Unveiling the genetic overlap and causal links between gastroesophageal reflux disease and asthma

Abstract

Background:

Gastroesophageal reflux disease (GERD) and asthma are commonly co-occurring conditions, with shared genetic factors identified. However, the specific loci and the influence of common genetic architecture remain undefined.

Methods:

We obtained genome-wide association study (GWAS) summary statistics for GERD (71 522 cases and 261 079 controls) and asthma (56 167 cases and 352 255 controls). Using linkage disequilibrium score regression (LDSC), we assessed genetic correlations between GERD and asthma. Bidirectional Mendelian randomization (MR) was performed to investigate potential causal relationships, followed by cross-trait GWAS meta-analysis and colocalization analysis to identify shared risk loci. Additionally, summary-data-based MR and transcriptome-wide association study were conducted to pinpoint common functional genes. Finally, we analyzed gene expression profiles in both healthy individuals and GERD patients using esophageal single-cell RNA sequencing (scRNA-seq) data.

Results:

We identified a significant genetic correlation between GERD and asthma (r_g = 0.37, P = 6.19 × 10^–38) and a significant causal effect of GERD on asthma [odds ratio (OR) = 1.22, P = 1.54 × 10⁻⁵]. Cross-trait meta-analyses revealed 56 shared risk loci between GERD and asthma, including 51 loci that were newly identified. Three loci (rs61937247, rs7960225, and rs769670) exhibited evidence of colocalization. Gene-level analyses pinpointed three novel shared genes (RBM6, SUOX, and MPHOSPH9) between GERD and asthma. scRNA-seq analysis uncovered heightened expression of these genes in immune cells of patients diagnosed with GERD.

Conclusion:

Our study has discovered novel shared genetic loci and candidate genes between GERD and asthma, providing further insights into the genetic susceptibility of comorbidity and potential mechanisms of the two diseases.

Introduction

Gastroesophageal reflux disease (GERD), characterized by the regurgitation of gastric contents into the esophagus, leads to a range of esophageal and extraesophageal symptoms^[1,2]. Asthma is a chronic respiratory condition marked by airway inflammation and hyperreactivity, affecting up to 9% of the global population, and is becoming more prevalent^[3]. Evidence for reciprocal comorbidity of GERD and asthma has grown in recent years^[4,5].

Epidemiological evidence shows that individuals with asthma have a 1.6-fold higher prevalence of GERD symptoms compared to controls, while individuals with GERD have a 1.2-fold higher prevalence of asthma compared to controls^[6]. A recent Mendelian randomization (MR) study demonstrated a significant bidirectional association between asthma and GERD, revealing that asthma increases the risk of GERD by 6%, while GERD increases the risk of asthma by 21%^[7]. Prior twin studies found that the phenotypic correlation for GERD and asthma was moderate at 0.14, and just over half of this was attributed to a shared genetic architecture^[8,9]. While accumulating evidence suggests shared genetic susceptibility between GERD and asthma, the underlying biological connections remain largely unexplored.

Notably, twin cohort studies have revealed elevated comorbidity rates potentially attributable to shared genetic and early environmental factors. However, further investigation within a general population cohort is warranted to identify broader comorbidity patterns that reflect diverse backgrounds and environmental influences. Elucidating the complex biological mechanisms underlying GERD and asthma comorbidity is crucial for understanding their pathogenesis and informing targeted therapeutic strategies.

Our study aims to pioneer the analysis of the shared genetic architecture between GERD and asthma by leveraging large-scale genome-wide association studies (GWASs) and advanced statistical methodologies. We performed cross-trait GWAS meta-analysis and colocalization analysis to identify and pinpoint significant and independent loci, and conducted summary-data-based Mendelian randomization (SMR) and transcriptome-wide association study (TWAS) to identify common functional genes. Finally, we analyzed gene expression profiles in both healthy individuals and GERD patients using esophageal single-cell RNA sequencing (scRNA-seq) data.

The work has been reported in line with the TITAN criteria^[10].

Methods

Datasets

We utilized summary-level data of GERD and asthma from two large GWAS conducted in European ancestry populations to investigate. For GERD, the genetic instrumental variables (IVs) were derived from the two latest and largest population-level GWAS meta-analyses on GERD (encompassing 71 522 cases and 261 079 control subjects)^[11]. BOLT-LMM algorithm (v2.3) was applied with recruitment age, genetic sex, and the first 10 principal components included as covariates in this dataset. For asthma, the genetic IVs were obtained from GWASs in the UK Biobank, which employed a broad definition of asthma (comprising 56 167 cases and 352 255 control subjects)^[12]. Asthma cases were restricted to UK Biobank participants without self-reported or clinically documented chronic lung diseases (e.g., COPD, emphysema), allergic conditions, or smoking history to minimize phenotypic overlap. The dataset using Scalable and Accurate Implementation of GEneralized mixed model (SAIGE), age, sex, and the first 20 ancestry-based principal components were included in the null logistic mixed model. Regarding the validation dataset, we utilized GWAS summary statistics on GERD (129 080 European ancestry cases and 473 524 European ancestry controls)^[13] and asthma data of European ancestry origin (19 954 cases and 107 715 controls)^[14].

The Genotype-Tissue Expression (GTEx)^[15] project provides tissue-specific gene expression data and regulation information. The data used in this study were obtained from the GTEx Portal. We utilized GTEx data from whole blood, thyroid, esophagus muscularis, esophagus gastroesophageal junction, stomach, spleen, small intestine terminal ileum, lung, artery coronary, ovary, uterus, prostate, and vagina tissues in our study^[16]. These tissue samples are used to explore the expression of corresponding tissue-related genes.

HIGHLIGHTS

• Our study identified a significant genetic correlation between GERD and asthma (LDSC: r_g = 0.37, P = 6.19 × 10⁻³⁸; GNOVA: r_g = 0.55, P = 4.49 × 10⁻⁵²).

• MR analyses suggest that GERD increases the risk of asthma (P_IVW = 1.54 × 10⁻⁵), but not vice versa.

• We discovered 56 genome-wide significant loci for GERD and asthma, including 51 novel loci, and identified RBM6, SUOX, and MPHOSPH9 as key novel genes.

• scRNA-seq indicated increased gene expression in immune cells of GERD patients, suggesting a mechanistic link through inflammation.

• The significant genetic correlation and causality between GERD and asthma highlights the need to delve into underlying mechanisms and reiterate the importance for clinicians to consider the co-occurrence of these conditions.

• Identifying shared SNPs and genes provides insights into the genetic architecture underlying both diseases, offering opportunities for novel therapeutic targets and risk prediction models.

• The discovery of cell-specific gene expression patterns stimulates further studies on their genetic basis and might provide an innovative research direction for future therapeutic strategies.

scRNA-seq data of proximal and distal esophageal biopsies from four patients diagnosed with GERD and six healthy donors were obtained from the Gene Expression Omnibus repository (GEO: GSE218607)^[17]. The overall study design is shown in Figure 1.

Figure 1.

The overall study design.

Genome-wide genetic correlation analysis

To explore the shared genetic basis of GERD and asthma, we employed linkage disequilibrium score regression (LDSC)^[18]. Using LDSC (Python 2.7), we assessed the genetic correlation among various subtypes. Single nucleotide polymorphisms (SNPs) that did not match the reference panel [with an INFO score ≤ 0.9 or MAF (minor allele frequency) ≤ 0.01] were omitted based on the linkage disequilibrium (LD) scores derived from the 1000 Genomes project phase III^[19]. Subsequently, bivariate LDSC was conducted to evaluate the genetic correlations between GERD and asthma. A genetic correlation with a P-value less than 0.05 was considered significant^[20,21].

Additionally, we used the Genetic covariance analyzer (GNOVA)^[22] to compute the genetic correlation and the SNP-based heritability for GERD and asthma. GNOVA explores genetic covariance using all genetic variants shared between two diseases. We then computed the genetic correlation based on the genetic covariance and variant heritability. Furthermore, we corrected the sample overlap of GWAS summary statistics. A threshold of P < 0.05 was regarded as strong evidence for MAF-stratified genetic correlation.

Local genetic correlation analysis

To measure the local genetic correlations (i.e., the genetic correlation between traits attributable to their shared genetic variance within a defined genomic region) between GERD and asthma, the Heritability Estimation from Summary Statistics (ρ-HESS)^[23] was applied. Relying on genomic references derived from genomes, the ρ-HESS method was employed to calculate the local SNP heritability for each trait and the genetic covariance between traits. The local genetic correlation estimates are then derived from the local single-trait SNP heritability and the local cross-trait genetic covariance estimates. The algorithm divided the entire genome into 1699 regions based on the LD patterns in the European population and quantified the correlations between traits due to genetic variation confined to a specific region. Subsequently, a Bonferroni correction was applied to adjust for multiple testing^[23].

MR analysis

To scrutinize the potential causal links between GERD and asthma, we performed bidirectional MR using “TwoSampleMR” R package^[24]. In MR analysis, we employed genetic variants strongly associated with the exposure as IVs to examine potential causal relationships between the exposure and the outcome^[25]. The MR estimations remain robust against confounding and reverse causal biases, as the genetic variants are randomly assigned during conception. To ensure validity, three pivotal assumptions must be satisfied: first, the genetic variant demonstrates a robust association with the exposure; second, it exhibits no association with any confounder affecting the exposure–outcome relationship; and third, its influence on the outcome is solely mediated through its relationship with the exposure^[26]. The inverse variance weighting (IVW)^[27] aggregates the estimate from each genetic variant and computes a precise causal estimation under the assumption that all genetic variants are valid or are invalid in such a way that the overall pleiotropy is balanced to be zero. MR Egger^[28] regression is employed to identify and correct for bias due to directional pleiotropy^[28]. The weighted median (WME)^[29] approach delivers a consistent estimate of the causal effect even when up to 50% of the estimate of genetic variants is invalid^[30]. The weighted mode^[31] categorizes the SNPs into clusters based on their estimated causal effects, and assesses evidence for causality using only the largest set of SNPs, which essentially eases the assumptions of MR and can pinpoint the true effect even if a majority of instruments are invalid SNPs^[31]. Additionally, the MR pleiotropy residual sum and outlier (MR-PRESSO)^[32] test was utilized to detect pleiotropic outlier variants. Subsequently, the four MR methods mentioned above were re-applied to estimate the causal effect, after excluding outlier SNPs in MR-PRESSO analyses. IVs for MR analyses were selected from GWAS summary statistics using three criteria: (1) genome-wide significance (P < 5 × 10⁻⁸); (2) linkage disequilibrium clumping via PLINK (r² < 0.001, 10 000 kb window) to ensure SNP independence; and (3) an F-statistic ≥10 (calculated as F = β²/SE²) to mitigate weak instrument bias, where β is the effect size and SE is the standard error of the SNP-exposure association. Causal Analysis Using Summary Effect (CAUSE)^[33] provides a test-statistic, an estimate of the causal effect, and a variant-level summary indicating how each variant contributes to the overall test, and which variants are likely to be acting through a shared factor. We executed CAUSE analysis utilizing the “cause” R package, version 1.2.0. This is the only method capable of distinguishing causality from both correlated and uncorrelated pleiotropy.

Cross-trait GWAS meta-analysis

To identify risk SNPs linked with combined phenotypes (GERD–asthma), we carried out a cross-trait meta-analysis of GWAS summary statistics utilizing Cross Phenotype Association (CPASSOC)^[34]. We performed a pairwise cross-trait meta-analysis using CPASSOC, accommodating variations in the degree of heritability for the two phenotypes. CPASSOC presumes the existence of heterogeneous effects across traits and computes the cross-trait statistic heterogeneity and P-value through a sample size weighted meta-analysis of GWAS summary data. Ultimately, the significant SNPs were characterized as SNPs that were significantly associated with both phenotypes. To identify the significant and independent loci, we utilized the threshold P_CPASSOC < 5 × 10^–8 and P_GERD < 1 × 10^–3 and P_Asthma < 1 × 10^–3 and the “clumping” function of PLINK (settings: clump_p1 = 5 × 10^–8, clump_p2 = 1 × 10^–5, clump_r2 = 0.2, and clump_kb = 500)^[35]. Functional enrichment analysis of protein-coding genes shared between GERD-asthma loci (excluding major histocompatibility complex variants) was performed using clusterProfiler v4.0^[36]. Gene Ontology biological processes and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were assessed with Benjamini-Hochberg false discovery rate (FDR) correction. Significantly enriched terms were defined as those with FDR <0.05.

Colocalization analysis

The pleiotropy of SNPs could elucidate the genetic correlation between GERD and asthma. Therefore, colocalization analysis was conducted to authenticate the shared genetic variants between GERD and asthma. We employed the coloc.abf function from the Coloc^[37] to scrutinize SNPs obtained from cross-trait GWAS meta-analysis for colocalization, either through shared common SNPs or unique SNPs within the same gene. Coloc utilizes a Bayesian algorithm to generate posterior probabilities for five mutually exclusive hypotheses concerning the sharing of causal variants in a genomic region^[38]. A locus was considered colocalized if PPH4 exceeded 0.75. Then we used “LocusZoom” for precise localization of causal SNPs^[39].

Shared functional genes

SMR^[40] was employed to pinpoint potential functional genes that underlie the statistical correlations for GERD and asthma. SMR is a technique that amalgamates summary statistics from GWAS and expression quantitative trait loci (eQTL) studies^[41] within the MR framework to probe for an association between gene expression and a target phenotype. SMR was executed in various tissues that exhibited significant SNP heritability enrichment of GERD and asthma. Utilizing genome-wide significant SNPs as IVs, the heterogeneity in dependent instruments (HEIDI) test was carried out to evaluate the existence of linkage in the observed association. SMR incorporates the HEIDI-outlier test to differentiate causality or pleiotropy from linkage^[38]. We used the enriched tissues from the result of functional mapping and gene annotation (FUMA) and disease-related tissues from GTEx v8 (Genotype-Tissue Expression version 8) and cis-eQTL data of whole blood from the eQTLGen consortium. Significant shared functional genes between GERD and asthma were characterized as functional genes that surpassed Bonferroni correction and the HEIDI-outlier test (P > 0.05, Nsnp > 10).

Transcriptome-wide association analysis

FUSION software was employed to scrutinize tissue-specific TWAS data derived from the GWAS summary data^[42]. A TWAS utilized pre-calculated gene expression weights in tandem with GWAS summary statistics to ascertain the correlations between genes and diseases. Cortical RNA sequence reference panels from the GTEx Consortium were amalgamated with GWAS summary statistics for TWAS^[15]. We prioritized the trait-related tissues by FUMA; thus, we prepared the eQTL data of whole blood, stomach, spleen, prostate, ovary, lung, thyroid, and esophagus-related tissues from GTEx v8. We applied the Bonferroni correction to identify significant expression–trait associations and selected genes that overlapped between GERD and asthma in the same tissue. Genes significantly associated with GERD-asthma comorbidity in SMR/TWAS analyses were validated at the protein level via LDSC regression using UK Biobank Pharma Proteomics Project (UKB-PPP) protein quantitative trait loci (pQTLs) data^[43], assessing genetic correlations between pQTLs and GERD/asthma risk.

scRNA-seq analysis

The raw count data were imported using the Read10X function in Seurat (version 5.1.0) for subsequent analysis^[44]. A Seurat object was created to store the scRNA-seq data. Normalization of the raw count data was performed using the “LogNormalize” method with a scale factor of 10 000 to account for variations in sequencing depth. Highly variable genes were identified using the “vst” method with a target of 2000 genes. This step enhances the downstream analyses by focusing on the most informative genes. All genes were scaled to ensure that the mean expression across cells is zero and the variance is one. Genes were retained if expressed in ≥3 cells, and cells were included with total gene counts ≥200. Individual datasets were normalized, scaled, and analyzed for variable features using the SCTransform workflow. Principal component analysis (PCA) was conducted using the highly variable genes to reduce dimensionality and capture the primary sources of variation in the dataset. Based on the PCA results, the cells were clustered using the nearest neighbor graph. Uniform Manifold Approximation and Projection was employed for visualization of the high-dimensional scRNA-seq data. Marker genes for each cluster were identified using the following criteria: a minimum fraction of expressing cells at 25% and a log fold change threshold of 0.25. This methodology ensures the identification of the most significant differentially expressed genes across clusters, facilitating further biological interpretation and functional analysis^[45]. The sc2GWAS provides trait-relevant information at single-cell resolution by effectively integrating GWAS summary statistics with scRNA-seq profiles^[46]. We leveraged sc2GWAS to integrate GWAS and scRNA-seq data^[47], exploring genetic links between specific cells and disease susceptibility.

The work has been reported in line with the Standards for the Reporting of Diagnostic accuracy studies criteria^[48].

No artificial intelligence tools were used in the research and manuscript development.

Results

Genetic correlation between GERD and asthma

To estimate the liability-scale SNP heritability for GERD and asthma, we used the LD score regression^[49] and the baseline-LD model^[50]. The SNP-based liability-scale heritability (h²) for GERD and asthma, as estimated using bivariate LDSC, was relatively low. The genetic correlation between GERD and asthma(r_g = 0.37, P = 6.19 × 10⁻³⁸) was significant. The results of bivariate LDSC show that GERD has a positive correlation with asthma (Table 1). To get a robust result, we conducted GNOVA and found a consistent result between GERD and asthma after correcting for sample overlap (Table 1). The ρ-HESS method^[50] was adopted to determine local genetic correlations across the genome between GERD and asthma (Supplementary Digital Content Table 1, available at: http://links.lww.com/JS9/F11), leading by 19q (chr19: 18409862.. 19877471) for GERD with asthma (P = 3.05 × 10⁻¹⁷).

Table 1.

Heritability and genetic correlation between GERD and asthma

	Method	GERD	asthma
Heritability h2	LDSC	0.0678	0.0472
Heritability h2	GNOVA	0.0769	0.0458
Heritability h2	ρ-HESS	0.0880	0.0423
Genetic correlation rg	LDSC	0.3688
Genetic correlation rg	GNOVA	0.5527
Genetic correlation rg	ρ-HESS	0.6020
Genetic correlation rg	MixeR	0.2200

GERD, gastroesophageal reflux disease; LDSC, linkage disequilibrium score regression; GNOVA, Genetic covariance analyzer; ρ-HESS: Heritability Estimator from Summary Statistics.

Causal association analysis via MR

To investigate the potential causal relationship between GERD and asthma, we performed a bidirectional MR study. All SNPs were strong instruments in the MR analysis. Various bidirectional MR methods were employed to ensure the robustness of the results. For MR analysis of GERD and asthma, we identified 18 GERD-associated SNPs and 67 asthma-associated SNPs in the GWAS summary statistics that were sufficiently strong independent genetic IVs (Fig. 1). On one hand, we found a strong causal effect estimate of genetically determined GERD on increased risk of asthma [OR, 1.22; 95% confidence interval (CI), 1.11-1.33; P_IVW = 1.54 × 10⁻⁵; Q_IVW = 0.56]. On the other hand, we detected a small but significant effect estimate of genetically determined asthma on increased risk of GERD (OR, 1.04; 95% CI, 1.01-1.07; P_IVW = 5.24 × 10⁻³; Q_IVW = 0.07). Cochran’s Q P-value reflects the probability of observing the measured heterogeneity among IVs estimates under the null hypothesis of no true heterogeneity. The Cochran’s Q-test in the IVW model and the MR Egger model indicate no heterogeneity of the selected IVs (Q_IVW > 0.05). The P-value of the MR Egger intercept test is greater than 0.05, indicating a lower likelihood of horizontal pleiotropy in causal estimation. Our findings further validated the causal effect of GERD on asthma, with estimates remaining directionally consistent across the IVW, WME, weighted mode, and MR Egger approaches (Fig. 2).

Figure 2.

The results of MR for the relationship between GERD and asthma. (A) Forest map of MR analysis results with GERD as exposure and asthma as outcome; (B) Forest map of MR analysis results with asthma as exposure and GERD as outcome.

To distinguish causality from pleiotropy, we further utilized CAUSE^[33]. Our findings indicated consistent evidence for a causal effect of GERD on asthma (P = 0.02). The low q values in the causal model (q = 0.04-0.09), representing the fraction of variants exhibiting correlated horizontal pleiotropy, implied restricted horizontal pleiotropy (Table 2). We examined the presence of potential outlier variants that might significantly influence the exposure but not the outcome by generating a scatter plot for the CAUSE test statistics on the causal association between GERD and asthma (Supplementary Digital Content Table 2, available at: http://links.lww.com/JS9/F11). Although CAUSE is not inherently sensitive to outliers, the presence of outlier variants could argue against the causal model. Nevertheless, we detected no outlier variants. As our CAUSE analyses revealed low shared model q values (q = 0.04), this might imply a high degree of causality between the two traits. The findings were further supported by CAUSE analysis and validated through multiple methods and replication datasets (Supplementary Digital Content Table 3, available at: http://links.lww.com/JS9/F11).

Table 2.

Results for MR analyses using the CAUSE method

Direction	Median causal effect (95% CI)	Median q (CI)	P causal vs sharing
GERD → asthma	0.18 (0.07, 0.28)	0.04 (0, 0.26)	0.02
Sharing model better fit for the data
Direction	Median causal effect (95% CI)	Median q (CI)	P causal vs sharing
Asthma → GERD	0.13 (−0.22, 0.50)	0.09 (0, 0.39)	0.06

GERD, gastroesophageal reflux disease; OR, odds ratio; CI, confidence interval.

Cross-phenotype GWAS meta-analysis identifies shared SNPs

We next performed cross-trait meta-analyses using CPASSOC to identify risk SNPs underlying the joint phenotypes of GERD and asthma. After excluding SNPs that were significant in the single-trait GWAS of GERD or asthma or were in LD (LD r² ≥ 0.02) with any previously reported significant SNPs, a total of 56 shared independent SNPs reached genome-wide significance in CPASSOC with the joint phenotype GERD-asthma (P_CPASSOC < 5 × 10⁻⁸ and P_GERD < 1 × 10⁻³ and P_Asthma < 1 × 10⁻³) (Supplementary Digital Content Table 4, available at: http://links.lww.com/JS9/F11).

Colocalization analysis of shared genetic variants

To determine whether the genetic variants driving the association in two traits are the same, we performed colocalization analysis. Among the 56 identified shared independent SNPs identified by CPASSOC, 51 loci were newly discovered in our research. Among the 51 novel pleiotropic loci, three loci (rs61937247, rs7960225, and rs769670) showed evidence of colocalization (PPH4 > 0.75) (Supplementary Digital Content Table 4, available at: http://links.lww.com/JS9/F11). The results of their precise localization are shown in Figure 3. Functional enrichment analysis of protein-coding genes mapped from shared loci revealed nominally significant associations with RNA polymerase II regulation and NF-κB signaling pathways (all P < 0.01), though these did not survive multiple testing correction, potentially reflecting limited statistical power from the modest number of overlapping genes (n = 23) while suggesting biologically plausible mechanisms at the epithelial-immune interface^[51,52] (Supplementary Digital Content Tables 5–6, available at: http://links.lww.com/JS9/F11).

Figure 3.

Results of precise localization. LocusZoom plots showing colocalization between GERD and asthma. The fine-mapped variants are shown with their rsID.

Identification of shared functional genes through integrative genomic analysis methods

To identify the identification of potential functional genes underlying GERD and asthma, we employed SMR to jointly analyze GWAS summary data for both conditions, along with eQTL summary data from GTEx (tissue shown SNP heritability enrichment in both GERD and asthma). We selected the eQTL summary data of phenotypically related tissues: whole blood, thyroid, esophagus muscularis, esophagus mucosa, esophagus gastroesophageal junction, brain hippocampus, stomach, spleen, small intestine terminal ileum, lung, coronary artery, ovary, uterus, prostate, vagina, and testis, then performed SMR analysis on these tissues in relation to the two diseases. The SMR analysis identified three significant genes (RBM6, SUOX, and MPHOSPH9) that were associated with both GERD and asthma, all of which passed the HEIDI-outlier test (Supplementary Digital Content Table 7, available at: http://links.lww.com/JS9/F11). RBM6 is enriched in nine tissues, SUOX is enriched in eight tissues, and MPHOSPH9 is enriched in two tissues. All three genes are enriched in the esophagus muscularis and thyroid. Among them, SUOX is shared between GERD and asthma in the esophagus muscularis (P_SMR = 1.54 × 10⁻⁶, P_HEIDI = 0.27; P_SMR = 1.07 × 10⁻¹⁰, P_HEIDI = 6.77 × 10⁻²) and the thyroid (P_SMR = 1.29 × 10⁻⁶, P_HEIDI = 2.24 × 10⁻¹; P_SMR = 3.31 × 10⁻¹², P_HEIDI = 8.58 × 10⁻²).

Results from tissue-specific TWAS revealed gene-level genetic overlap; the TWAS P-values listed in the table have all passed Bonferroni correction (Supplementary Digital Content Table 8, available at: http://links.lww.com/JS9/F11). After multiple corrections, three genes were shared by GERD and asthma. All significant genes obtained in the results were validated either through validation datasets or by two methodological validations. The results led by SUOX (P_GERD = 3.43 × 10⁻⁷, Z_GERD = 5.10; P_asthma = 2.92 × 10⁻¹⁴, Z_asthma = −7.60) shared by GERD and asthma in the esophagus muscularis and (P_GERD = 3.43 × 10⁻⁷, Z_GERD = 5.10; P_asthma = 2.92 × 10⁻¹⁴, Z_asthma = −7.60) in the thyroid. A notable highlight from our findings is the highest prevalence of overlapping genes between thyroid and esophagus muscularis tissues, encompassing 13 positive genes across both tissues. In addition, we performed pQTLs-LDSC to assess genetic correlations between SUOX and the risk of GERD or asthma. Results showed significant genetic correlations: SUOX pQTLs were positively correlated with GERD (r_g = 0.30, P = 1.31 × 10⁻³) and asthma (r_g = 0.19, P = 8.60 × 10⁻³).

scRNA-seq analysis in healthy and GERD-affected individuals

To explore the differences in gene expression between diseased and healthy tissues in the proximal and distal esophagus, we analyzed publicly available esophageal scRNA-seq data from patients diagnosed with GERD and healthy control individuals. Each cluster was detected and identified using established markers and/or previous literature (Fig. 4). We observed a significant increase in RBM6 expression within the monocyte, fibroblast, and macrophage clusters compared to healthy individuals. Additionally, the SUOX and MPHOSPH9 genes are expressed in macrophages, monocytes, and fibroblasts in diseased individuals, but at a lower proportion and expression level compared to the RBM6 gene in these cell types. In the disease group, there was an increased presence of immune-related cell types, which may indicate an inflammatory response under disease conditions. To identify genetically anchored cell-disease associations, we implemented sc2GWAS integration. Asthma susceptibility showed the strongest genetic association with Cytotoxic CD4+ T cells (Norm Score = 3.39, P = 3 × 10⁻³). GERD pathogenesis was primarily driven by microglial cells (Norm Score = 5.52, P = 1 × 10⁻³) and monocytes (Norm Score = 5.65, P = 1 × 10⁻³) (Supplementary Digital Content Figure 1A–D, available at: http://links.lww.com/JS9/F10). These cell populations converge on the shared inflammatory networks, suggesting synergistic contributions to comorbid pathogenesis.

Figure 4.

Results of scRNA-seq analysis. Single-cell sequencing analysis revealed that genes were predominantly expressed in superficial, suprabasal, and basal cells in both healthy individuals and patients with GERD. (A) and (B) In healthy samples, gene expression was mainly in epithelial and endothelial cells. (C) and (D) In patients diagnosed with GERD, there was an increased presence of immune-related cells (monocytes, macrophages, and fibroblast clusters), indicating an inflammatory response.

Discussion

Our study aimed to investigate the significant shared loci, potential shared genes, and causal relationships between GERD and asthma. We found significant global genetic correlation between GERD and asthma and a positive causal effect of GERD on asthma. Additionally, three functional genes (RBM6, SUOX, and MPHOSPH9) are identified as associated with GERD and asthma. Finally, our scRNA-seq analysis revealed increased expression of three functional genes in immune cells from patients diagnosed with GERD, indicating a mechanistic connection between GERD-induced inflammation and asthma.

Our MR results indicated a significant causal effect estimate of genetically determined GERD on the increased risk of asthma in European populations, not vice versa. While the asthma→GERD association reached nominal significance in IVW analysis (OR = 1.04, P = 5 × 10⁻³), its small effect size further attenuates clinical significance and may be influenced by unmeasured confounders such as obesity and smoking. Methodologically, the lack of support from CAUSE and limited replication in independent datasets indicate that the statistically significant signal likely reflects residual pleiotropy or confounding rather than true causality. Methodologically, the lack of support from CAUSE and limited replication in independent datasets indicate that the statistically significant signal likely reflects residual pleiotropy or confounding rather than true causality. A recent MR study indicated no significant associations between genetically predicted adult-onset asthma and GERD^[53]. Additionally, the lack of significant regions passing multiple testing likely reflects the polygenic architecture of GERD and asthma, with shared genetic influences dispersed across many small-effect loci rather than concentrated in discrete regions^[38]. This aligns with their complex pathophysiology, involving interactions between immune regulation and epithelial barrier function^[54], modulated by polygenic variants across diverse pathways.

We found three novel loci: rs61937247, rs7960225, and rs769670, shared between GERD and asthma, which were further validated through colocalization. It has been reported that the locus rs769670 is related to the MAML3 gene. MAML3 is a pleiotropic gene associated with smoking-related traits and GERD^[55]. GERD activates the Notch signaling pathway^[56], which is also critical in asthma pathogenesis: Notch activation impairs regulatory T cell (Treg) function, promoting Th2 cytokine production and allergic asthma^[57]. We reasonably hypothesize MAML3 contributes to GERD-asthma comorbidity via Notch modulation: its association with GERD may involve Notch activation in esophageal inflammation, while Notch-mediated Treg dysfunction could drive asthmatic processes.

At the gene-tissue level, our analysis has revealed potential biological mechanisms linking GERD and asthma. SUOX is a key enzyme in sulfur amino acid metabolism; a deficiency of SUOX causes isolated sulfite oxidase deficiency^[58]. Disrupted sulfite-sulfate balance may contribute to esophageal mucosal injury and airway inflammation by compromising antioxidant defenses or modulating inflammatory pathways. As a critical enzyme in sulfite detoxification^[59,60], SUOX presents dual opportunities: enhancing its activity may mitigate sulfite-induced epithelial injury in GERD^[61], while inhibiting SUOX-mediated sulfur metabolism could dampen Th2-driven airway inflammation in asthma^[62]. Experimental validation of these proposed mechanisms is required. SUOX modulation offers dual therapeutic potential: enhancing its activity could correct sulfite-mediated oxidative damage in GERD epithelium^[59], while strategic inhibition may attenuate Th2-driven airway inflammation through sulfur metabolite regulation. RBM6 has been identified as a novel asthma-related gene in prior studies^[63]. It exhibits functional associations with macrophage biology^[55,64]. Besides, RBM6 is a pleiotropic gene implicated in GERD^[65], involved in alternative splicing and cellular stress response regulation^[66]. Mechanistic evidence shows that RBM6 interacts with MLKL to stabilize mRNA encoding adhesion molecules^[67], a process critical for maintaining epithelial barrier integrity. Targeting RBM6 presents novel opportunities to disrupt inflammatory cascades by normalizing macrophage polarization states^[68,69] and reinforcing epithelial barrier integrity via mRNA stabilization of adhesion molecules. MPHOSPH9 encodes a protein that regulates cell cycling^[70], and its transcript MPHOSPH9-OT1 has been shown to control CXCL8/IL-8 levels, key pro-inflammatory cytokines involved in epithelial barrier dysfunction. Functional studies demonstrate that MPHOSPH9-OT1 knockdown/overexpression significantly reduces/increases IL-8 mRNA and secretion, highlighting its role in maintaining cellular cytokine homeostasis essential for epithelial barrier integrity^[71]. MPHOSPH9-directed strategies hold promise for restoring mucosal homeostasis through IL-8 pathway regulation^[72], directly addressing neutrophil-mediated barrier dysfunction in both tissues.

To validate and refine the expression patterns of candidate genes at single-cell resolution, we analyzed publicly available scRNA-seq datasets. Our analysis focused on characterizing cell-type-specific expression profiles of these genes across distinct cellular subpopulations. Early-stage GERD may initiate Th1-driven pro-inflammatory responses^[73], but sustained epithelial injury triggers Th2 polarization. Prolonged reflux stimulates esophageal epithelial cells to release Th2-attracting chemokines and cytokines, recruiting and activating Th2 cells. These cells secrete type 2 cytokines, key mediators of asthmatic airway hyperresponsiveness, including eosinophilia, mucus hypersecretion, and bronchial hyperreactivity^[74]. Notably, Tregs, which suppress excessive Th2 responses via FOXP3^[75], may be functionally impaired in chronic GERD, exacerbating Th2 dominance. Our identified genetic variants, which potentially regulate chemokine signaling or Treg function, could amplify this cascade. They may reinforce esophageal-airway crosstalk by enhancing Th2 infiltration or weakening suppression, thereby linking genetic susceptibility to the shared Th2-mediated inflammatory pathway underlying both conditions.

Our findings provide novel insights into the overlapping genetic architecture between GERD and asthma, advancing our comprehension of their shared genetic predispositions. Several limitations need to be recognized. First, as the samples in our study were derived solely from Europeans, our conclusions may not be applicable to other ethnic groups, which could give rise to external validity bias. Genetic architecture, including allele frequencies and linkage disequilibrium patterns, varies across populations of different ethnicities. Comprehensive validation in diverse cohorts will be essential to delineate ethnic-specific risk profiles. Second, while integrative genomic analyses identified candidate mediators of GERD-asthma comorbidity, our findings rely on statistical associations and lack direct experimental validation. These candidates require functional testing in preclinical models to clarify their roles in shared pathogenesis.

Conclusion

Our study has discovered novel shared genetic loci and candidate genes between GERD and asthma, providing further insights into the genetic susceptibility of comorbidity and potential mechanisms of the two diseases. This enhanced understanding of the biological architecture linking GERD and asthma comorbidity can guide the future development of personalized treatment strategies and targeted therapies.

Ethical approval

Not applicable. The datasets generated and analyzed during this study are available from the corresponding author on reasonable request. The datasets used in this study were obtained from publicly available sources.

Consent

Not applicable. The datasets used in this study were obtained from publicly available sources.

Sources of funding

This work was funded by the National Natural Science Foundation of China Regional Innovation and Development Joint Foundation (U23A20408), the National Natural Science Foundation of China (82571984, 82171698, 82570637, 82170561, 81741067, and 81300279), Guangzhou Basic and Applied Basic Research Scheme - Project for Pilot Voyage (2024A04J6573), the Climbing Program for Young Scientific and Technological Talents of Southern Medical University (2024PFPYA008), the Natural Science Foundation for Distinguished Young Scholars of Guangdong Province (2021B1515020003), Foreign Distinguished Teacher Program of Guangdong Science and Technology Department (KD0120220129), Natural Science Foundation of Guangdong Province (2022A1515012081), the Climbing Program of Introduced Talents and High-level Hospital Construction Project of Guangdong Provincial People’s Hospital (DFJH201803, KJ012019099, KJ012021143, and KY012021183), Guangdong Basic and Applied Basic Research Foundation (NO. 2024A1515012854), Guangzhou Science and Technology Program Jointly Funded by Municipal Schools and Institutes (NO. 2023A03J0338), and in part by VA Clinical Merit and ASGE clinical research funds (FWL).

Author contributions

Y.Z.: Conceptualization, Methodology, Visualization, Writing – original draft; Y.L.: Data curation, Formal analysis, Visualization, Writing – original draft; W.H.: Data curation, Validation, Visualization, Writing – original draft; S.T.: Data curation, Validation, Writing – original draft; R.Z.: Data curation, Writing – original draft; Y.L.: Methodology, Writing – original draft; F.W.L: Conceptualization, Writing – review & editing; K.C.: Methodology, Writing – review & editing; W.S.: Conceptualization, Writing – review & editing; H.C.: Conceptualization, Data curation, Funding acquisition, Project administration, Supervision, Writing – review & editing

Conflicts of interest disclosure

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Guarantor

Prof. Felix W Leung, E-mail: felixleung@socal.rr.com; Prof. Kequan Chen, E-mail: yinyuedegushi@126.com; Prof. Weihong Sha, E-mail: shaweihong@gdph.org.cn; and Prof. Hao Chen, E-mail: chenhao@gdph.org.cn.

Research registration unique identifying number (UIN)

Not applicable. The datasets used in this study were obtained from publicly available sources.

Provenance and peer review

Not commissioned, externally peer-reviewed.

Data availability statement

The datasets generated and analyzed during this study are available from the corresponding author on reasonable request. The GWAS summary statistics for GERD and asthma used in this study were obtained from publicly available sources: GERD GWAS data: https://pubmed.ncbi.nlm.nih.gov/31527586/, https://pubmed.ncbi.nlm.nih.gov/34187846/. Asthma GWAS data: https://pubmed.ncbi.nlm.nih.gov/34103634/, https://pubmed.ncbi.nlm.nih.gov/29273806/. The scRNA-seq data used in this study are available in the GEO database under accession number GSE218607.

Presentation

None.

Assistance with the study

We gratefully acknowledge the contributions from publicly available databases and the participants who contributed to those studies.