Updates on Immune Mechanisms in Aspirin-Exacerbated Respiratory Disease

Abstract

Aspirin-exacerbated respiratory disease (AERD) has fascinated and frustrated specialists in allergy/immunology, pulmonology, and otorhinolaryngology for decades. It generally develops in previously healthy young adults and is unremitting and challenging to treat. The classical triad of asthma, nasal polyposis, and pathognomonic respiratory reactions to aspirin and other cyclooxygenase (COX)-1 inhibitors is accompanied by high levels of mast cell activation, cysteinyl leukotriene (cysLT) production, platelet activation, and severe type 2 respiratory inflammation (T2I). The “unbraking” of mast cell activation and further cysLT generation induced by COX-1 inhibition reflect an idiosyncratic dependency on COX-1-derived products, likely prostaglandin E₂, to maintain a tenuous homeostasis. While cysLTs are clear disease effectors, little else was known about their cellular sources and targets, and the contributions from other mediators and T2I effector cells to disease pathophysiology were unknown until recently. The applications of targeted biological therapies, single cell genomics, and transgenic animal approaches have substantially advanced our understanding of AERD pathogenesis and treatment and have also revealed disease heterogeneity. This review covers novel insights into the immunopathogenesis of AERD from each of these lines of research, including the roles of lipid mediators, effector cell populations, and inflammatory cytokines, discusses unanswered questions regarding cause and pathogenesis, and considers potential future therapeutic options.

Introduction.

Over the last few years, new insights into the mechanisms of immune dysregulation that drive aspirin-exacerbated respiratory disease (AERD) have come from a variety of study types, including clinical trials using targeted biologics and therapeutics, investigations of mouse models of respiratory inflammation, and the application of novel technologies like single-cell RNA sequencing. Some of these recent advances in our understanding of AERD pathogenesis have provided insights into the mechanism of action by which our current therapies provide therapeutic benefit, and others have provided the rationale for the development of new therapies directed against selected targets. Though there is much more yet to be learned, there is now considerable knowledge of the involvement of particular lipid mediators, respiratory effector cells, antibody overproduction, and cytokine-driven inflammation in AERD.

Lipid mediators.

Once regarded principally as regulators of smooth muscle tone and vascular permeability, arachidonic acid-derived lipid mediators are now well known to regulate features of innate and adaptive immunity. Long-recognized disturbances in lipid mediator production and receptor function associated with AERD now can be considered in the greater context of the dysregulated type 2 inflammation (T2I) that underlies the disease.

Prostaglandin (PG)E₂.

PGE₂ is the most ubiquitous COX product, being generated in some quantity by most cell types, although the most prominent sources are macrophages, epithelial cells, and fibroblasts. Either constitutive COX-1 or inducible COX-2 can drive PGE₂ production by converting arachidonic acid to PGH₂, which then serves as a substrate for conversion to PGE₂ by one of three different PGE₂ synthase (PGES) enzymes (1). PGE₂ stabilizes and prevents the activation of mast cells (2) and group 2 innate lymphoid cells (ILC2s)(3), and prevents the formation of cysLTs by inhibiting the translocation of 5-lipoxygenase (5-LO) (4). Several studies link AERD to diminished production (or increased catabolism) of PGE₂ in the respiratory tract (5, 6), associated with diminished expression and function of COX-2 and/or microsomal (m)PGES-1, the dominant PGES isoform that is upregulated with inflammation (7)(Fig. 1). Additionally, expression of the PGE₂-selective EP₂ receptor by hematopoietic cells (8, 9) and fibroblasts (7, 10) is reduced in AERD. Since COX-1 is exquisitely sensitive to inhibition by aspirin, a chronic reduction in COX-2/mPGES-1-derived PGE₂ could account for aspirin-induced reactions in AERD, by removing a critical COX-1-dependent “braking” function. Consistent with this hypothesis, inhaled PGE₂ not only blocks changes in lung function in response to aspirin challenges in AERD, but also blocks the characteristic aspirin-induced increase in urinary leukotriene (LT)E₄ (11). Mice lacking mPGES-1 (Ptges^−/− mice) or EP₂ receptors (Ptger2^−/− mice) display higher levels of airway T2I, mast cell activation, ILC2 expansion, and cysLT production than do wild-type controls after intranasal instillation with dust mite antigens, and display AERD-like physiologic reactions to aspirin (12). While the causes of impairments in the COX-2/mPGES-1 system and EP₂ expression are unknown, it is likely that they contribute to the T2I, mast cell activation, and overproduction of cysLTs that characterize AERD.

Figure 1. Summary of the pathogenesis of the chronic inflammation in the respiratory tissue in AERD.

IL-33 and TSLP derived from dysplastic epithelial basal cells activate Th2 cells and ILC2s to produce IL4, IL-13, and IL-5, as well as activating mast cells. Those cytokines in turn activate and recruit B cells, plasma cells, mast cells and eosinophils and also feedback to target the epithelium. Proinflammatory lipids CysLTs and PGD₂ are overproduced and act in concert with IL-33 to activate ILC2s, the anti-inflammatory lipid, PGE₂, is underproduced, contributing to ILC2 expansion and unchecked mast cell activation.

ILC, innate lymphoid cell; CysLT, cysteinyl leukotriene; PG, prostaglandin; TSLP, thymic stromal lymphopoietin

cysLTs.

CysLTs form from conversion of arachidonic acid to LTA₄ by 5-LO, which is then conjugated to reduced glutathione by LTC₄ synthase (LTC₄S) to form LTC₄. Mast cells, basophils, eosinophils, epithelial brush cells, and platelet-adherent granulocytes all have the capacity to generate LTC₄, which is converted extracellularly to the potent bronchoconstrictor LTD₄ (13) and its metabolite LTE₄. Urinary levels of LTE₄, a marker of systemic cysLT production, are characteristically elevated in patients with AERD relative to aspirin-tolerant controls, and increase further during NSAID-induced reactions (14). Additionally, patients with AERD are highly sensitive to bronchoconstriction induced by inhaled LTE₄ (15). Treatment with zileuton, which inhibits 5-LO and decreases production of leukotrienes, attenuates the sinonasal symptoms (16), lung function changes (17), and gastrointestinal and cutaneous symptoms that occur during aspirin-induced reactions in AERD. Additionally, a post hoc analysis of the Phase 3 studies of zileuton for the treatment of asthma revealed that the subgroup of asthmatic patients with comorbid aspirin sensitivity experienced a 22.8% increase in FEV1 from baseline, which was significantly greater than that in the aspirin-tolerant subgroup (18). These findings verify the importance of cysLTs as drivers of both the ongoing airway disease and the acute aspirin-induced reactions in AERD. Moreover, both overproduction and end-organ hyperresponsiveness likely contribute to the cysLT-driven features of AERD.

Animal studies suggest that cysLTs play a much broader role than previously appreciated in driving the exaggerated features of T2I in AERD. All three cysLTs have potent and distinct bioactivities in lung T2I, displaying differential preferences for their receptors, CysLT₂R, CysLT₁R and CysLT₃R/GPR99, respectively(19–21). As is the case for humans with AERD (16, 17), treatment of Ptges^−/− mice (which display AERD-like reactions to inhaled lysine aspirin and selective hyperresponsiveness to LTE₄) with the 5-LO inhibitor zileuton blocks the aspirin-induced changes in both lung function and mast cell activation (12). Genetic deletion of LTC₄S attenuates the eosinophilic lung inflammation, ILC2 expansion, and upregulation of IL-33 that occurs in Ptges^−/− mice treated with intranasal dust mite (20). Notably, respective deletions of CysLT₁R, CysLT₂R, or CysLT₃R also reduce lung eosinophilia and ILC2 expansion in these AERD-like mice, but only CysLT₂R deletion inhibits the upregulation of IL-33, limiting the expansion of IL-33⁺ epithelial cells (20). Moreover, LTE₄ acts at CysLT₃R to drive the expansion of brush/tuft cells (22), a unique population of epithelial chemosensory cells that are the dominant source of IL-25 in the lung and gut mucosa (22, 23), and are also capable of generating both LTC₄ and prostaglandin D₂ (PGD₂) (24, 25) Additionally, LTC₄ and LTD₄ activate and elicit proliferation of mouse lung ILC2s by engaging CysLT₁R (26)(Fig. 1), and synergize with IL-33 and IL-25 in this process (27, 28). Collectively, these studies suggest that cysLTs and their receptors function in a network that amplifies T2I through synergistic actions on ILC2s and epithelial-derived cytokines. The high levels of cysLTs in AERD, the substantial therapeutic effects of zileuton, and the robust features of respiratory inflammation in this disease suggest that this putative network may be especially important in driving disease features. Further, many of these leukotriene-dependent features may depend on cysLT receptors that are not targeted by the currently available drugs zafirlukast and montelukast, which only block CysLT₁R. This may explain the comparatively modest clinical effects of CysLT₁R blockade in AERD relative to the more robust effects of zileuton (29).

PGD₂.

PGD₂ is the dominant COX product of mast cells and tuft/brush cells and is also generated in lesser amounts by eosinophils, T helper type 2 (Th2) cells (30), and ILC2s(31). The levels of urinary PGD₂ metabolites in patients with AERD exceed those in aspirin-tolerant controls, and often paradoxically increase in parallel with LTE₄ during reactions to aspirin (32). Interestingly, a subset of patients with AERD display both especially high baseline levels of urinary PGD₂ metabolites and dramatic further increases during aspirin reactions, which is usually associated with their development of a pruritic macular rash and abdominal pain (33). PGD₂ induces chemotaxis of Th2 cells (34) and ILC2s (35) and elicits their production of cytokines (35, 36), and also elicits eosinophil and basophil chemotaxis (34). These actions all depend on the chemoattractant receptor-homologous molecule expressed on T-helper type 2 cells (CRTH2), one of two known PGD₂ receptors. In a 12-week placebo-controlled study of 43 patients with CRSwNP, 27 of whom had AERD, treatment with the selective CRTH2 antagonist AZD1981 had no effect on nasal polyp score, measures of olfaction, or measures of sinus disease symptoms compared with placebo (37). Further, in a 16-week study of the CRTH2 inhibitor fevipiprant for the treatment of CRSwNP, fevipiprant did not significantly improve nasal polyp score, nasal congestion, measures of olfaction, or measures of sinus disease symptoms compared with placebo, though there was only one patient with AERD was randomized to the treatment group(38). Neither study assessed baseline systemic or respiratory tract PGD₂ levels in the patient population, so it is not known whether these were the most appropriate target populations(39). Though these results do not completely exclude a role for PGD₂ in CRSwNP in general or AERD in particular, they do suggest that signaling through the PGD₂-CRTH2 pathway is unlikely to be a major contributor to the ongoing respiratory inflammation in CRSwNP/AERD. Whether CRTH2 blockade might attenuate features of reactions to aspirin or benefit the “PGD₂-high” subgroup remains to be determined.

Effector cell-related features.

Severe T2I of both the upper and lower respiratory tract is typical of AERD, and accounts for the clinical severity of the asthma and CRSwNP in the syndrome. Recent clinical trials and the application of molecular tools have challenged dogma regarding the contributions of particular cell types, while revealing previously unrecognized additional effectors.

Tissue eosinophilia.

Eosinophilia is a prominent histologic feature of CRSwNP and asthma, especially in AERD (40, 41). Although eosinophils are regarded as drivers of tissue dysfunction and remodeling, recent studies prompt a reassessment of these assumptions in AERD. Dexpramipexole is a small orally available molecule that depletes blood and respiratory tissue eosinophils. In an open-label study of 16 patients with CRSwNP, several of whom had AERD, dexpramipexole induced a 97% reduction from baseline in nasal polyp tissue eosinophilia. Nonetheless, there was no significant reduction in nasal polyp score, nor any significant improvement in measures of olfaction or measures of sinus disease symptoms (42). These ‘negative’ findings bring the pathogenetic role of tissue eosinophils into question. Notably, while polyp neutrophils were not affected by dexpramipexole, polyp mast numbers in the biopsies actually increased significantly over the treatment period, consistent with a possible compensatory role for mast cells disease drivers. The lack of effects on symptoms and polyp size induced by eosinophil depletion with dexpramipexole contrasts with the modest therapeutic efficacy of anti-IL-5 treatments, such as benralizumab, mepolizumab, or reslizumab, suggesting that IL-5 acts on cells other than eosinophils to support features of T2I.

Platelet activation.

Compared with aspirin-tolerant asthmatic controls, the serum of patients with AERD contains high levels of platelet activation markers (43). Additionally, nasal fluid from patients with AERD has higher levels of platelet-derived microparticles than does nasal fluid from aspirin-tolerant controls with CRSwNP (44). During inflammatory responses, activated platelets form complexes with circulating leukocytes. Platelet-adherent leukocytes become primed for adhesion and activation. Moreover, platelets strongly express LTC₄S, and adherent platelets convert granulocyte-derived LTA₄ to LTC₄ through a transcellular pathway (45). The frequencies of platelet-adherent granulocytes in the blood of patients with AERD exceed those in the blood of aspirin tolerant-controls (43, 46), and correlate with the levels of urinary LTE₄, suggesting that transcellular platelet-derived cysLTs contribute significantly to systemic cysLTs generation. A placebo-controlled cross-over trial of prasugrel, which inhibits platelet activation by blocking the P2Y12 receptor, was undertaken with 40 AERD patients to determine whether prasugrel would inhibit platelet activation and attenuate the severity of respiratory symptoms induced during aspirin challenge. While prasugrel produced no benefit for the group as a whole, a subgroup analysis revealed 5 patients for whom prasugrel pretreatment completely blocked all reaction symptoms. Those 5 responders had higher baseline indices of platelet activation and less evidence of aspirin-induced mast cell activation, suggesting a contribution from P2Y12 signaling and platelets in this subset (47). Because prasugrel did not decrease the frequency of platelet-leukocyte aggregates, it is likely that the P2Y12 pathway does not account for the aberrant platelet phenotype of AERD.

Mast cell hyperplasia and activation.

Several lines of evidence indicate that mast cells play a central role in AERD. Mast cells accumulate in the nasal and bronchial mucosa of patients with AERD at higher concentrations than in aspirin-tolerant controls (41, 48). The levels of mast cell products in the urine and nasal fluid from patients with AERD are elevated compared with aspirin-tolerant controls, as are the levels of mast cell-derived microparticles in their nasal fluids (44), indicative of steady-state mast cell activation. Mast cell products increase further during reactions to aspirin (32, 33, 49), likely reflecting the depletion of PGE₂. Treatment with cromolyn, a mast cell stabilizing drug, rapidly increases baseline lung function (50) and reduces sputum eosinophil counts (51) in subjects with AERD, suggesting that persistent mast cell activation drives several steady-state features of AERD. Moreover, cromolyn pretreatment also blocks changes in lung function during aspirin challenges, and blocks the characteristic increase in urinary LTE₄ (52). A recent study demonstrated a high frequency of proliferating (Ki67+) mast cells in AERD sinonasal tissues compared with aspirin-tolerant controls (48) and single cell RNA-sequencing revealed a high degree of activation in nasal polyp mast cells, with enrichment for both IgE- and IL-33-driven transcriptional signatures. Moreover, treatment of human cord blood mast cells with IL-4 ex vivo elicited a transcriptional signature that closely mirrored the intraepithelial population of mast cells that is the most richly expanded population in AERD tissues. These studies indicate that both innate and adaptive immune pathways drive mast cell activation in AERD, and that proliferation is the principal mechanism responsible for the marked mast cell hyperplasia observed in this disease.

ILC2 recruitment.

ILC2s are tissue resident poised effector cells capable of generating large quantities of IL-5, IL-13 and other cytokines in response to both innate cytokines (IL-33, thymic stromal lymphopoietin (TSLP), IL-25) and lipid mediators. Nasal aspirin challenges in AERD elicit a reduction in circulating ILC2s, accompanied by an increase in ILC2s in the nasal tissue (53), which parallels the peaks of urinary LTE₄ and PGD₂ metabolites. Given that PGD₂ and LTE₄ cooperatively activate human ILC2s and elicit their chemotaxis via CRTH2 and CysLT₁R, respectively (35, 54), it is possible that their recruitment in AERD is driven by the release of lipid mediators from the respiratory tract. This may also result in incremental tissue eosinophilia, as studies have noted a similar reduction in blood eosinophil counts and influx of nasal eosinophil counts associated with aspirin challenges (47, 55), suggesting that ILC2s and eosinophils may be recruited by overlapping mechanisms. Interestingly, the antibody-mediated depletion of ILC2s markedly inhibited all physiologic and biochemical features of aspirin-induced reactions in a mouse model (20), suggesting that products of ILC2s may be essential to the clinical phenotype.

Plasma cell expansion.

While the majority of mechanistic studies in AERD focus on innate effectors, the adaptive immune system also likely plays an important role. Plasma cells are found in nasal polyp tissues. A transcriptionally distinct population of plasma cells was identified within AERD nasal polyps, which was notable for strong expression of the alpha subunit of the IL-5 receptor (IL-5Rα)(56). Flow cytometry confirmed the presence of IL-5Rα+ plasma cells at higher frequencies in nasal polyps from AERD patients than in those from aspirin-tolerant CRSwNP. These plasma cells also expressed IGHG4 and IGHE, encoding the heavy chains of IgG₄ and IgE, respectively. Interestingly, the levels of both IgG4 and IgE in sinonasal tissue from patients with AERD far exceeded those in aspirin-tolerant CRSwNP, but there was no relationship between the nasal and serum levels of these antibodies. Further, nasal IgE levels were highest and IgG4 levels the lowest in patients whose nasal polyps regrew the most quickly after surgery. This study highlights a potentially unique role for plasma cells, locally generated antibodies, and IL-5Rα in AERD pathophysiology, although the immune stimuli driving these responses and the role of IL-5 in plasma cell ontogeny remain to be determined.

Epithelial dysplasia.

Epithelial cells have emerged as major contributors to the pathogenesis of T2I. A recent study used single cell RNA-sequencing to define the functional and transcriptional differences between epithelial cells from CRSsNP and CRSwNP tissue (57). Although the study was underpowered to detect differences between AERD and aspirin-tolerant CRSwNP, several key findings emerged with respect to the epithelium. First, epithelial basal cells, the principal cellular site of IL33 and TSLP expression, were markedly expanded in polyp tissue compared to non-polypoid sinus tissue. Second, the expanded basal cells bore a strong imprint of IL-4 receptor alpha (IL-4Rα)-driven signature transcripts (e.g., ALOX15, POSTN, CCL26). Third, analyses of developmental trajectories and chromatin architecture revealed that polyp epithelial basal cells display evidence of “blocked” differentiation, characterized by persistent stem-ness and loss of normal development into secretory cells, mediated by chromatin remodeling. A subsequent study revealed that ALOX15 is specifically upregulated in AERD nasal polyps relative to that in aspirin-tolerant CRSwNP (58), and that its expression correlates with that of 15-hydroxyprostaglandin dehydrogenase (PGDH) by mast cells. 15-hydroperoxyeicosatertraenoic acid (15-HETE), the product of ALOX15, is converted by PGDH into 15-oxo-ETE, a potent eosinophil chemoattractant, suggesting a mechanism by which epithelial-mast cell cross talk may facilitate tissue eosinophilia in AERD. Moreover, esterified 15-HETE was recently shown to regulate ferroptosis (59), a form of programmed cell death in epithelial cells that is driven by lipid peroxidation (60). Thus, nasal polyposis in general, and AERD in particular, may involve disturbances in epithelial homeostasis, redox balance, and regenerative capacity, while driving T2I through production of IL-33 and TSLP. The therapeutic efficacy of targeted anti-IL-4Rα antibodies, such as dupilumab, as treatments for CRSwNP and AERD could reflect a partial restoration of these functions (61).

Lessons learned from targeted therapeutics.

The advent of biologics targeting cytokines and antibodies involved in T2I, and the investigation of high-dose aspirin as a treatment regimen, has not only provided therapeutic breakthroughs for patients with AERD, but has also provided the opportunity to test existing hypotheses regarding the pathogenesis of the disease. As a result, clinical trials have yielded important and, in some instances, unexpected information regarding the causal mechanisms of AERD.

IgE.

Although AERD is not thought to be driven by classical allergen-induced pathways, some studies do support a potentially significant role for IgE in its pathogenesis. In an open-label study, treatment of AERD patients with omalizumab, a monoclonal anti-IgE antibody, dramatically decreased urinary levels of both LTE₄ and PGD₂ metabolites (62) and induced improvement in baseline symptoms. Subsequently, a placebo-controlled trial of omalizumab pretreatment followed by aspirin challenge in AERD showed that 5 of the 7 patients randomized to omalizumab then had no respiratory reaction during aspirin challenge (63). These findings support a role for IgE, likely derived locally from plasma cells in the respiratory tract (56), in driving both the steady-state eicosanoid formation and the mast cell-dependent clinical reactions to aspirin, and are consistent with the enrichment for IgE-driven activation transcripts in nasal polyp mast cells (48). The mechanisms by which IgE can induce a chronically activated mast cell population, and the antigen specificity of the locally produced IgE remain to be determined. As the entry criteria for the omalizumab studies included at least one positive skin test to environmental allergens, it is unclear whether the efficacy of anti-IgE is limited to patients with AERD who are also atopic, or whether locally produced respiratory IgE of unknown specificity drives significant mast cell activation and pathology.

IL-5.

IL-5, produced by activated Th2 cells, ILCs, and mast cells, is elevated in the respiratory tissues of patients with AERD relative to controls and is required for the development of tissue eosinophilia. Neutralization of IL-5 with the humanized antibody mepolizumab reduces blood and tissue eosinophilia in patients with severe asthma and CRSwNP and improves symptomatic control over placebo. Open label studies reveal that mepolizumab provides moderate symptomatic benefit for patients with AERD (64, 65), but does not consistently block NSAID-induced reactions (66). Notably, single cell RNA-sequencing of nasal polyps has revealed that several non-eosinophil cell types strongly express IL5RA transcript, including mast cells, ciliated epithelial cells, and plasma cells. These findings indicate that, contrary to dogma, the actions of IL-5 are not limited to eosinophils. Inhibition of IL-5 signaling on these other IL-5Rα⁺ cell types could account for the therapeutic benefit provided by mepolizumab, considering the lack of benefit that dexpramipexole afforded despite its complete eosinophil depletion in CRSwNP and AERD (42). A recent case-control study of 18 AERD patients on mepolizumab and 18 matched AERD patients not treated with mepolizumab revealed that mepolizumab decreased production of inflammatory eicosanoids, especially PGD₂, and upregulated tight junction-associated nasal epithelial cell transcripts, potentially reflecting the actions of IL-5 on mast cells and epithelial cells, respectively (65). However, the function of the IL-5Rα and the role of IL-5 signaling on these non-eosinophil airway cells still remains to be determined.

IL-4/IL-13.

IL-4 and IL-13 are critical to the pathogenesis of T2I, signal through a shared common receptor subunit, IL-4Rα, and are elevated in the nasal fluid of AERD patients compared to controls(67). RNA-sequencing has revealed that mast cells and Th2 cells are the principal sources of IL-13 in the nasal polyps (57). Treatment with dupilumab, a humanized antibody against IL-4Rα, leads to robust improvements in upper and lower respiratory tract symptoms in AERD patients, including improved sense of smell, reduced nasal polyp size, and improved lung function (68–70). Patients with AERD generally report onset of clinical improvement with dupilumab within the first month of initiating treatment (71). Additionally, perioperative initiation of dupilumab decreases the rate of polyp regrowth after endoscopic sinus surgery (72). An open label study of 22 AERD patients treated with dupilumab for three months showed a rapid and sustained improvement in both upper and lower airway symptoms and sense of smell. Concomitantly, dupilumab reduced nasal fluid albumin levels, suggesting an improvement in epithelial/endothelial barrier function, and decreased serum and nasal fluid IgE levels. Additionally, dupilumab markedly decreased nasal and urinary levels of LTE₄, while increasing nasal fluid concentrations of PGE₂. Since IL-4 and IL-13 suppress the expressions of COX-2, mPGES-1, and EP₂ receptors by human bronchial epithelial cells ex vivo (73), it is tempting to speculate that IL-4Rα blockade could at least partially restore PGE₂-dependent homeostasis of the cysLT pathway. Consistent with this hypothesis, treatment with dupilumab can increase aspirin tolerance and often completely prevent aspirin-induced reactions in AERD (74). Given the broad distribution of IL-4Rα expression, the efficacy of dupilumab in AERD likely reflects its ability to block IL-4Rα on multiple relevant cell types, including mast cells, granulocytes, epithelial cells, endothelial cells, Th2 cells, and B/plasma cells(71). Whether its effects are disease modifying or dependent on indefinite IL-4Rα inhibition remains to be determined.

Epithelial cytokines.

IL-25, IL-33, and TSLP are alarmin-type cytokines that are produced and released by epithelial barrier cells in response to injury or activation. These cytokines act directly on ILC2s, mast cells, and Th2 cells to drive T2I. Sinonasal tissues from patients with AERD contain high concentrations of TSLP (32) and IL-33 (75), both of which are expressed principally by epithelial basal cells and (in the case of IL-33) endothelial cells (57). Platelets, which contain pre-formed IL-33 (76), are another potential source in polyp tissue. Nasal polyp mast cells are strongly enriched for IL-33-driven activation signature genes (48) and blockade of IL-33 dramatically attenuates the aspirin-induced AERD-like responses of Ptges^−/− mice (75). Nasal polyp levels of TSLP mRNA correlate with mast cell-related transcripts and with urinary PGD₂ metabolites, and TSLP stimulates human mast cells to generate PGD₂ ex vivo (32). Tezepelumab, an anti-TSLP humanized monoclonal antibody (77), was recently FDA approved for the treatment of severe asthma, and antibodies against IL-33 and its receptor are currently under investigation (78, 79). Clinical trials are necessary to verify the roles of TSLP and IL-33 in the pathogenesis of AERD.

Aspirin desensitization.

Following a reaction to aspirin or any COX-1-active NSAID, patients with AERD are transiently refractory to the effects of aspirin and can be desensitized and escalate to a daily dose of up to 650 mg twice daily. For about 60–70% of patients with AERD, this daily aspirin dose significantly improves symptoms of sinonasal dysfunction and asthma, and can reduce the rate of polyp regrowth (80). Despite these clinical benefits, treatment with high-dose aspirin does not diminish the characteristic biomarkers of inflammation. In a prospective study of 40 individuals with AERD who underwent desensitization, treatment with high-dose aspirin for 8 weeks resulted in substantially increased urinary concentrations of LTE₄ (concomitantly with significantly reduced urinary levels of PGE₂ metabolites), along with increases in serum tryptase (indicative of mast cell activation and/or mast cell burden), higher blood eosinophil counts, and increased exhaled nitric oxide concentrations, which may be suggestive of increased airway eosinophilia (81). Notably, none of these parameters changed in aspirin-tolerant controls. These observations indicate that the mechanism(s) responsible for the clinical benefit of high-dose aspirin are likely at least partly due to a loss of end-organ reactivity to cysLTs, as reported previously (15, 82). The depletion of PGE₂, unique to AERD, may be consistent with the lack of COX-2/mPGES-1 function in their respiratory tract (83) and could account for the observed increases in tryptase, LTE₄, and nitric oxide.

Remaining unknowns and hypotheses.

Although advances have been made in our understanding of the immune mechanisms that underlie AERD, many unknowns still remain (Table I). For example, there are only two known potential risk factors for the development of the disease. Prolonged tobacco smoke exposure may be one risk, as a patient exposed to environmental tobacco smoke during both childhood and adulthood has an odds ratio of 5.09 for developing AERD (95% CI, 2.75–9.43)(84). However, a study of adults in Japan revealed that smoking cessation 1–4 years prior was positively associated with the development of AERD compared with the development of aspirin-tolerant asthma (85). Female sex is another risk factor, identified as such as early as 1968 when Samter and Beer described the first large cohort of patients with AERD, 57% of whom were female(86). More recent work has confirmed this female predominance, and has shown that on average, women develop the disease a few years earlier than men, with 6.8% of female AERD patients reporting the onset of their disease prior to 18 years of age [REF Bensko, once Epub online available]. The reasons for these sex-based differences are still unknown, although several plausible mechanisms are reported. For example, androgens suppress 5-LO translocation in granulocytes (87), and thus granulocytes from females generate significantly more leukotrienes than do those from males. Androgens also suppress the proliferation and activation of ILC2s (88). Finally, COX-2 expression and PGE₂ production are both significantly lower in women than in men (89, 90). Each of these sex differences are potentially relevant to the unique pathophysiology of AERD. However, aside from female sex, and possibly tobacco smoke exposure, other risk factors that predispose to the development of AERD remain unknown.

Table I.

Knowns and unknowns in the causative immune mechanisms in AERD.

Known or Suspected		Unknown
Risk factors for the development of AERD
	Female sex	Why is there a female predominance, and why do women tend to develop the disease earlier than men?
	Tobacco smoke exposure	What is the mechanism by which tobacco may increase risk of developing AERD?
	Age	Why is AERD so uncommon in children?
Inciting trigger for AERD
	Respiratory infection	Is there a universal bacterial or viral infection that precipitates the initial immune dysregulation?
	Autoimmunity	Is there a predisposition for the development of autoantibodies in patients with AERD?
Causes of ongoing inflammation and chronic olfactory dysfunction
	CysLTs and mast cell activation	What drives the chronic overproduction of cysLTs and mast cell activation?
	Epithelial dysfunction	Is the epithelial barrier dysregulation a primary cause of the inflammation in AERD, or a result of it?
	Elevations of inflammatory cytokines and decreased levels of PGE2 in the respiratory tissue	What are the cellular sources of these type 2 inflammatory cytokines, and what are their effects on their respective effector cell targets? Do inflammatory cytokines or low levels of PGE2 directly cause olfactory dysfunction?
Mechanisms of COX-1 inhibitor-induced reactions
	Decreased tissue levels of PGE2 and EP2/4 receptors	What drives the deficiency in respiratory PGE2 production or signaling?
	Further increase in cysLTs during reactions	Which cells drive the COX-1 inhibitor-induced rise in cysLTs?
	Variability in reaction severity and development of extra-pulmonary symptoms	Is AERD a local respiratory disease or a systemic disease that can involve the GI tract and skin as well?
Mechanisms of benefit of high-dose aspirin
	~60% of AERD patients derive benefit from high-dose aspirin (delay of polyp regrowth)	Why is daily aspirin not beneficial for all AERD patients, and what biomarkers predict who will derive benefit?
	Aspirin doses of ≥325mg can provide benefit	Why is such a high dose of aspirin needed for the respiratory benefit?
	Other non-aspirin NSAIDs do not provide respiratory benefit	Why do other COX-1 inhibitors provide no respiratory benefit?

The inciting trigger for the development of AERD is also unknown. For most patients, the disease begins quite suddenly, often described as though it was “a cold that never went away”. Decades ago, Szczeklik proposed the hypothesis that AERD may be initiated by an upper respiratory viral infection (91). While both viral (92) and bacterial pathogens, particularly Staphylococcus aureus (93, 94), have been implicated in the pathogenesis of CRSwNP, none have proven causality, and none are specific to AERD. Nonetheless, it seems plausible that infection-induced damage to the respiratory epithelium could result in exposure of the self-antigens and allow for the development of an autoimmune process. Several studies have found the presence of class-switched autoantibodies within CRSwNP tissue, with measurable autoantibody levels against a number of self-antigens, including against nuclear antigens (anti-dsDNA), B2-microglobulin, and collagen (95–97). However, whether these autoantibodies play a causative role in AERD is not known.

The cause of the severe chronic inflammation in AERD remains unknown. Further, nasal polyps in these patients have frustratingly high rates of post-operative recurrence (98), without known causes of polyp recurrence. The adult onset of AERD, along with its minimal heritability, and intransient nature are consistent with a contribution from epigenetic changes, perhaps as a result of a prior insult. In particular, epigenetic modifications may impact the control of lipid mediator biosynthesis that is fundamental to AERD, which could in turn contribute to the chronic inflammation and polyp regrowth. One study reported hypomethylation of PGDS and ALOX5AP in whole nasal polyp tissue from subjects with AERD relative to aspirin-tolerant controls, which could affect the production of PGD₂ and cysLTs, and hypermethylation of PTGES, which could affect the synthesis of PGE₂ (99). Additionally, PTGER2 expression by polyp fibroblasts from patients with AERD controlled by histone acetylation, and an epigenetically mediated reduction of EP₂ expression could contribute to the PGE₂ resistance observed in the disease (10). Still, the nature of the insult driving the development of these epigenetic changes is unknown. The profound olfactory dysfunction, often leaving patients with decades of complete anosmia, is an additional source of frustration, without clear causes. One recent study of 22 patients with AERD found that baseline severity of smell loss was correlated with lower nasal PGE₂ levels and treatment-induced improvement in olfaction related to increases in nasal PGE₂ (71). Nonetheless, a direct immunologic link between PGE₂ and olfactory function is elusive. Elevations in mucous levels of IL-2, IL-5, IL-6, IL-10, and IL-13 have also been shown to associate with loss of smell (100). However, the induction of an acute transient systemic inflammation with an experimental endotoxemia model, which does trigger a rise in some inflammatory cytokines, was not associated with a reduction in smell (101), suggesting that the cytokine associations may be correlative but not causative.

Conclusions.

Our understanding of the immune mechanisms that underlie AERD has been considerably advanced in the last decade, through investigations into the mechanistic changes induced by targeted therapies, the application of single cell genomics of the respiratory tissue, and the preclinical study of interventions afforded by the availability of transgenic animal models. The pathogenic roles of both mast cell activation and epithelial dysregulation have been further elucidated, along with the particular hypersensitivity to cysLTs and reliance on PGE₂ that is evident in AERD. The importance of type 2 cytokines as drivers of the chronic disease has also been clearly established. However, many questions remain in regard to the initial disease triggers and the causes of the persistent inflammation. As further investigations into the immunopathogenesis of AERD reveal additional clues, there may be future therapeutic options that have yet to be explored.

Abbreviations:

AERD

Aspirin-exacerbated respiratory disease

COX

Cyclooxygenase

T2I

Type 2 inflammation

CysLTs

Cysteinyl leukotrienes

PG

Prostaglandins

ILC2

Group 2 innate lymphoid cells

5-LO

5-lipoxygenase

Th2

T helper type 2

CRTH2

chemoattractant receptor-homologous molecule expressed on T-helper type 2 cells

CRSwNP

Chronic rhinosinusitis with nasal polyps

CRSsNP

Chronic rhinosinusitis without nasal polyps

TSLP

thymic stromal lymphopoietin