Aspirin sensitivity; lessons in the regulation (and dysregulation) of mast cell function

Abstract

The idiosyncratic activation of mast cells (MCs) in response to the administration of nonselective cyclooxygenase (COX) inhibitors is a cardinal feature of aspirin exacerbated respiratory disease (AERD). Older studies using mast cell stabilizing drugs support a critical role for MCs and their products in driving the severe eosinophilic inflammation and respiratory dysfunction that is typical of AERD. Since patients with AERD react to all nonselective COX inhibitors regardless of their chemical structure, the mechanism of MC activation is not due to classical, antigen-induced cross-linking of IgE receptors. Recent studies in both human subjects and animal models have revealed a complex and multifactorial process culminating in dysregulation of MC function, and an aberrant dependency on COX-1-derived prostaglandin E₂ to maintain a tenuous homeostasis. This article reviews the factors most likely to contribute to MC dysregulation in AERD, and the potential diagnostic and therapeutic implications.

Introduction

Mast cells abound in all barrier tissues, including the respiratory mucosa. Although their physiologic functions in healthy humans are unknown, animal studies strongly support a key role in the initiation and amplification of innate immune responses to microbial pathogens (1). The functions of MCs in type 1 hypersensitivity responses that underlie anaphylaxis, urticaria, and acute allergen-driven episodes of rhinitis and atopic asthma is well-established. MC-derived products such as histamine, cysteinyl leukotrienes (cysLTs), prostaglandin D₂ (PGD₂), cytokines and chemokines have functions in allergic diseases that are validated or strongly supported by experimental evidence (2). The ability of blocking antibodies against the Fc portion of IgE to block allergen-induced early and late phase responses in the airways of atopic asthmatics (3) further validates the importance of this classical activation pathway.

Aspirin exacerbated respiratory disease (AERD) is a distinctive syndrome, typically of adult onset, that is characterized by severe eosinophilic sinonasal disease with recurrent nasal polyposis, asthma (which is often but not uniformly severe), and pathognomonic clinical reactions to the ingestion of aspirin or any other drug that inhibits the enzymatic function of cyclooxygenase (COX)-1 (4). It affects ~7% of all asthmatic subjects in the United States and as many as 14% of subjects with severe asthma (5). The pathophysiology of AERD involves dysregulated MC function, and MC-derived products contribute substantially to the baseline immunopathology of the disease (6). Moreover, the pathognomonic clinical reactions to COX-1 active drugs are associated with marked activation of MCs by mechanisms that are poorly understood but are not due to IgE-dependent immune recognition of the drug (7–9). This review summarizes the potential significance of MCs in AERD, and the recent studies that highlight potential mechanisms responsible for their dysregulation and activation in this context.

Evidence to support a key role for MCs in AERD.

When activated, MCs generate LTC₄, the parent of the cysLTs, through the 5-lipoxygenase (5-LO)/LTC₄ synthase (LTC₄S) pathway (10). This pathway is shared with eosinophils, basophils, and monocytes. MCs also metabolize arachidonic acid through the COX pathway to generate especially large quantities of PGD₂ (11). CysLTs and PGD₂ synergize to activate human innate group 2 lymphoid cells (ILC2s) and T helper type 2 (Th2) cells. CysLTs are the most potent known constrictors of human airways (12), elicit mucous secretion (13), and activate IL-25-expressing airway brush cells (14). The terminal cysLT metabolite, LTE₄, can be measured in the urine, as can stable metabolites (PGD-M) of PGD₂, as time-weighted reflections of LTC₄ and PGD₂ generation, respectively. In AERD, urinary levels of both LTE₄ and PGD-M exceed those in aspirin tolerant controls by several fold (15, 16), consistent with ongoing MC activation. The administration of aspirin induces sharp additional increases in both mediators, accompanied by increases in plasma histamine (which may also reflect concomitant basophil activation) and (in some individuals) serum tryptase, confirming MC activation. The numbers of MCs are increased in the bronchial biopsies and sinonasal mucosa of patients with AERD compared with aspirin tolerant controls (17, 18). Thus, AERD involves relative MC hyperplasia, persistent steady-state MC activation (“leaky MCs”), and marked unbraking of MC activation with COX-1 blockade. Importantly, the unbraking phenomenon occurs with all drugs (regardless of structure) that interfere with COX-1, indicating that reactions are not manifestations of IgE-dependent “drug allergy”, but rather reflect an idiosyncratic dependency on one or more COX products to maintain a tenuous homeostasis over MCs and other effector cells.

Several older studies using classical MC-stabilizing cromone drugs strongly suggest that both the steady-state MC “leak” and the incremental activation with COX-1 inhibition contribute substantially to AERD pathophysiology. Although the administration of cromolyn has no acute effect on baseline lung function measurements in aspirin tolerant asthmatic control subjects, it increased FEV1 in subjects with AERD within minutes of administration (19). In another study, treatment with cromolyn for one week significantly decreased the numbers of sputum eosinophils and sputum levels of eosinophilic cationic protein in subjects with AERD (20). The administration of cromolyn or nedocromil to subjects with AERD blocks the reduction in FEV1 that occurs with aspirin challenges (21), as well as the accompanying increase in urinary LTE₄ levels (22). Collectively, these studies strongly support the importance of MCs and their mediators in both the steady-state respiratory dysfunction and the clinical reactions to COX-1 inhibition in AERD, emphasizing the importance of understanding the idiosyncratic mechanisms responsible for the dysregulated function of MCs in this disease.

Factors contributing to MC activation in AERD.

1. Dysregulation of the prostaglandin E₂/EP₂ receptor system.

PGE₂ is a ubiquitous COX product with MC-stabilizing properties. It is generated constitutively at steady-state and inducibly at very high levels during inflammatory responses. Epithelial cells, fibroblasts, and macrophages all generate substantial quantities of PGE₂ when stimulated ex vivo with IL-1β or LPS. This response reflects inducible expression of COX-2 and microsomal PGE₂ synthase-1 (mPGES-1), the principal terminal synthase responsible for generating PGE₂ during inflammatory responses. Compared with constitutively expressed COX-1, COX-2 is much less sensitive to inhibition by aspirin (23). Consequently, PGE₂ derived from COX-2/mPGES-1 during inflammatory responses may be maintained even in the face of low dose aspirin. PGE₂ signaling through the E prostanoid 2 (EP₂) receptor blocks MC degranulation in response to IgE receptor cross-linking (24), and dampens leukotriene production by inducing protein kinase A-dependent phosphorylation of 5-LO (25). Thus, PGE₂ may function to restrain MC activation and cysLT generation during inflammatory responses.

Impaired function of the COX-2/mPGES-1 system may potentiate the risk of AERD (Fig. 1). Nasal polyp levels of PGE₂ are markedly lower than in non-polypoid sinonasal tissue (26), and fibroblasts from nasal polyps show weak induction of COX-2 and mPGES-1 expression compared with cells from nonpolyp controls (27). These levels tend to be lowest in samples from subjects with AERD (28). Impaired expression of COX-2 and/or mPGES-1 could render the maintenance of local PGE₂ generation disproportionately dependent on COX-1, predisposing to depletion below a critical threshold when aspirin is administered. Indeed, clinical reactions generally occur at low threshold doses (80-160 mg) of aspirin that are typically selective for COX-1. The administration of a single dose of PGE₂ by inhalation blocks aspirin-induced changes in lung function and increases in urinary LTE₄ in subjects with AERD (29), consistent with a major role for endogenous PGE₂ as the critical COX-1-derived brake. The effect of depleting PGE₂ with COX-1 inhibition may be potentiated by reduced levels of EP₂ receptor expression by MCs (and other 5-LO-expressing leukocytes) in subjects with AERD, as suggested by immunohistochemical staining of sinonasal and lung tissues (30, 31). Mice lacking either mPGES-1 or EP₂ receptors (Ptges^−/− and Ptger2^−/− mice, respectively) display AERD-like features, including aspirin-induced mast cell activation and increases in cysLTs, supporting the likely importance of perturbations in these systems as disease-causing lesions (32). Some studies support an epigenetic basis for these perturbations (33, 34), consistent with the acquired nature and persistence of AERD.

Figure 1. Biosynthesis of PGE₂ and its potential dysregulation in AERD.

A. Under physiological conditions, PGE₂ derives from either constitutively expressed COX-1 or inducibly expressed COX-2. COX-derived PGH2 can be converted to PGE₂ by three different terminal PGE₂ synthase (PGES) enzymes; cytosolic (c)PGES, and microsomal (m)PGES-1 and -2. During inflammation, the concomitantly induced expressions of COX-2 and mPGES-1 markedly increase the production of PGE₂. Because COX-2 is much less sensitive to aspirin than is COX-1, this increase is relatively resistant to low dose aspirin. PGE₂ elicits signaling through EP receptors 1-4. EP₂ and EP₄ activate the cAMP/PKA pathway that stabilizes MCs and blocks leukotriene formation by phosphorylating 5-LO. B. In AERD, the impaired expression and function of the COX-2/mPGES-1 system renders PGE₂ production disproportionately dependent on COX-1 and susceptible to inhibition by low dose aspirin. The effects of PGE₂ depletion is potentiated by diminished EP₂ receptor expression, leading to unbraking of MCs and 5-LO.

Whereas low-dose aspirin-induced depletion of COX-1-derived PGE₂ likely permits the triggering of MC activation during provocative challenges, high dose aspirin (325-650 mg twice daily, a dose sufficient to block COX-2) following desensitization can paradoxically improve sinonasal function and reduce the frequency of polyp recurrent in patients with AERD (35). Cahill and colleagues (36) demonstrated that patients with AERD on high dose aspirin display significantly increased total serum tryptase levels, as well as elevated urinary levels of LTE₄, compared with their pre-challenge baselines, suggesting paradoxical amplification of MC “leak” even in the face of substantial clinical improvement on aspirin. This increase in MC activation markers was paralleled by sharply reduced levels of PGE₂ metabolites in the urine, again consistent with the dependence of MCs on PGE₂ to maintain a tenuous homeostasis in AERD. Since the tryptase was not fractionated in this study, it is not currently possible to ascertain whether ongoing MC degranulation is increased, or whether depletion of PGE₂ dysregulates transcriptional control of tryptase synthesis. Notably, while PGD₂ production increases during reactions, the desensitized state on high dose aspirin is accompanied by sharp reductions in urinary PGD-M levels (7). Nasal polyp MCs express markedly higher levels of mRNA encoding COX-2 than COX-1 (16), potentially explaining why PGD₂ production resists inhibition during provocative challenges until the aspirin dose is escalated. The suppression of PGD₂ production (with its attendant activating and chemotactic effects for eosinophils, basophils, group 2 innate lymphoid cells (ILC2s) and Th2 cells) could contribute to the therapeutic benefit of aspirin (7).

2. Dysregulation of cysLT synthesis and cysLT receptor function.

CysLTs contribute substantially to the baseline pathophysiology of AERD (37), and are essential to the pathognomonic reactions to aspirin (38). The dysregulated synthesis of the cysLTs in AERD (15) is paralleled by strongly upregulated expression of the type 1 cysLT receptor (CysLT₁R) on hematopoietic cells, including MCs, in the airways (39), suggesting a mechanism driving a state of end-organ hyperreactivity to endogenous cysLTs. Indeed, subjects with AERD display a markedly higher sensitivity to bronchoconstriction induced by inhaled LTE₄ than do aspirin tolerant asthmatic controls (40). Both hyperreactivity to LTE₄ and overexpression of CysLT₁R are reduced toward levels seen in aspirin tolerant controls after treatment with high dose aspirin (39, 40). A study by Fischer and colleagues strongly suggested that endogenous cysLTs are required for the idiosyncratic MC activation mechanism that characterizes AERD. In this study, the administration of the 5-LO inhibitor zileuton not only blocked incremental sinonasal symptoms in response to the challenge, but prevented the increase in nasal lavage fluid levels of tryptase (41). Although endogenous cysLTs are not required for IgE-dependent degranulation of MCs in ex vivo systems (42), exogenous cysLTs are sufficient to elicit calcium flux, cytokine production, and PGD₂ generation from human cord blood MCs (42, 43). Moreover, a recent study demonstrated that inhalation of LTE₄ by asthmatic (mostly aspirin-tolerant) subjects induced markedly increased urinary levels of PGD₂, indicative of MC activation in vivo. This response was blocked by the CysLT₁R antagonist montelukast (44). While cysLTs are very likely to signal directly to MCs to induce their activation, animal models suggest additional effects related to cysLT-dependent control of MC-active cytokines (see below).

3. The role of IgE and FcεRI.

While the rates of atopy in subjects with AERD vary depending on the study (45, 46), the disease frequently occurs in subjects who lack IgE with specificity to inhalant allergens (some of whom nonetheless have total IgE above the normal range)(47). A study by Taniguchi and colleagues demonstrated that treatment of patients with AERD with recombinant humanized anti-IgE (Omalizumab) in an open label protocol markedly improved symptomatic control of both asthma and sinonasal disease, while decreasing the urinary levels of LTE₄ and PGD-M by >75% each (48). In a small placebo-controlled trial, Lang and colleagues demonstrated that omalizumab treatment for 16 weeks prevented clinical symptoms and increases in urinary LTE₄ in 5 of 7 subjects challenged with aspirin (49), whereas all four placebo-treated subjects experienced reactions. Thus, despite the lack of evidence implicating classical allergen-induced MC activation in AERD physiology, these studies suggest that ongoing signaling through MC-associated IgE receptors permit and amplify activation responses to depletion of PGE₂. The nature of the antigens driving the IgE production remains to be determined, though local production of antibodies against Staphylococcal antigens (50) and autoantigens (51) have been reported in nasal polyps. It is unknown whether the levels of these antibodies distinguish AERD subjects from aspirin tolerant controls.

4. The role of innate cytokines.

MCs express receptors for both IL-33 and thymic stromal lymphopoietin (TSLP). Both cytokines are abundant in nasal polyps (52), with mRNA for each localizing principally to epithelial basal cells (53). IL-33 potently induces the production of type 2 cytokines and chemokines by MCs ex vivo (54), and strongly induces COX-2-dependent production of PGD₂ (16, 55). TSLP strongly potentiates these IL-33-induced effects (16, 54), suggesting that these resident tissue-derived cytokines may synergize to influence MC function in vivo. In AERD, the levels of TSLP expression in the sinonasal tissue correlate strongly with urinary levels of PGD-M (and with MC-specific transcripts such as tryptase, carboxypeptidase A3, and hematopoietic PGD₂ synthase), suggesting that TSLP conditions MCs for augmented function in this disease (16). One study reported that the quantities of IL-33 protein in nasal polyp extracts from patients with AERD substantially exceeded those in non-AERD control samples (56).

IL-33 may be especially important in regulating MC function in AERD. In AERD-like Ptges^−/− mice, antibody-mediated blockade of IL-33 or blockade of suppressor of tumorigenicity 2 (ST2, the IL-33 receptor) with a soluble Fc fusion protein completely prevents MC activation (with attendant release of proteases, production of cysLTs and PGD₂) as well as the increases in airway resistance occurring in response to aspirin challenges (56). These mice display high levels of lung IL-33 protein following the induction of airway inflammation by house dust mite allergen. This induced expression of IL-33 in Ptges^−/− mice requires endogenous cysLTs and signaling through the type 2 cysLT receptor (CysLT₂R) (57). Moreover, cysLTs not only drive IL-33 expression by alveolar type 2 cells, but also induce the recruitment and activation of IL-33+ platelets, the latter of which are necessary for the MC response to aspirin challenge. Platelets release substantial amounts of pre-formed, cytosolic IL-33 in response to stimulation by LTC₄ (58). Notably, large numbers of extravasated platelets are present in nasal polyps from patients with AERD compared with aspirin tolerant controls (59), and stimulation of freshly excised human nasal polyp tissue with LTC₄ ex vivo results in the release of substantial amounts of IL-33 (58). It is tempting to speculate that the high levels of IL-33 protein found in AERD polyps may partially reflect this cysLT-driven platelet store.

Notably, while IL-33 alone is insufficient to elicit degranulation of MCs ex vivo, it does cause degranulation of mouse MCs that are passively sensitized with IgE, and can cause anaphylaxis in mice when IgE is present (60). This finding suggests a potential interaction between IL-33/ST2 and IgE-driven pathways that are permissive for the leaky MC phenotype in AERD, and may help to explain both the incremental degranulation of MCs with aspirin challenge and therapeutic responses to omalizumab.

Pathophysiologic and therapeutic implications

MC activation in AERD appears to be driven by complementary dysregulation of systems that respectively induce stabilization (PGE₂/EP₂) or activation (IgE, IL-33, cysLTs) of MCs (Fig. 2). Blockade of 5-LO with zileuton provides incremental improvement in sinonasal function for patients already treated with glucocorticoids (37). It seems possible that the marked effect of zileuton on signs, symptoms, and biomarkers of reactions to aspirin may reflect inhibition of both direct (CysLT₁R-mediated) (44) and indirect (CysLT₂R-driven, IL-33-mediated) cysLT-dependent pathways that facilitate MC activation. The potential efficacy of dual-specific antagonists of CysLT₁R and CysLT₂R could exceed that of the existing CysLT₁R-selective drugs. Stable PGE₂ analogues that selectively stimulate the EP₂ receptor could inhibit the activation of MCs (and other EP₂ receptor-bearing effector cells) and limit cysLT production without the cough-inducing effects that preclude the use of PGE₂ itself, though reduced expression of EP₂ may limit efficacy in some individuals. The encouraging results of omalizumab in small trials (48, 49) suggests a need for larger studies, and compels further study of the antigens potentially responsible for driving IgE synthesis. The high levels of PGD₂ production in AERD suggests that antagonists of the type 2 PGD₂ receptor (DP₂, also known as CRTH2) may have efficacy by suppressing tissue eosinophilia. Biologics under development that target IL-33 and ST2 should provide opportunities to validate potential causative pathways implicated in experimental models, as well as potential efficacy. Lastly, the MC-depleting tyrosine kinase inhibitor imatinib was recently shown to improve lung function and airway reactivity in a population of refractory asthmatics not selected for the presence or absence of AERD (61). While imatinib has other potential targets (e.g., platelet-derived growth factor receptors) that may account for its benefits independent of MCs, it is possible that it may offer efficacy in AERD given the strong MC activation signatures associated with the disease.

Figure 2. Hypothetical factors contributing to dysregulation of MC activation in AERD.

Occupancy of FcεRI, increased signaling through CysLT₁R, and locally derived innate cytokines (TSLP and IL-33) may conspire to promote an ongoing “leak” of PGD₂ and other MC products that facilitate eosinophilic pathology and baseline respiratory tissue dysfunction in AERD. PGE₂ -dependent maintenance of homeostasis is compromised by poor COX-2/mPGES-1 and EP₂ receptor function. Aspirin and other COX-1 inhibitors remove the brake provided by residual PGE₂, markedly destabilizing the system to permit marked MC activation and associated end-organ responses during pathognomonic reactions.

Bullet points:

• Diminished levels of COX-2 and mPGES-1 expression likely impair PGE2 generation in AERD and render residual PGE₂ production highly dependent on COX-1

• Reduced levels of EP₂ receptor expression on mast cells and other effector cells may exacerbate the effects of depleting PGE2 with aspirin and other COX-1 active drugs.

• The high levels of cysLT production in AERD are accompanied by increased levels of CysLT₁R expression on hematopoietic cells.

• Desensitization to aspirin downregulates CysLT₁R expression and decreases end-organ reactivity to cysLTs, which may help explain clinical improvement despite substantial further increases in cysLT generation.

• Local IgE produced against Staph or autoantigens in nasal polyps may synergize with IL-33 to promote ongoing MC activation, which is likely exaggerated in AERD due to reduced PGE₂/EP₂ signaling and enhanced cysLT generation.

What we do know:

Mast cells contribute substantially to both the steady-state respiratory dysfunction and to the pathognomonic reactions to aspirin and other COX-1-active drugs in AERD

PGE2 maintains an essential brake on mast cell activation

Local IgE and endogenous cysteinyl leukotrienes contribute to the idiosyncratic activation mechanism in AERD

What we do not yet know:

The role of innate type 2 cytokines (IL-33, TSLP) in conditioning mast cells for activation in AERD

The molecular basis of the impaired function of the endogenous PGE2/EP2 receptor system

Abbreviations used in this article:

AERD

aspirin exacerbated respiratory disease

COX

cyclooxygenase

cysLTs

cysteinyl leukotrienes

CysLT₁R

type 1 cysLT receptor

CysLT₂R

type 2 cysLT receptor

DP2

type 2 receptor for PGD₂

EP₂

E prostanoid 2 receptor

FcεRI

high affinity receptor for immunoglobulin E

FEV1

forced expiratory volume in 1 second

IL-1β

interleukin 1 β

IL

interleukin

ILC2

group 2 innate lymphoid cell

LT

leukotriene

LTC₄S

leukotriene C₄ synthase

MCs

mast cells

mPGES-1

microsomal PGE₂ synthase-1

PG

prostaglandin

PGD-M

stable urinary PGD₂ metabolite

ST2

suppressor of tumorigenicity 2 (IL-33 receptor)

Th2

T helper type 2

TSLP

thymic stromal lymphopoietin

5-LO

5-lipoxygenase