Aspirin Intolerance: Experimental Models for Bed-to-Bench.

Aspirin is the oldest non-steroidal anti-inflammatory drug (NSAID), and it sometimes causes asthma-like symptoms known as aspirin-exacerbated respiratory disease (AERD), which can be serious. Unwanted effects of aspirin (aspirin intolerance) are also observed in patients with food-dependent exercise-induced anaphylaxis, a type I allergy disease, and aspirin-induced urticaria (AIU). However the target and the mechanism of the aspirin intolerance are still unknown. There is no animal or cellular model of AERD, because its pathophysiological mechanism is still unknown, but it is thought that inhibition of cyclooxygenase by causative agents leads to an increase of free arachidonic acid, which is metabolized into cysteinyl leukotrienes (cysLTs) that provoke airway smooth muscle constriction and asthma symptoms. As the bed-tobench approach, to confirm the clinical discussion in experimental cellular models, we have tried to develop a cellular model of AERD using activated RBL-2H3 cells, a rat mast cell like cell line. Indomethacin (another NSAID and also causes AERD), enhances in vitro cysLTs production by RBL-2H3 cells, while there is no induction of cysLTs production in the absence of inflammatory activation. Since this suggests that all inflammatory cells with activation of prostaglandin and cysLT metabolism should respond to NSAIDs, and then I have concluded that aspirin intolerance should be separated from subsequent bronchoconstriction. Evidence about the cellular mechanisms of NSAIDs may be employed for development of in vitro AERD models as the approach from bench-to-bed.

CLINICAL ASPECTS OF ASPIRIN INTOLERANCE

Aspirin Intolerance and Clinical Aspects of AERD

Aspirin-exacerbated respiratory disease (AERD) is also known as aspirin-induced asthma or aspirin-intolerant asthma (AIA). It is a respiratory disorder that features nasal obstruction, rhinorrhea, and acute asthma attacks with airway constriction, which is induced by non-steroidal anti-inflammatory drugs (NSAIDs) such as aspirin (acetylsalicylic acid; Fig. ) [1-3] or indomethacin (Fig. ) [4], other medications [5] such as sodium succinate-conjugated corticosteroids (e.g. hydrocortisone sodium succinate; Fig. ) food additives such as tartrazine (a yellow food dye; Fig. ) or parahydroxybenzoates (antiseptic food additives known as paraben; Fig. ), and food components such as sulfites, mint, and salicylate (Fig. ) [6, 7]. Hypersensitivity to NSAIDs and these causative substances is known as aspirin intolerance.

According to the asthma guideline of the US National Heart, Lung and Blood Institute, 21% of adult asthma patients and 5% of pediatric asthma patients have AERD [8], while the Japanese asthma guideline [9] states that 5-10% of adult asthma patients and 3.6-7.8% of pediatric asthma patients have AERD. According to both guidelines, the majority of AERD patients are adults.

Females are affected twice as often as males, but other factors such as the family history, ethnicity, or geographic region have no influence [10], suggesting that genetic disorders are not related to the pathogenesis of AERD. Nasal polyps and dysosmia are frequent in AERD patients [9, 11, 12].

Asthma attacks triggered by aspirin intolerance can be serious, and it was reported that 40% of patients receiving emergency treatment for asthma have AERD [13, 14]. Since AERD has been reported in patients without known prior exposure to the triggering compound, it seems that the underlying mechanism is not type I allergy [10]. Evidence shows that treatment with chromoglycate is effective for AERD [15], suggesting that mast cells may play a pivotal role in this condition.

Aspirin Intolerance in Other Diseases

Aspirin intolerance is also observed in a condition called aspirin-induced urticaria (AIU), in which urticaria and/or angioedema develops at 1-6 h after exposure to COX-1-inhibiting NSAIDs in 12% of patients with chronic urticaria [16-18]. AERD and AIU account for the majority of patients with aspirin intolerance.

Harada et al. [19] reported that aspirin intolerance was observed in all of their patients with food-dependent exercise-induced anaphylaxis (FDEIA), a condition combining food allergy and respiratory disorder that is mostly related to wheat or crustaceans. FDEIA differs from AERD, since many patients are teenage boys and 40% of them have atopic diseases, suggesting that FDEIA is associated with type I allergy. While 10% of patients develop asthma attacks with exercise several hours after intake of a causative food, they have no symptoms if they do not exercise [20].

These differences between AERD and FDEIA may indicate that the type of allergy (atopic or non-atopic) is not important, or may suggest that non-atopic immune activation underlies the atopic characteristics of FDEIA. Accordingly, it is possible that aspirin intolerance should be separated from the subsequent bronchoconstriction in AERD or FDEIA.

PATHOPHYSIOLOGICAL ASPECTS OF AERD

Arachidonic Acid Metabolism and AERD

As shown in Fig. (), the enzyme phospholipase A2 releases fatty acids from cell membrane phospholipids. Arachidonic acid is one of the fatty acids released and it is metabolized into various substances, including prostaglandins (PGs), leukotrienes (LTs), and thromboxanes (TXs), which are thought to make a major contribution to the pathogenesis of inflammatory diseases.

PGD2 and PGE2 are arachidonic acid metabolites produced by cyclooxygenase (COX, also called prostaglandin G/H synthase, PGHS, EC 1.14.99.1) which metabolizes arachidonic acid to PGG2 by its cyclooxygenase activity and then metabolizes PGG2 to PGH2 by its hydroperoxidase activity (Fig. ). Metabolites of COX are known to contribute to inflammation. In 1990s, two subtypes of the COX enzyme were found. One of these was named COX-1 and was found to be constitutively expressed by cells. The other was named COX-2 [21, 22], and this was found to be induced by physiological and experimental inflammatory stimuli, such as the carcinogenic promoter 12-O-tetradecanoylphorbol 13-acetate, thapsigargin, or calcium ionophore A23187 [23-26]. Transcription of COX-2 protein is controlled by various transcription factors, including nuclear factor-κB (NF-κB) and AP-1 [27-31], via degradation of the inhibitory protein, IκB [32]. Specific COX-2 inhibitors were developed to avoid the side effect of gastrointestinal ulceration, and it was revealed that selectivity is due to difference of tertiary protein structure between COX-1 and COX-2 [33-36].

Arachidonic acid is also metabolized to leukotriene A4 (LTA4) by another pathway involving 5-lipoxygenase (5-LOX), after which LTA4 is converted to LTB4 and LTC4. Then LTC4 is metabolized to LTD4, and LTE4 as shown in Fig. (). LTC4, LTD4, and E4 contain cysteine residues, and thus are called cysteinyl leukotrienes (cysLTs). Cysteinyl LTs were originally discovered as slow reacting substance of anaphylaxis (SRS-A), which was extracted from the lung tissues of antigen-sensitized guinea pigs, and was shown to constrict airway smooth muscle from these animals more potently, slowly, and continuously than histamine via an antihistamine- resistant mechanism [37].

Urinary concentrations of cysLTs are elevated in AERD patients, even when they have no asthma symptoms [38], and cysLT inhibitors, such as 5-LOX inhibitor [39], or cysLT receptor blockers [40-42] are reported to be safe and effective for AERD, indicating involvement of cysLTs in the mechanism underlying this disease.

Target of Aspirin/NSAIDs and Mechanism of AERD

Aspirin is an anti-inflammatory analgesic agent that has been marketed since 1899 and is the oldest chemically synthesized drug (Fig. ). In the 1970s, it was found that the mechanism of action of aspirin and other acidic NSAIDs involves inhibition of PG production [43-45]. In the 1990s, the molecular mechanism of aspirin was finally revealed to be irreversible acetylation of Ser-530 on COX [46], with inhibition of both COX-1 and COX-2 [23], while most NSAIDs such as indomethacin (Fig. ) reversibly inhibit COX.

Almost all (80-100%) the oral dose of aspirin (100-300 mg) is immediately absorbed from the stomach or upper small intestine and the maximum serum concentration is reached within 35 min. Then it is metabolized into salicylic acid (Fig. ) and conjugated with a glucuronate or sulfate. A sodium salt of the one of the metabolites, gentisic acid (Fig. ), is reported to have an anti-inflammatory effect [47].

Interestingly, selective COX-2 inhibitors have been reported to be safe for AERD patients [48-51], indicating the pivotal involvement of COX-1 in the mechanism of AERD.

In 1990s, it was hypothesized that suppression of prostaglandin production via inhibition of COX leads to accumulation of excess arachidonic acid both inside and outside of cells, after which the excess arachidonic acid is metabolized into cysLTs that cause asthma attacks [38, 52-55].

BED-TO-BENCH APPROACH OF THE EXPERIMENTAL MODELS WHICH PROPOSING OR CONFIRMING THE PATHOPHYSIOLOGICAL MECHANISMS OF AERD

The pathogenic mechanism of AERD is still not clear. Most of the research performed has been clinical because useful animal or cellular models do not exist. Some attempts to develop experimental models of AERD are ongoing based on clinical information.

Cellular Model of Arachidonic Acid Metabolism Using a Cultured RBL-2H3 Mast Cell Line

We have tried to develop a cellular model of AERD based on inhibition PG production by NSAIDs, and we previously assessed the effect on cysLT production in the RBL-2H3 mast cell line derived from rat basophilic leukemia [56]. We choose indomethacin (≦3 µM, Fig. ) as the NSAID to avoid unexpected protein acetylation by using aspirin [57]. We found that cysLT production increased when RBL-2H3 cells were pretreated overnight with dinitrophenol (DNP)-specific immunoglobulin E (IgE) and then treated with DNP-conjugated human serum albumin, creating a cellular model of type I allergy [58-60]. Treatment with indomethacin at concentration up to 3 μM did not cause dose-dependent changes of cysLT production, while there was almost 90% inhibition of PGD2 production. LTB4 production was significantly increased by indomethacin at 1-3 μM. Our observation regarding LTB4 is supported by a study of Planaguma et al. [61] using rat Kupffer cells; adding aspirin (up to 5 mM) doubled LTB4 production while inhibiting PGE2 production approximately 60%.

Then we tried adding arachidonic acid (1-10 μM) with DNP-HSA to the culture medium of RBL-2H3 cells. Cysteinyl LT production was significantly increased by adding arachidonic acid combined with allergic stimulation, while there was no significant change without allergic stimulation [57].

As mentioned above, AERD is not thought to be a type I allergic disease, so we also tested non-allergic stimulation of RBL-2H3 cells with calcium ionophore A23187 and obtained the similar results.

Our observations indicate that induction of cysLT production by indomethacin may require some inflammatory stimulation of mast cells. If AERD is fundamentally a problem with arachidonic acid metabolism, it leads to another question: “Why don't all patients with type I allergy who show activation of mast cells and increased cysLT production have aspirin intolerant?”

NSAIDs Increase cysLT Production by Mast Cells

In our experiments described above [57], we obtained the interesting result that 0.1 μM of indomethacin slightly, but significantly increased the production of LTC4, and LTE4. We also tested indomethacin from another source to avoid the influence of contamination by other factors, and the same result was observed.

Togo et al. [62] reported that aspirin and salicylate (both at 100-300 μM) increased LTC4 production by the RBL-2H3 mast cell line and mouse bone marrow-derived mast cells, and suggested that the mechanism involved calcium influx following activation of cytosolic phospholipase A2. We cannot conclude whether their observation corresponds with ours or not, and it is also unclear whether the effect of these NSAIDs is related to AERD.

Possible Involvement of Cytokines in PG Production

Liu et al. [63] studied mice with knockout of microsomal PGE synthase-1 (mPGES-1), which associates with COX-2 and metabolizes PGH2 to PGE2 (Fig. ). After pretreatment with dust mite extract and challenge with lysine-aspirin (1.2 mg/12 µL via a ventilator), there was a marked increase of respiratory tract resistance and elevation of the levels of cysLTs, histamine, and MCP-1 (an indicator of mast cell degranulation) in bronchoalveolar lavage fluid. Ketorolac (a selective COX-1 inhibitor) and celecoxib (a selective COX-2 inhibitor) failed to increase respiratory tract resistance in this model. However, an EP2 receptor agonist significantly inhibited these responses, and EP2 receptor knockout mice showed smaller increases than mPGES-1 knockout mice. They recently observed that interleukin (IL)-33, a type 2 helper T (Th2) cytokine released from necrotic cells as “alarmin” [64], was highly expressed in the lung epithelium of mPGES-1 knockout mice pretreated with mite extract and in the nasal polyps of AERD patients [65, 66]. It was reported that IL-33 is produced by mast cells and that it increases the production of IL-4, IL-5, and IL-6 by bone marrow-derived murine mast cells and a murine mast cell line (MC/9) when these cells received IgE sensitization followed by antigen treatment. It was also reported that the IL-33 receptor (ST-2) is expressed by mast cells after atopic stimulation [67].

Another class of ST-2 positive memory-type Th2 cells, known as pathogenic Th2 cells, was recently proposed to have a role in inflammatory diseases including asthma [68]. Though the relationship between the pathogenic Th2 cells and AERD is unclear, T cells or their cytokines could activate mast cells in AERD patients.

The observation that mPGES knockout mice respond to lysine-aspirin [63] may indicate that inhibition of cytosolic PGES (cPGES) associated with COX-1 (Fig. ) could be at least partly involved in AERD, suggesting that cPGES activators could be effective for this condition.

Nakatani et al. [69] studied the effect of cPGES knockout, and reported that cPGES-null mice undergo perinatal death with low body weight, hypoplastic skin changes, and alveolar collapse in the lungs at 18.5 days post-coitum. In contrast, heterozygous knockout mice seemed normal, apart from the cPGES protein level. They did not provide information about the effect of NSAIDs on cPGES heterozygous knockout animals.

We previously observed that auranofin, a gold compound for rheumatoid arthritis, induced a COX-1-dependent increase of PGE2 production without significant induction of COX-1 protein within 4 h after inflammatory treatment of rat peritoneal macrophages, while nitric oxide production was not increased [25, 26, 31]. It was also reported that auranofin induces p23, a co-chaperone of heat shock protein 90 that was later identified as cPGES (Fig. ) [70]. If reduction of PG synthesis via cPGES/COX-1 plays an important role in AERD, auranofin or other cPGES activator(s) could be effective for this condition.

Findings in our cellular model suggested that mast cells may require inflammatory activation for overproduction of cysLTs with COX inhibition. Some classes of T cells and some cytokines may be involved the pathogenesis of AERD by activating mast cells. Therefore, in vitro experiments may be able to identify inflammatory mediators that activate mast cells and PGES activators, with the results having clinical implications.

Other Targets of Aspirin, the Bench-to-Bed Approach

The mechanism through which NSAIDs act on COX is very clear and explains many inflammatory phenomena, but we could not explain all the effects of NSAIDs through inhibition of COX activity.

As described above, some causative substances of AERD may not strongly inhibit COX-1-dependent PG synthesis. We previously observed that the ligands of two well-known nuclear receptors, peroxisome proliferator-activated receptor (PPAR) α and γ, independently reduced cysLT release from stimulated RBL-2H3 mast cell line [58-60]. This may suggest the possibility of a role of some unknown target(s) of NSAIDs and other causative substances.

Aspirin is reported to be associated with phospholipase A2 [71], which is the enzyme that liberates arachidonic acid from membrane phospholipids (Fig. ).

Aspirin and other NSAIDs are also reported to affect various mitochondrial enzyme complexes, including NADH-ubiquinone oxidoreductase (complex I), succinate dehydrogenase (complex II), cytochrome bc 1 complex (complex III) and cytochrome c oxidase (complex IV) [72, 73]. Antimycin, a specific complex III inhibitor, is reported to inhibit histamine release from experimentally stimulated human mast cells obtained by bronchoalveolar lavage [74]. Antimycin was also reported to increase reactive oxygen species in the mitochondria of a pollen extract-treated A549 human alveolar epithelial cell line and the authors suggested that damage to mitochondrial respiration-related proteins may exacerbate inflammation [75].

I found that the amino acid sequence of the acetylated by aspirin in COX (FSLKG from 529 to 533) [46], is also in the primary protein structure [76] of human corona virus HKU1 near the nonstructural protein (nsp) 4 domain which has a similar amino acid alignment to nsp4 of murine hepatitis virus. I could not find a nomenclature for this 5 amino acid site in HKU1, and I cannot predict whether aspirin and other NSAIDs act via this site in humans.

Salicylic acid (Fig. ) is secreted by plants in response to pathogens and induces resistance to viruses [77]. The salicylate receptors in plants are nonexpressors of PR gene 1 (NPR1), which is a homolog of IκB in mammals [78], and has NPR1-like proteins 3, and 4 as paralogues [79, 80]. In tumor necrosis factor-treated COS-1 cells, it has been demonstrated that sodium salicylate (1-20 mM) inhibits the degradation of IκBα, but not IκBβ, via inhibition of IκB kinase activity [81, 82].

Recently, Honjo et al. reported the discovery of 36 genes related to thermal nociception in Drosophila larvae [83]. Twenty of 36 genes have human orthologs, and some of them are not known to be related to pain sensation. This is interesting because pain sensation in the respiratory tract may be related to cough in asthma patients. PGE2 is known to enhance pain, which is why NSAIDs are analgesics. It is possible that PGE2 or aspirin acts directly or indirectly on these gene products to enhance nociceptive sensation in inflammation.

CONCLUSION

In traditional Japanese medicine, a disease with a high risk of occurrence in the future, is called “mibyou” [84, 85]. I think that AERD patients who have never experienced an asthma attack could be considered as “mibyou” patients. While asthma attacks associated with AERD are frequently serious, we cannot define whether a person is at risk of AERD or not until asthma occurs following the intake of a causative substance.

If we could determine the mechanism of AERD and identify the risk of asthma during the “mibyou” period, we could prevent AERD and reduce the number of asthma deaths. Further clinical investigations of AERD are needed, and additional basic studies based on the clinical information could also help clinicians to understand the pathophysiological features of this disease.