogenesis.

Conclusion: This integrative multi-omics analysis uncovers lactylation-related genes as causal drivers of migraine through immunometabolic pathways. Targeting lactylation-regulated metabolic and immune mechanisms may offer novel precision therapeutic strategies for migraine, particularly in patients with inflammatory or metabolic endophenotypes.

Supplementary Information: The online version contains supplementary material available at 10.1186/s10194-025-02075-3.

Keywords: Immune mediation; Lactylation; Mendelian randomization; Migraine; Multi-omics; Single-cell RNA sequencing.

Fig. 1

Schematic overview of the analytical workflow. Genetic instruments (P < 5×10⁻⁸) for lactylation-related genes were identified from publicly available methylation (mQTL), expression (eQTL), and protein (pQTL) quantitative trait loci datasets following functional normalization and quality control. Two-sample Mendelian randomization (MR) analyses were conducted using inverse-variance weighted (IVW), MR-Egger regression, weighted median, and MR-PRESSO methods. Sensitivity analyses were performed to assess the robustness and consistency of causal estimates. Summary-data-based Mendelian randomization (SMR) results were evaluated using the HEIDI test and colocalization analysis, with PPH4 > 0.60 indicating evidence for a shared causal variant. Genes with consistent associations across at least two omics layers were prioritized for immune cell mediation analysis using GWAS summary statistics of 731 immune phenotypes. Finally, single-cell RNA sequencing was conducted to examine cell-type–specific expression of candidate genes in peripheral blood mononuclear cells (PBMCs) from migraine patients and healthy controls. Abbreviations: mQTL, methylation quantitative trait loci; eQTL, expression quantitative trait loci; pQTL, protein quantitative trait loci; GWAS, genome-wide association study; MR, Mendelian randomization; SMR, summary-data-based Mendelian randomization; HEIDI, heterogeneity in dependent instruments; PPH4, posterior probability for a shared causal variant; PBMCs, peripheral blood mononuclear cells

Fig. 2

Conceptual framework for evaluating immune-mediated pathways linking lactylation-related genes to migraine risk. A The total causal effect of lactylation-related genes on migraine (path c) was estimated using Mendelian randomization (MR). Reverse causality (path d) was assessed in sensitivity analyses. B A two-step MR-based mediation model was implemented to partition the total effect into an indirect effect through immune cell traits (paths a and b) and a direct effect independent of immune mediation (path c). Note: Path c in panel (A) represents the total effect, while in panel (B) it denotes the direct effect independent of immune mediation. Abbreviations: MR, Mendelian randomization

Fig. 3

SMR and colocalization results for CpG methylation sites associated with migraine. Summary-data-based Mendelian randomization (SMR) and colocalization analysis of DNA methylation (CpG sites) in relation to migraine risk. A Results from the primary analysis using the FinnGen cohort as the outcome dataset. B Replication results from the GWAS Catalog dataset (GCST90000016). Each dot represents an SMR association between a CpG probe and migraine risk. Horizontal lines denote 95% confidence intervals (CIs). Only associations passing the HEIDI test (P_HEIDI > 0.01) and showing strong colocalization (PPH4 > 0.60) are shown. Abbreviations: SMR, summary-data-based Mendelian randomization; PPH4, posterior probability for a shared causal variant; OR, odds ratio; CI, confidence interval

Fig. 4

SMR colocalisation of lactylation-related gene expression with migraine. Forest plots display odds ratios (OR) and 95 % confidence intervals (CI) from summary-data-based Mendelian randomisation (SMR) using cis-eQTL instruments. A Primary analysis using gene expression quantitative trait loci (eQTL) and migraine GWAS data from the FinnGen cohort. B Replication results using the GWAS Catalog dataset (GCST90000016). Only genes with nominal significance (P < 0.05) are shown. Horizontal bars represent 95% confidence intervals (CIs) for odds ratios (ORs). Genes with strong colocalization support (PPH4 > 0.60) are considered to have evidence of shared causal variants. Abbreviations: CI, confidence interval; eQTL, expression quantitative-trait locus; OR, odds ratio; PPH4, posterior probability of shared causal variant; SMR, summary-data-based Mendelian randomisation

Fig. 5

Two-sample MR analysis of lactylation-related gene expression and migraine risk. Forest plots displaying inverse-variance weighted Mendelian randomization (IVW-MR) estimates for the association between genetically predicted expression of EP300, SIRT1, and SLC16A1 and migraine risk. A Primary analysis using eQTL exposure data and migraine GWAS summary statistics from the FinnGen cohort. B Replication analysis using an independent GWAS dataset (GCST90000016). The inverse-variance weighted (IVW) method was used to estimate causal effects. Odds ratios (ORs) and 95% confidence intervals (CIs) represent the effect of a one–standard-deviation increase in gene expression on migraine risk. All shown associations reached nominal significance (P < 0.05). Abbreviations: MR, Mendelian randomization; IVW, inverse-variance weighted; OR, odds ratio; CI, confidence interval; eQTL, expression quantitative trait loci; GWAS, genome-wide association study

Fig. 6

Forest plot of Mendelian randomization associations between immune cell traits and migraine risk. Forest plot displaying representative 21 immune cell traits with significant associations with migraine risk (P < 0.05) based on inverse-variance weighted (IVW) Mendelian randomization analysis. Traits were selected from a panel of 731 immune phenotypes and include markers from B cells, TBNK cells, conventional dendritic cells (cDCs), monocytes, and maturation-stage T cells. Each point estimate represents the odds ratio (OR) per genetically predicted increase in the immune trait, with 95% confidence intervals shown. Abbreviations: MR, Mendelian randomization; IVW, inverse-variance weighted; OR, odds ratio; CI, confidence interval; AC, absolute count; cDC, conventional dendritic cell; TBNK, T, B, and NK cells; EM, effector memory; TD, terminally differentiated

Fig. 7

t-SNE projection of unsupervised PBMC clusters. Single-cell transcriptomes from three migraine patients and two healthy controls were integrated and clustered with the Seurat shared-nearest-neighbor algorithm, yielding 13 transcriptionally distinct groups (labelled 0– 12). Each dot denotes one peripheral blood mononuclear cell (PBMC) and is coloured by its cluster identity. These clusters were subsequently used for cell-type annotation and differential expression analysis. The x- and y-axes correspond to the first and second t-SNE dimensions. Abbreviations: t-SNE, t-distributed stochastic neighbour embedding; PBMC, peripheral blood mononuclear cell

Fig. 8

t-SNE projection of annotated immune cell types in peripheral blood mononuclear cells. t-SNE plot visualizes immune cell clusters derived from single-cell RNA sequencing of peripheral blood mononuclear cells (PBMCs) pooled from migraine patients and healthy controls. Cells are grouped by transcriptional similarity and annotated as T cells, CD4⁺ T cells, natural killer (NK) cells, B cells, and monocytes based on reference-based classification. Each dot represents one cell, colored according to its assigned identity. Cells lacking confident annotation are labeled as “NA.” The two axes correspond to the first and second t-SNE dimensions. Abbreviations: t-SNE, t-distributed stochastic neighbor embedding; PBMCs, peripheral blood mononuclear cells; NK, natural killer; NA, not annotated

Fig. 9

Single-cell t-SNE maps of EP300, SIRT1 and SLC16A1 expression in peripheral blood mononuclear cells. t-SNE plots depict log-normalised transcript counts for EP300, SIRT1 and SLC16A1 in peripheral blood mononuclear cells (PBMCs) pooled from migraine patients and healthy controls. Each dot represents one cell; colour denotes expression intensity from low (blue, 0) to high (red, ≥ 3). EP300 is detected in multiple clusters, whereas SIRT1 and SLC16A1 show uniformly low expression levels. Abbreviations: t-SNE, t-distributed stochastic neighbour embedding; PBMCs, peripheral blood mononuclear cells

Fig. 10

Single-cell expression of EP300 in major peripheral immune cell subsets from migraine patients and healthy controls. Violin plots display log-normalized expression levels of EP300 in five immune cell types: A B cells, B T cells, C CD4⁺ T cells, D natural killer (NK) cells, and E monocytes. Expression data are shown separately for cells from migraine patients and healthy controls. Higher EP300 expression is observed in B cells and T cells from migraine samples, while expression in monocytes appears lower in migraine patients compared to controls. CD4⁺ T cells show comparable levels across groups. EP300 expression is minimal or absent in NK cells in both conditions. Abbreviations: NK, natural killer; MI, migraine