Reduced ALDH2 in the respiratory tract associates with dysregulated alcohol metabolism and respiratory reactions in AERD

Abstract

Background:

Patients with aspirin-exacerbated respiratory disease (AERD) frequently experience alcohol-induced respiratory reactions, though the immunologic mechanism is unknown. Many East Asian patients experience intense alcohol-induced skin flushing, due to an alcohol dehydrogenase 2 (ALDH2) polymorphism that leads to impaired alcohol metabolism and acetaldehyde buildup.

Objective:

Determine whether AERD-related alcohol reactions are attributable to ALDH2 deficiency in the respiratory tract.

Methods:

AERD patients (n=600) completed a survey to assess alcohol-induced symptoms and disease severity; urinary leukotriene E4 and prostaglandin D2 was quantified in a subgroup (n=72). ALDH2 protein and transcript were measured in surgically excised sinus tissue and sorted sinus epithelial cells, and in cultured respiratory epithelial cells pre- and post-IL-4/IL-13 stimulation. ALDH2 expression was assessed in nasal scrapings from AERD patients before and after dupilumab and mepolizumab.

Results:

Alcohol-induced upper respiratory, lower respiratory, and skin flushing symptoms were reported by 79.6%, 45.1% and 38.8% of AERD patients. Patients who experienced alcohol-induced respiratory reactions had faster post-operative polyp regrowth, worse baseline sinus disease and asthma control, and higher urinary eicosanoids than did alcohol non-reactors. Treatment with dupilumab or daily aspirin improved alcohol-induced symptoms for most patients. Nasal polyp ALDH2 protein and nasal epithelial cell ALDH2 transcripts were lower in AERD patients than in aspirin-tolerant controls. In vitro stimulation with IL-4/13 decreased ALDH2 expression in epithelial cell cultures. In vivo inhibition of IL-4Rα with dupilumab increased nasal ALDH2 transcript.

Conclusion:

Acquired ALDH2 enzyme deficiency within the respiratory tract in AERD, likely due to high local levels of IL-4 and IL-13, may prevent the degradation of alcohol-derived acetaldehyde, leading to mast cell activation and alcohol-induced respiratory reactions.

Capsule Summary:

This study identifies local ALDH2 deficiency as a mechanism for alcohol-induced reactions in AERD and demonstrates that IL-4/IL-13 blockade with dupilumab reverses this deficiency and improves patient-reported alcohol sensitivity.

Introduction:

Aspirin-exacerbated respiratory disease (AERD) is defined by the clinical triad of asthma, chronic rhinosinusitis with nasal polyposis (CRSwNP), and respiratory reactions to aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs).(1) It is characterized by eosinophilic inflammation and chronic activation of mast cells within the respiratory mucosa.(2) While all AERD patients develop NSAID-induced respiratory reactions, many AERD patients also develop alcohol-induced respiratory reactions, as do some patients with CRSwNP and asthma but no NSAID sensitivity.(3–5) These reactions can develop within minutes of consuming a single alcoholic beverage which most commonly leads to upper respiratory symptoms and often lower respiratory symptoms as well. Alcohol hyperreactivity is more prevalent in patients with AERD and CRSwNP patients with severe sinus disease and comorbid asthma - the severity of their reactions to aspirin correlates to the severity of their reactions to alcohol.(4,5)

Some of the current treatments for CRSwNP and asthma may partially mediate the alcohol-related symptoms in patients with AERD. Endoscopic sinus surgery followed by aspirin desensitization and high-dose aspirin therapy led to an improvement in patient-reported alcohol tolerance for 86.5% of AERD patients.(6) In a case series of six patients with AERD and a history of alcohol-induced respiratory symptoms, all six reported complete resolution of alcohol-induced symptoms after treatment with the anti-IL-4Rα biologic dupilumab.(7) Interestingly, one patient was eventually transitioned from dupilumab to the anti-IL-5Rα biologic benralizumab and reported a recurrence of her alcohol-induced sinonasal symptoms upon switching biologics. In another study of eight AERD patients, premedication with montelukast plus cetirizine before oral alcohol challenges reduced their alcohol-induced nasal and lower respiratory symptoms when compared to placebo pretreatment.(8)

The underlying mechanism of these alcohol-induced respiratory reactions in AERD has remained a mystery, but accumulating evidence suggests a central role for acetaldehyde, a toxic metabolite of ethanol. In some patients, most prevalent in East Asian populations, a genetic polymorphism in acetaldehyde dehydrogenase (ALDH2) leads to reduced acetaldehyde metabolism resulting in elevated acetaldehyde levels following alcohol consumption. This causes the phenomenon known as “alcohol flush,” characterized by vasodilation associated with facial flushing, and in some cases respiratory symptoms including nasal congestion and bronchoconstriction.(9)

The development of alcohol-induced skin flushing and respiratory reactions in ALDH2-deficient patients correlate with the extent to which ethanol ingestion raises their blood acetaldehyde levels.(10,11) Though the precise mechanism by which acetaldehyde induces their symptoms is not clear, several prior interventional studies have provided crucial mechanistic insights. An early study of patients with alcohol flush showed that a single dose of 640 mg of aspirin, taken one hour prior to ethanol ingestion, markedly reduced the alcohol-induced facial flushing.(12) Given that aspirin blocks production of cyclooxygenase-derived prostaglandin (PG)D₂ and that PGD₂ can induce both vasodilatory skin flushing(13) and bronchoconstriction(14), these results hinted that a mast cell product may be involved in the symptoms of ALDH2 deficiency-related alcohol flushing. Pretreatment with H1/H2 receptor antihistamines has also been shown to partially block the alcohol-induced skin flushing, implicating yet another mast cell product.(15) The alcohol-induced respiratory reactions experienced by ALDH2-deficient asthmatic patients have also been linked to mast cell activation. This was demonstrated in a Japanese study of asthmatics in whom alcohol challenges caused a ≥20% reduction in lung function, which was accompanied by an increase in both blood acetaldehyde and histamine levels.(11) Furthermore, pretreatment with either the antihistamine cyproheptadine or disodium cromoglycate, a mast cell stabilizing agent, partially suppressed these respiratory reactions. Another study of patients with alcohol-exacerbated asthma found that alcohol challenges caused an increase in urinary levels of the PGD₂ metabolite, 9α,11β-PGF₂, but not a corresponding increase in urinary levels of leukotriene E₄ (LTE₄).(16) While the divergence between PGD₂ and LTE₄ release may be seemingly contradictory, there is evidence to suggest that acetaldehyde’s activating effects on mast cells do not include the release of cysteinyl leukotrienes (cysLTs). In vitro, acetaldehyde causes histamine release from human airway mast cells, and some degranulation, but does not induce their release of cysLTs.(17) In vivo, pretreatment with the CysLT1R inhibitor pranlukast failed to abrogate acetaldehyde-induced bronchoconstriction in asthmatics, suggesting that cysLTs may not be the primary mediators by which acetaldehyde, and likely alcohol, causes respiratory reactions.(18) Further, disodium cromoglycate, which suppressed alcohol-induced reactions in asthmatics, prevents the mannitol challenge-induced rise in urinary PGD₂ metabolites but does not prevent the parallel rise in LTE₄, suggesting that its stabilizing effects on mast cells are more restricted to the release of PGD₂.(19)

Building upon these findings, we hypothesized that in AERD patients impaired acetaldehyde metabolism within the respiratory tract, potentially due to altered ALDH2 enzyme activity, may lead to alcohol-related mast cell activation and the development of alcohol-induced respiratory symptoms.

Methods:

Collection of surgically excised sinus samples:

Diseased ethmoid sinus tissue or concha bullosa tissue was collected from patients between the ages of 18 and 75 who underwent endoscopic sinus surgery at Brigham and Women’s Hospital (Boston, MA) for CRSwNP with AERD, NSAID-tolerant CRSwNP, chronic rhinosinusitis without nasal polyps (CRSsNP) or anatomic nasal blockage (concha bullosa). The local Institutional Review Board approved the study and all subjects provided written informed consent.

Sinus tissue was split with one segment immediately preserved in RNAlater (Qiagen, Valencia, CA) and the remainder placed in RPMI medium (Corning, Corning, NY) with 10% FBS and 1 U/mL penicillin-streptomycin and transported to the lab on ice. For protein extraction, a segment of tissue was transferred into Cell Lytic M Cell lysis reagent (Sigma-Aldrich, St Louis, MO) with 2% protease inhibitor (Roche, Indianapolis, IN) and homogenized using a gentleMACS Dissociator (Miltenyi Biotec, San Diego, CA). The homogenized fluid was spun down with supernatant stored at −80°C. Additional sinus tissue segments for transcript analysis were chopped, digested in RPMI-1640 medium with 10% FBS containing type IV collagenase and DNase I at 37°C for 30 min, and triturated using 16g and 18g needles every 15 min. The resultant cell suspension was filtered, washed, and resuspended. Participants undergoing sinus surgery also had nasal epithelial cells (EpCs) noninvasively collected with a FLOQSwab cytobrush (Copan, Murrieta, CA) from the inferior turbinate for culture.

Nasal EpC Culture for protein assessments:

Nasal EpCs were cultured for 2 weeks on 6-well plates coated with 0.03 mg/mL collagen in Pneumacult Ex-Basal Media + 50X Supplement (StemCell Technologies) with 0.096 g/mL hydrocortisone, 1 U/mL penicillin-streptomycin, 30 mg/mL gentamicin, 15μg/mL amphotericin B, 1 μM A-83–01 (Tocris Bioscience), 0.2 μM DMH-1 (Tocris Bioscience), 5 μM Y27632, and 2.5 μM DAPT. Cells were stimulated with 10 ng/mL of IL-4 and IL-13 for 96 hours before lifting with Accutase (Millipore Sigma, Burlington, MA). Cells were resuspended in Cell Lytic M Cell lysis reagent and protein lysates were collected as above and assessed for ALDH2 protein by ELISA (Novus Biologicals, Centennial, Colorado).

ALDH2 protein measurement in sinus tissue and EpC cultures:

Protein lysate supernatants from nasal polyp tissue of AERD patients who had previously reported alcohol-induced respiratory reactions or from ethmoid sinus tissue of CRSsNP were assessed for ALDH2 protein by ELISA (Novus Biologicals). Total protein was measured with the Pierce^® BCA Protein Assay kit (ThermoFisher).

ALDH2 transcript analysis in sinus EpC subsets:

Cell suspensions from excised sinus tissues were sorted using the gating strategy shown in Fig E1. In brief, cells were blocked with Fc receptor blocker (BioLegend), then incubated with fluorophore-labeled antibodies to human EpCAM (BioLegend), NGFR (Abcam), BCAM (MBL International), CD31 (BioLegend), CD45 (BioLegend), and CD90 (BioLegend) for 30 min. 7AAD (BioLegend) was added before cell sorting. Live EpCAM⁺ Lineage⁻ EpCs were sorted into three populations: BCAM^high EpCAM^low NGFR⁺ basal EpCs, EpCAM^high NGFR⁺ transitional EpCs, and EpCAM^high NGFR⁻ differentiated EpCs. 1000 cells were collected for each population, mixed with 5μl TCL buffer (Qiagen), and stored at −80°C. RNA libraries were prepared using Smart-Seq2 and 38bp paired-end sequencing was performed by the Harvard-MIT Broad Institute. Reads were pseudo-aligned and transcript expression quantified using Kallisto with the Ensembl GRCh38 transcriptome.

Nasal and bronchial air-liquid interface (ALI) cultures:

Nasal ALI cultures were grown from basal EpCs from 6 donors with concha bullosa and no reported alcohol-induced sensitivity by seeding 100K cells per well in 6.5 mm Transwell plates (Costar). After confluence was achieved, the media in the upper chamber was aspirated and ALI medium was added to the lower chamber every other day. ALI medium components were as follows: ALI Complete Base Medium (STEMCELL Technologies) containing 100uL of PneumaCult^™-Ex Plus 50X Supplement, 50uL StemCell Hydrocortisone (STEMCELL Technologies), and 20 uL StemCell Heparin (STEMCELL Technologies). Mature ALIs were stimulated with 10 ng/mL IL-4 and IL-13 for 7 days, then lysed with TCL buffer, and Smart-Seq2 paired-end bulk RNA-seq was performed by the Harvard-MIT Broad Institute. Alternatively, bulk RNA-seq FASTQ files and metadata from human bronchial EpC ALI cultures stimulated with 10 ng/mL of IL-13 for 7 days were downloaded from the NCBI Gene Expression Omnibus.(20) FASTQ files from both nasal and bronchial ALIs were pseudo-aligned and quantified using Kallisto, and differential expression testing was performed using DESeq2 with donor as a covariate.

ALDH2 transcript analysis in a human single cell RNA-seq dataset:

Metadata and scRNA-seq FASTQ files containing epithelial and immune cells from the Wang upper airway dataset(21) were downloaded from the Genome Sequence Archive [HRA000772] and aligned to Gencode GRCh38 using 10x Genomics Cell Ranger v6.1.2. Cells with >12,000 UMIs, >15% mitochondrial reads, or <500 genes detected were filtered out. RNA counts were scaled using the NormalizeData function. Principal component analysis was performed on top variable genes and PCs were subsequently corrected using the Harmony package to correct for disease (“Public_Description”) and sex (“Gender”). UMAPs were generated using Harmonized principal components, and the dataset was sequentially re-clustered into various immune and epithelial populations.

Nasal EpC scrapings:

Bulk RNA-seq data of EpCs scraped from the inferior turbinate of 15 patients with AERD who were treated with dupilumab for 3 months as part of a previously published study(22) were analyzed specifically for ALDH2 transcript expression at baseline (pre-dupilumab) and again after 1 or 3 months of dupilumab treatment. Additionally, from a separate study, ALDH2 transcript expression in nasal scraping transcripts was analyzed from 10 AERD patients who were on no biologic and 18 AERD patients who had been on mepolizumab for at least 3 months. (23)

Survey study:

A newly developed survey was electronically sent to all 2566 participants in the Brigham and Women’s Hospital AERD Registry through REDCap in December 2024. After 2 weeks, all completed survey responses were downloaded for analysis. Participants who reported that they had ever regularly consumed alcohol were then asked about history of alcohol-induced respiratory and flushing reactions by asking “Has drinking alcohol ever triggered any of the following symptoms (stuffy nose/nasal congestion, runny nose, shortness of breath, wheezing, and/or skin flushing or redness)?”. Respondents were asked to check all symptoms that applied and answered questions about time of onset and severity for each symptom. Patients who had ever used high-dose aspirin therapy, omalizumab, dupilumab, tezepelumab, mepolizumab/benralizumab, or zileuton for at least 3 months were asked if their alcohol-induced symptoms changed while on those therapies. Baseline data from the participants’ initial registry enrollment was available for the patient-reported post-operative polyp regrowth rate, Asthma Control Test (ACT) scores, and Sino-Nasal Outcome Test (SNOT-22) scores.

Urinary eicosanoids:

For 72 of the AERD patients who completed the alcohol symptom survey, urine had previously been collected and analyzed for eicosanoid levels as part of a prior clinical trial or study(24–26). Briefly, baseline urine samples from participants who had been off NSAIDs for at least 1 week were analyzed by chromatography–mass spectrometry at Vanderbilt University, as described previously(27). Concentrations of LTE₄ and the major urinary PGD₂ metabolite (PGD-M) were measured and reported as nanograms per milligram of creatinine.

Statistical analyses:

Differences between groups with non-normally distributed data (measurements of baseline respiratory disease severity, urinary eicosanoids, and nasal tissue ALDH2 protein levels) were analyzed by Mann-Whitney or Kruskal-Wallis, with data shown as median plus interquartile range. Statistical analysis of ALDH2 transcript levels in sorted EpC subsets was performed with one-way ANOVA followed by uncorrected Fisher’s LSD test for multiple comparisons between the 3 patient subgroups, with data shown as mean plus standard deviation. ALDH2 transcript and protein levels in cultured EpCs and ALDH2 transcript in nasal scrapings pre- and post-dupilumab were compared with paired t-tests and shown as individual values or mean plus standard deviation.

Results:

Prevalence and characteristics of alcohol-induced symptoms in AERD:

A total of 600 study participants who had ever regularly consumed alcohol responded to the survey and had complete survey responses. There was a female predominance (73.0% female), consistent with the demographics of the entire AERD registry (73.1% female); 94% of the respondents were Caucasian, and only 4 respondents were of Asian descent. Overall, 86.2% of all AERD respondents reported experiencing at least one type of alcohol-induced symptom, though this was slightly higher in females vs males (90.4% vs 74.7%, p<0.0001). Upper respiratory reactions were the most common, with stuffy nose/congestion and/or runny nose reported by 79.6%, whereas lower respiratory symptoms of shortness of breath and/or wheezing were reported by 45.1%, and alcohol-induced skin flushing was reported by 38.8% of respondents (Figure 1A). A small minority (2.3%) of AERD patients reported skin flushing as their only alcohol-induced symptom. The majority of patients described the onset of these symptoms within 15–60 minutes after drinking alcohol, with the onset of flushing generally being the fastest (48.4% who developed flushing noted that it began within 15 minutes of drinking) and wheezing being the most delayed (30.3% who developed wheezing noted that it began >1 hour after drinking). For patients with alcohol-induced flushing, the majority (95.7%) noted that it occurred on their face and/or neck, 31.7% on their chest, and only 5.2% flushed on their limbs or elsewhere on their body. Baseline markers of respiratory disease severity were compared between patients who did or did not report developing alcohol-induced respiratory symptoms (Figure 1B). Patients who experienced alcohol reactions also reported a faster historical rate of nasal polyp regrowth after sinus surgery (n=389, median 8 months vs 18 months, p=0.0146), more substantial impairment in sinus-related quality of life (n=523, median SNOT-22 score of 50.0 vs 39.5, p<0.001) and poorer asthma control (n=563, median ACT score of 21.0 vs 23.0, p=0.0085) than AERD patients who did not experience any alcohol-induced reactions.

Figure 1. Characterization of alcohol-induced symptoms and relationship to disease severity and urinary eicosanoids.

A. Reported rates of alcohol-induced symptoms among survey participants with AERD. B. Baseline respiratory disease severity in patients with or without alcohol-induced respiratory symptoms are shown as median + interquartile range. C. Reported change in alcohol-induced symptom severity for patients treated with dupilumab, high-dose aspirin, zileuton, omalizumab, or anti-IL-5/5Rα biologics. D. Basal urinary levels of LTE₄ and PGD-M in AERD patients who reported no alcohol-induced symptoms, only upper respiratory, or both upper and lower respiratory symptoms. Data shown are individual points and median + interquartile range.

Survey participants who had used daily aspirin, omalizumab, dupilumab, tezepelumab, mepolizumab/benralizumab, or zileuton for at least 3 months were asked if their alcohol-induced symptoms had changed (“better/improved”, “same”, or “worsened/more severe”) while on those therapies. The majority of patients who had been on dupilumab (79% of n=169) and on high-dose aspirin therapy (59% of n=128) reported improvements in their alcohol symptoms, with 46% of the 33 patients on zileuton, 37% of the patients on omalizumab (n=30), and 36% of the patients on anti-IL-5/5Rα biologics (n=39), reporting improvement (Figure 1C). Only 6 patients had been on tezepelumab for ≥3 months, so its effects could not be determined.

Relationship between alcohol-induced respiratory reactions and urinary eicosanoids:

Of the 600 AERD patients who completed the alcohol survey, urinary levels of LTE₄ and PGD-M were available for 72 of them. Of those, 13 reported no reactions to alcohol, 42 experienced only upper respiratory symptoms, and 17 reported both upper respiratory and lower respiratory symptoms induced by alcohol. Basal levels of both urinary LTE₄ and PGD-M were lowest in the no-reaction group, trended higher in the upper respiratory reactors, and were both significantly higher in the upper and lower alcohol reaction patients (no alcohol reaction group vs upper and lower reactors, p=0.0076 for LTE₄ and p=0.0289 for PGD-M) (Figure 1D).

Sinus tissue ALDH2 protein and nasal EpC ALDH2 transcript expression:

ALDH2 protein levels in excised sinus tissue from AERD patients who reported alcohol-induced respiratory symptoms (n=10) were significantly lower than in that from aspirin-tolerant controls who tolerated alcohol (n=10) (median 172.0 vs 319.3 pg/μg total protein, p=0.0029) (Figure 2A). Exploration of the ALDH2 transcript expression in a publicly-available human scRNA-seq dataset(21) identified ALDH2 expression in a variety of cells in the upper airway tissues including EpCs, stromal cells (particularly fibroblasts), and immune cells (particularly macrophages) (Figure 2B). Bulk RNA-seq analysis of our data of sorted sinus EpC subsets identified the highest ALDH2 expression in EpCs from non-polyp CRSsNP control sinus tissue, particularly in the transitional and basal EpCs (Figure 2C). Compared to CRSsNP controls, ALDH2 expression was significantly lower in all EpC subsets from aspirin-tolerant CRSwNP patients and was lowest in all EpC subsets from AERD patients (Figure 2C).

Figure 2. Sinus tissue and nasal epithelial cell ALDH2.

A. ALDH2 protein levels in ethmoid sinus tissue from alcohol-tolerant patients with CRSsNP (n=10) or alcohol-sensitive patients with AERD (n=10) are shown as individual points and median + interquartile range. B. ALDH2 transcript expression in single-cell RNA-sequencing dataset of sinus mucosa cells (21). C. ALDH2 transcript levels in flow-sorted differentiated, transitional, and basal EpC subsets from surgically excised sinus tissue from CRSsNP (n=4), NSAID-tolerant CRSwNP (n=5), and AERD (n=5) patients.

Effect of IL-4/IL-13 and IL-4Rα inhibition on nasal ALDH2:

In vitro stimulation with IL-4 and/or IL-13 significantly suppressed ALDH2 mRNA expression in ALI cultures derived from human nasal and bronchial EpCs (Figure 3A). Stimulation with IL-4 and IL-13 also significantly decreased ALDH2 protein in submerged cultured nasal EpCs (Figure 3B). Bulk RNA-seq of inferior turbinate cell scrapings taken from 15 AERD patients before and again after treatment with dupilumab for 1–3 months showed that in vivo inhibition of IL-4Rα led to an increase in normalized ALDH2 transcript compared to their pre-treatment baseline levels for 11 of the 15 patients (baseline mean 24.8 ±22.5 vs on-dupilumab mean 53.4 ±32.8, p=0.0151) (Figure 3C). The normalized ALDH2 transcript level of similar scrapings from 10 AERD controls on no biologic and 18 AERD patients on mepolizumab did not show significant difference between them (80.7 ±46.5 vs 91.1 ±48.6, p=0.5864).

Figure 3. ALDH2 transcript and effects of IL-4/13.

A. ALDH2 transcript in cultured healthy nasal mucosa-derived ALIs with or without IL-4/13 (left) and in healthy lung-derived bronchial ALIs with or without IL-13 (right). B. ALDH2 protein levels in healthy nasal mucosa-derived submerged EpCs with or without IL-4/13. Data are shown as individual values. C. Bulk RNA-seq data of nasal scraping cells from paired patient samples (n=15) show the dupilumab-induced in vivo change in ALDH2 from baseline to after 1–3 months of dupilumab. Data are shown as individual values and means; analysis with paired t-tests.

Discussion:

As in prior smaller survey studies from our group and others(3,4), this larger study confirms that most patients with AERD experience alcohol-induced respiratory reactions and that patients who develop reactions to alcohol have more severe respiratory disease overall (Figure 1A,B). Further, we show that for more than a third of those patients, their alcohol-induced respiratory reactions are also accompanied by facial flushing (Figure 1A). It has been previously noted that elevations in basal urinary LTE₄ and PGD-M levels in AERD correlate with the severity of both aspirin-induced respiratory symptoms and aspirin-induced skin symptoms (27). This study now demonstrates that urinary levels of LTE₄ and PGD-M are also higher in the AERD patients who develop alcohol reactions than in the non-reactors (Figure 1D). This is likely because the burden of activated mast cells in the respiratory tissues is highest in the subset of patients who have the most severe disease overall, and those patients are in the group that is prone to the most severe alcohol-induced reactions. Our survey respondents noted several therapies that, when used to treat their underlying CRSwNP and asthma, did provide some perceived symptomatic improvement in their alcohol-induced reactions. Of these, dupilumab was perceived to be effective by the highest percentage of participants, followed by high-dose aspirin, zileuton, and omalizumab (Figure 1C), which is in line with previously published case series of AERD patients who reported similar treatment efficacies. (6,7,28) Interestingly, all of these agents have effects on mast cell mediator release, as dupilumab and omalizumab both inhibit mast cell activation, zileuton blocks both LTE₄ and PGD₂ production (29), and aspirin blocks PGD₂ production.

Prior studies of alcohol-induced flushing and bronchoconstriction in patients with ALDH2 polymorphisms showed that the subsequent ALDH2 enzyme deficiency leads to dysregulated alcohol metabolism and elevated systemic levels of acetaldehyde, with the release of mast cell mediators as the likely underlying cause of their alcohol reactions. Therefore, we sought to determine if abnormalities in ALDH2 function play a similar role in alcohol-induced reactions in AERD. We measured ALDH2 protein levels in the sinus tissue of patients with AERD who had experienced alcohol reactions, and compared it to that in samples from control patients without respiratory disease. Results revealed that AERD patients have significantly less ALDH2 in their nasal tissue with levels of only about half of that found in healthy controls (Figure 2A). To further study the relationship between ALDH2 and the development of alcohol-induced respiratory symptoms, we examined single-cell RNA sequencing data of sinus tissue cells and noted high expression of ALDH2 in structural cells, especially EpCs (Figure 2B). Regardless of the maturation or differentiation stage of the EpCs, ALDH2 transcript expression was notably reduced in the nasal epithelium of AERD patients compared to that of aspirin-tolerant CRSwNP patients or non-polyp CRS patients (Figure 2C). To explore the potential causes of this reduction in ALDH2, we cultured EpCs collected from nasal brushings of healthy patients with no alcohol-induced symptoms. Given that increased local tissue levels of IL-4 and IL-13 are associated with the development of CRSwNP and AERD, we created an in-vitro AERD-like EpC phenotype by stimulating the cells with IL-4 and IL-13 for 96 hours. There was a significant reduction in ALDH2 transcript following that stimulation, which was paralleled by the results of an equivalent IL-13-induced stimulation experiment with cultured human bronchial EpCs (Figure 3A). The consequence of that IL-4/13-induced reduction in transcript expression was then confirmed when stimulation of cultured basal nasal EpCs with IL-4 and IL-13 also led to decreased cellular expression of ALDH2 protein (Figure 3B). Given that the anti-IL-4α biologic dupilumab should effectively inhibit the signaling pathways for both IL-4 and IL-13, we then looked at whether patients with AERD had any change in their ALDH2 expression following treatment with dupilumab. Indeed, in the 15 patients for whom we had bulk RNA sequencing data from scrapings of their inferior turbinate nasal tissue before and again after 1–3 months of dupilumab, there was a significant dupilumab-induced increase in ALDH2 (Figure 3C), which was not seen in patients treated with mepolizumab.

On the basis of findings from this and previous studies, we propose the following mechanism of alcohol-induced respiratory symptoms in AERD patients. Ethanol, when ingested orally, is decomposed by alcohol dehydrogenase to yield acetaldehyde, which is then degraded by the ALDH2 enzyme to yield acetic acid. In patients for whom ALDH2 enzyme activity is globally reduced, as is the case in patients with an ALDH2 polymorphism, ethanol ingestion leads to elevated systemic levels of acetaldehyde, which then induces the release of histamine and PGD₂ from the mast cells of the skin, causing flushing. Some acetaldehyde is also present in pulmonary circulation and spreads through the airway (30), inducing activation of mast cells within the airway mucosa, which can cause associated bronchoconstriction in susceptible patients. In AERD, our findings suggest that at least their upper airways have locally reduced ALDH2 enzyme levels, likely due to the effects of chronically elevated IL-4 and IL-13 within their respiratory tissues. In AERD patients, we suspect that ethanol ingestion may lead to some elevation in blood levels of acetaldehyde, given that 38.8% do also report skin flushing, but we suspect that there is a more notable acetaldehyde accumulation within the respiratory tract, likely due to decreased metabolic enzyme levels within local respiratory tissue. In combination with the high mast cell burden prevalent in the respiratory mucosa in AERD patients, the ethanol-induced acetaldehyde accumulation then may cause mast cell degranulation and release of histamine and PGD₂, and possibly cysLTs, resulting in the classic alcohol-induced respiratory reactions observed in AERD. The association of both more severe respiratory disease (Figure 1B) and higher urinary levels of LTE₄ and PGD-M (Figure 1D) with the development of alcohol-induced respiratory reactions is in line with the proposed mechanism of mast cell overactivation, which likely underlies the baseline disease severity, the urinary eicosanoids, and the alcohol-induced symptoms.

In summary, these findings provide novel mechanistic insight into alcohol-induced respiratory reactions in AERD, demonstrating that reduced ALDH2 expression in the airway epithelium—likely driven by IL-4 and IL-13—contributes to localized acetaldehyde accumulation and subsequent mast cell activation. The observed dupilumab-induced restoration of ALDH2 expression suggests that targeting IL-4/IL-13 signaling may mitigate alcohol-induced reactions in AERD, paralleling its known benefits for overall disease control. Clinically, these results highlight the potential for biologic therapies to improve not only chronic respiratory inflammation but also specific inflammatory triggers such as alcohol, which significantly impact patient quality of life.

Key Messages:

• Alcohol-induced respiratory symptoms in AERD are associated with higher urinary eicosanoid levels and more severe respiratory disease.

• ALDH2 deficiency within the respiratory tract in AERD impairs acetaldehyde metabolism and may underlie their alcohol-induced respiratory reactions.

• IL-4 and IL-13 downregulate ALDH2 in airway epithelial cells, and treatment with dupilumab restores nasal epithelial ALDH2 expression in vivo and improves alcohol tolerance in AERD patients.

Abbreviations used:

AERD

aspirin-exacerbated respiratory disease

ACT

Asthma Control Test

ALDH2

aldehyde dehydrogenase 2

ALI

air-liquid interface

CRSwNP

chronic rhinosinusitis with nasal polyps

CRSsNP

chronic rhinosinusitis without nasal polyps

CysLT

cysteinyl leukotriene

EpC

Epithelial cell

IL

interleukin

LT

leukotriene

NSAID

nonsteroidal anti-inflammatory drug

PG

prostaglandin

SNOT-22

sinonasal outcome test-22