Local and Systemic Production of Pro-Inflammatory Eicosanoids Is Inversely Related to Sensitization to Aeroallergens in Patients with Aspirin-Exacerbated Respiratory Disease

Aspirin-exacerbated respiratory disease (AERD) is characterized by overproduction of the pro-inflammatory eicosanoids. Although immunoglobulin E-mediated sensitization to aeroallergens is common among AERD patients, it does not belong to the defining disease characteristics. In this study of 133 AERD patients, we sought to find a relationship between sensitization to aeroallergens and local (leukotriene E4, prostaglandin E2 and prostaglandin D2) and/or systemic (leukotriene E4) production of arachidonic acid metabolites. Interestingly, a negative association between pro-inflammatory eicosanoid levels in induced sputum supernatant or urine and sensitization to aeroallergens was observed. This inverse relationship might suggest the presence of a protective effect of atopic sensitization to aeroallergens against stronger local airway inflammation and higher systemic AERD-related inflammatory activity.

Because specific immunoglobulin E (IgE) antibodies to aspirin have not been identified and all the structurally different cyclooxygenase-1 inhibitors cause respiratory reactions in these patients, AERD is not an IgE-mediated allergy, despite the allergic-like reactions to aspirin and NSAIDs that define the disease. On the other hand, atopic sensitization to classical aeroallergens is rather common among AERD individuals, and some AERD patients with and or without a coexisting atopy exhibit higher total serum IgE levels [11][12][13][14][15]. Furthermore, beneficial clinical and biochemical (reduced LTE 4 production) effects of omalizumab have been observed in AERD subjects [16,17]. Interestingly, omalizumab is a humanized recombinant monoclonal antibody interfering with IgE-dependent MC activation [18,19].
Considering the involvement of IgE-related mechanisms in AERD, we sought, for the first time, to find a relationship between atopic sensitization to aeroallergens and local and systemic eicosanoid production in subjects with AERD.

Clinical Evaluation
Asthma control was evaluated based on ACT according to GINA 2021 guidelines, where scores of 20-25 correspond to well-controlled asthma, scores of 16-19 to partially controlled, and those of 5-15 to uncontrolled disease [24]. Further, according to GINA 2021 guidelines, asthma severity was assessed retrospectively after at least 2-3 months of treatment, from the level of treatment required to control symptoms and exacerbations. Thereafter, based on this assessment, the patients were classified as those with "mild asthma"-asthma well controlled with step 1 or step 2 treatment, "moderate asthma"asthma well controlled with step 3 or step 4 treatment, and "severe asthma"-uncontrolled despite optimized treatment, this group corresponded to step 5 treatment [24,25].

Spirometry
Standard spirometry and skin prick tests were performed.

Skin Prick Tests
Skin prick tests for aeroallergens were used to ascertain atopic sensitization status to aeroallergens. A wide panel of aeroallergens (ALK-Abelló, Madrid, Spain), including those most prevalent in our region was used, including the following: house dust mites (Dermatophagoides pteronyssinus, Dermatophagoides farinae, Lepidoglyphus destructor), pollens (Phleum, Lolium, Cupressus, Platanus, Olea, Chenopodium, Plantago, Artemisia, Parietaria, Salsola kali), molds (Alternaria, Aspergillus, Cladosporium, Penicillium) and animal epithelia (dog, cat, and hamster). Histamine was used as a positive control and saline was used as a negative control. Histamine had to produce a wheal at least three millimeters in diameter. A positive skin prick test to a certain single allergen was defined by a wheal that was equal to or larger than the histamine control [26]. Overall skin prick testing was considered positive and equal to the presence of sensitization to aeroallergens if reactivity to at least one of the aeroallergens was detected.

Induced Sputum Collection
Induced sputum collection was performed according to ERS recommendations [27]. Study subjects inhaled hypertonic saline solution; concentration of solution was increasing from 3% to 5% using ultrasonic nebulizer (Ultraneb 2000; DeVilibiss, Somerset, PA, USA). Induced sputum was manually separated from saliva, after which cytospin slides were obtained. Those were used for differential cell count.

Induced Sputum Supernatant Eoicosanoids
The concentrations of prostaglandin PGE 2 and PGD 2 were evaluated in ISS using gas chromatography/mass spectrometry (GC-MS), while those of LTE 4 were analyzed by high-performance liquid chromatography/tandem mass spectrometry (HPLC-MS/MS). These analytical method details were presented elsewhere [30].

Peripheral Blood Parameters
Total IgE levels in blood were measured by a fluoroenzyme immunosorbent assay (UniCAP; Pharmacia Diagnostics, Uppsala, Sweden).

Urinary Leukotriene E 4
Samples of urine were collected in the mornings. Enzyme-linked immunosorbent assay (Cayman Chemical Co., Ann Arbor, MI, USA) was used to assess LTE 4 excretion in urine. Finally, the results were presented in picograms per mg of urinary creatinine.

Statistical Analysis
Categorical variables were presented as numbers and percentages, and analyzed using Pearson's chi-squared test or Freeman-Halton extension to the Fisher exact test, as appropriate. Continuous variables were tested for the normality of the distribution by Shapiro-Wilk test. Those having normal distribution were expressed as mean ± standard deviation and otherwise as median (interquartile range). For the former, pairwise comparisons were made by Student's t-test, while for the latter by Mann-Whitney U-test. Spearman's rank correlation coefficient was calculated to assess the relationship between two continuous variables. Multiple linear regression models, including potential confounders as independent predictors, were used to verify the effects of aeroallergen sensitization on local and systemic eicosanoids, with all continuous variables log-transformed before entering the model. A p-value of less than 0.05 was considered statistically significant.

Characteristics of the Study Groups
One hundred and thirty-three patients with stable AERD participated in the study. Their basic demographic, clinical and laboratory characteristics are given in Table 1. Moreover, eighty stable asthmatics tolerating aspirin well (ATA) were included as controls. This group is also characterized in Table 1.

General Correlation Analysis in the Whole AERD Group
In the whole AERD group (n = 133), % forced expiratory volume in 1 s (%FEV 1 ) negatively correlated with IS eosinophils, which in turn positively correlated with blood eosinophil count and inversely with IS neutrophils. Moreover, both blood and IS eosinophils positively correlated with total serum IgE levels ( Figure 1). Finally, positive correlations of IS and/or blood eosinophils with ISS LTE 4 , ISS PGD 2 or urine LTE 4 , a positive correlation of sputum neutrophils with ISS PGE 2 , and an inverse association between urine LTE 4 and %FEV 1 were observed ( Figure 1).

Characteristics of the Study Groups
One hundred and thirty-three patients with stable AERD participated in the stu Their basic demographic, clinical and laboratory characteristics are given in Tabl Moreover, eighty stable asthmatics tolerating aspirin well (ATA) were included controls. This group is also characterized in Table 1.

General Correlation Analysis in the Whole AERD Group
In the whole AERD group (n = 133), % forced expiratory volume in 1 s (%FE negatively correlated with IS eosinophils, which in turn positively correlated with bl eosinophil count and inversely with IS neutrophils. Moreover, both blood and eosinophils positively correlated with total serum IgE levels ( Figure 1). Finally, posi correlations of IS and/or blood eosinophils with ISS LTE4, ISS PGD2 or urine LTE positive correlation of sputum neutrophils with ISS PGE2, and an inverse associa between urine LTE4 and %FEV1 were observed ( Figure 1).

Analysis of Correlations between Pro-Inflammatory Eicosanoids and Sensitization to Aeroallergens in the Whole AERD Group
Interestingly, in the whole AERD group (n = 133), atopic sensitization aeroallergens inversely correlated with ISS LTE4, ISS PGD2 and urine LTE4 values, des the presence of an expected positive correlation of aeroallergen sensitization with t serum IgE levels, the latter indirectly linked to ISS LTE4, ISS PGD2 or urine LTE4 le through the correlation with blood and sputum eosinophils ( Figure 1). Total serum

Analysis of Correlations between Pro-Inflammatory Eicosanoids and Sensitization to Aeroallergens in the Whole AERD Group
Interestingly, in the whole AERD group (n = 133), atopic sensitization to aeroallergens inversely correlated with ISS LTE 4 , ISS PGD 2 and urine LTE 4 values, despite the presence of an expected positive correlation of aeroallergen sensitization with total serum IgE levels, the latter indirectly linked to ISS LTE 4 , ISS PGD 2 or urine LTE 4 levels through the correlation with blood and sputum eosinophils ( Figure 1). Total serum IgE levels themselves showed only a tendency (both p-values of 0.07) to positively correlate with ISS LTE 4 or urine LTE 4 concentrations ( Figure 1).
The negative associations of atopic sensitization to aeroallergens with ISS LTE 4 , ISS PGD 2 or urine LTE 4 levels remained significant in multiple linear regression models, including atopic sensitization to aeroallergens, total serum IgE concentrations and potential confounders as independent predictors (Table 2). In addition, in the same models, statistical tendencies of total serum IgE levels to positively correlate with ISS LTE 4 and urine LTE 4 concentrations observed in univariate analyses became significant ( Table 2). Table 2. Multiple regression models verifying the effects of aeroallergen sensitization on local and systemic eicosanoids in the whole group of patients with aspirin-exacerbated respiratory disease (n = 133).

Correlation Analysis in the Control (ATA) Group
In the control group, comprising 80 asthmatics who tolerated aspirin well (ATA), %FEV 1 negatively correlated with both IS neutrophils and eosinophils. In turn, IS eosinophils positively correlated with blood eosinophil count and total serum IgE levels. Atopic sensitization to aeroallergens positively correlated with total serum IgE concentrations and inversely with IS neutrophils (Supplementary Figure S1).
In addition, serum or IS eosinophils positively correlated with not only ISS LTE 4 , ISS PGD 2 or urine LTE 4 levels but also ISS PGE 2 . However, no correlation between ISS or urine eicosanoids and aeroallergen sensitization was observed (Supplementary Figure S1).

Correlation Analysis in a Subgroup of AERD Patients with Eosinophilic Asthma Phenotype
Similarly to the full AERD group, in a subgroup of AERD patients with eosinophilic asthma phenotype (n = 61), atopic sensitization to aeroallergens inversely correlated with ISS LTE 4 , ISS PGD 2 and urine LTE 4 concentrations ( Figure 2). All three associations remained significant in multiple linear regression models, with aeroallergen sensitization, total serum IgE levels and potential confounders as independent predictors (Table 3).

Correlation Analysis in a Subgroup of AERD Patients with Non-Eosinophilic Asthma Phenotype
In AERD patients with non-eosinophilic asthma phenotype (n = 66), only ISS PG concentrations inversely correlated with atopic sensitization to aeroallergens (Figure and this association was not resistant to the adjustment of total serum IgE levels a several potential confounders (Table 4). In the same model, a marginally signific association of total serum IgE levels and ISS LTE4 concentrations appeared (Table 4).
Furthermore, ISS PGE2 levels positively correlated with IS neutrophils, while LTE4 concentrations positively with IS neutrophils and eosinophils and negatively w %FEV1 (Figure 3 and Supplementary Figure S3).  Furthermore, blood eosinophil count positively correlated with IS eosinophils, which negatively correlated with IS neutrophils and %FEV 1 . Blood eosinophils positively correlated with urine LTE 4 , IS eosinophils with ISS urine LTE 4 , and ISS neutrophils with ISS PGE 2 concentrations. Finally, %FEV 1 positively correlated with IS neutrophils and negatively with urine LTE 4 levels ( Figure 2 and Supplementary Figure S2).

Correlation Analysis in a Subgroup of AERD Patients with Non-Eosinophilic Asthma Phenotype
In AERD patients with non-eosinophilic asthma phenotype (n = 66), only ISS PGE 2 concentrations inversely correlated with atopic sensitization to aeroallergens (Figure 3), and this association was not resistant to the adjustment of total serum IgE levels and several potential confounders (Table 4). In the same model, a marginally significant association of total serum IgE levels and ISS LTE 4 concentrations appeared (Table 4).

Discussion
The major aim of this study was to answer a scientific question as to whether a links between IgE-mediated, atopic sensitization and local and systemic eicosano production exist in AERD.
However, to validate the generalization potential of the main findings of t investigation, we first analyzed the relationships between the basic pathophysiologi parameters of AERD-related inflammation, such as involved immune cells, eicosano or IgE, and compared them to the previous observations by other researchers [7,31]. schematically illustrate this part of our results, shown in more detail in Figure 1 and Ta 2, Figure 4 was created, with eosinophil-neutrophil balance representing a cent reference line, to which all the other parameters relate.  Furthermore, ISS PGE 2 levels positively correlated with IS neutrophils, while ISS LTE 4 concentrations positively with IS neutrophils and eosinophils and negatively with %FEV 1 (Figure 3 and Supplementary Figure S3).

Discussion
The major aim of this study was to answer a scientific question as to whether any links between IgE-mediated, atopic sensitization and local and systemic eicosanoid production exist in AERD.
However, to validate the generalization potential of the main findings of this investigation, we first analyzed the relationships between the basic pathophysiological parameters of AERD-related inflammation, such as involved immune cells, eicosanoids or IgE, and compared them to the previous observations by other researchers [7,31]. To schematically illustrate this part of our results, shown in more detail in Figure 1 and  Figure 4. A scheme summarizing the major findings of the study based on the data given in Figure  1 and In brief, a negative correlation between sputum eosinophils and neutrophils was observed as a central pathophysiological feature. While IS eosinophils positively correlated with ISS PGD2 and ISS or urine LTE4 (major pro-inflammatory arachidonic acid metabolites [32,33]), a similar type of relationship was observed between sputum neutrophils and ISS PGE2, a central anti-inflammatory eicosanoid mediator [7,32]. Furthermore, sputum eosinophils negatively correlated with FEV1 and thus positively with airway obstruction. Finally, serum blood and sputum eosinophils correlated with each other and total serum IgE levels. Taken together, this part of our results corroborates the observations made by others in similar groups of patients [7,31], suggesting an overall high generalization potential of our findings, including those trying to answer scientific questions asked here for the first time.
In predisposed subjects, aspirin and other NSAIDs cause respiratory symptoms underlain by the allergic-like reactions defining the disorder. However, since no specific IgE antibodies against any of those drugs have been identified, AERD is not an IgEmediated allergy [11]. Despite substantial progress made in our understanding of the pathobiology of the AERD, neither the triggering cause nor the detailed mechanisms of aspirin/NSAID-induced reactions have been determined. However, although atopy is not a defining feature of the disease, atopic sensitization to aeroallergens and/or related characteristics, such as elevated total serum IgE levels, is common in subjects with AERD, Figure 4. A scheme summarizing the major findings of the study based on the data given in Figure 1 and Table 2: mutual relationships between inflammatory and clinical parameters characterizing our group of 133 patients with aspirin-exacerbated respiratory disease are shown. Eosinophil-neutrophil balance represents a central reference line to which all the other parameters relate to, with only atopic sensitization shown with regard to its relationship to either total serum immunoglobulin E (IgE) or leukotriene E 4 (LTE 4 )/prostaglandin D 2 (PGD 2 ). Atopic sensitization to aeroallergens inversely correlated with induced sputum supernatant (ISS) LTE 4 , ISS PGD 2 and urine LTE 4 values, despite the presence of a positive correlation of aeroallergen sensitization with total serum IgE levels, the latter indirectly linked to ISS LTE 4 , ISS PGD 2 or urine LTE 4 levels through the correlation with blood and sputum eosinophils. For details, please refer to the Results, Sections 3.2 and 3.3, and to the Discussion and Conclusions. %FEV 1 ; % forced expiratory volume in 1 s; PGE 2 , prostaglandin E 2 .
In brief, a negative correlation between sputum eosinophils and neutrophils was observed as a central pathophysiological feature. While IS eosinophils positively correlated with ISS PGD 2 and ISS or urine LTE 4 (major pro-inflammatory arachidonic acid metabolites [32,33]), a similar type of relationship was observed between sputum neutrophils and ISS PGE 2 , a central anti-inflammatory eicosanoid mediator [7,32]. Furthermore, sputum eosinophils negatively correlated with FEV 1 and thus positively with airway obstruction. Finally, serum blood and sputum eosinophils correlated with each other and total serum IgE levels. Taken together, this part of our results corroborates the observations made by others in similar groups of patients [7,31], suggesting an overall high generalization potential of our findings, including those trying to answer scientific questions asked here for the first time.
In predisposed subjects, aspirin and other NSAIDs cause respiratory symptoms underlain by the allergic-like reactions defining the disorder. However, since no specific IgE antibodies against any of those drugs have been identified, AERD is not an IgE-mediated allergy [11]. Despite substantial progress made in our understanding of the pathobiology of the AERD, neither the triggering cause nor the detailed mechanisms of aspirin/NSAID-induced reactions have been determined. However, although atopy is not a defining feature of the disease, atopic sensitization to aeroallergens and/or related characteristics, such as elevated total serum IgE levels, is common in subjects with AERD, as it is with classic asthmatics [12][13][14][15]. Moreover, elevated levels of total serum IgE have been observed in AERD patients even in the absence of atopy [11]. Further evidence potentially supporting the involvement of IgE-related mechanisms in the pathogenesis of aspirin hypersensitivity has been provided by recent trials on the effectiveness of omalizumab, a humanized recombinant monoclonal antibody interfering with IgE-dependent MC activation [18,19], conducted in patients with AERD. While detailed mechanisms remain to be elucidated [33,34], those studies demonstrated that omalizumab therapy inhibited urinary PGD 2 and LTE 4 overproduction, and attenuated upper/lower respiratory tract symptoms during an oral aspirin challenge, resulting in improved aspirin tolerance in the majority of AERD patients treated with omalizumab [16,17].
Thus, and considering that the overproduction of the pro-inflammatory eicosanoids by eosinophils, platelet-leukocyte aggregates and MCs biochemically characterizes AERD [7][8][9][10], we checked whether there is an association between IgE-mediated sensitization to aeroallergens and the local and/or systemic levels of LTE 4 , PGD 2 and PGE 2 . Surprisingly, opposite to our expectations, urinary and sputum LTE 4 and sputum PGD 2 , the major pro-inflammatory metabolites of arachidonic acid [32,33], negatively correlated with atopic sensitization to aeroallergens both in crude and adjusted analyzes ( Figure 4). Interestingly, this observation might suggest that, in yet an unknown mechanism, the presence of atopic sensitization exerts, in patients with AERD, a kind of protective, local airway inflammationdiminishing and overall disease inflammatory activity-reducing effects. Moreover, despite a positive correlation between total serum IgE levels and atopic sensitization to aeroallergens, statistical tendencies of total serum IgE levels to positively correlate with ISS LTE 4 and urine LTE 4 observed in crude models became significant in multivariate analyzes. This finding might, in turn, suggest the presence of a dual, bidirectional relationship between IgE system and AERD-related local and systemic inflammation. In this scenario, aeroallergen-specific IgE molecules, corresponding to the presence of atopic sensitization, would protect, whereas allergen-non-specific, polyclonal IgE, reflected by total serum IgE mass, would either contribute to or, most probably, be only an epiphenomenon related to AERD-related inflammation and/or stimulation by other factors, such as staphylococcal enterotoxins [35]. Even though the presence of atopic sensitization and total serum IgE levels obviously correlate, different patterns of the associations they show with other traits, such as eicosanoids, cannot be surprising due to many possible reasons. For example, it is mostly varying genetic mechanisms that influence total and specific IgE production [36,37], and it has been shown that specific serum IgE against certain allergens can account for up to 20-25% of total serum IgE, driving the overall value of total blood serum IgE [38]. Furthermore, as partially mentioned above, even though total serum IgE levels are known to correlate with the presence of asthma or its severity, also independently of atopy [39,40], it is not fully elucidated if they mechanistically contribute to or are only a secondary epiphenomenon of asthma [36,37], possibly resulting from a polyclonal stimulation by other factors, e.g., staphylococcal enterotoxins [35].
While eosinophilic inflammation in upper and lower airways, together with MC activation, play a key role in the pathogenesis of AERD [23,41,42], non-eosinophilic airway inflammation-related mechanisms may also contribute in some AERD patients [22,[43][44][45][46]. Therefore, we performed our analyzes also after substratification of AERD individuals according to the IS cytology-determined type of airway inflammation. Interestingly, similarly to the whole AERD group, but with a stronger magnitude, atopic sensitization to aeroallergens was found in AERD patients with eosinophilic asthma inversely associated with urinary and sputum LTE 4 and sputum PGD 2 . On the contrary, no robust association between arachidonic acid metabolites and atopic sensitization was observed in the subgroup of AERD patients with non-eosinophilic airway inflammation; the only correlation observed for ISS PGE 2 in a crude analysis disappeared when tested in a multivariable model. Thus, the effect observed in the whole AERD group was probably driven by the subgroup of patients with eosinophilic asthma.
Our study has several strengths and weaknesses. The major strengths include, for AERD, a unique analysis of IS cytology and ISS eicosanoids, and a comparatively high number of well-characterized AERD subjects included in the study. The major limitation is a single time-point analysis, while the longitudinal design would make it possible to determine a predictive effect of sensitization to aeroallergens at the baseline for the levels of local and systemic eicosanoids during the follow-up.

Conclusions
In summary, to the best of our knowledge for the first time, we studied and observed the total serum IgE-independent reverse association between the presence of atopic sensitization to aeroallergens and the levels of major local and systemic pro-inflammatory arachidonic acid metabolites in AERD patients. Our results might suggest the presence of the protective effect of atopic sensitization to aeroallergens against stronger local (airway) and overall/systemic AERD-related inflammation. However, further studies, e.g., those on the interplay between systemic and local IgE, including specificities directed against some nasal bacteria antigens [6,47], are required to fully elucidate the mechanisms underlying our observation. In addition, whether our findings could translate into clinical effects, e.g., those on disease severity, should be investigated in larger clinical studies.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jpm12030447/s1, Figure S1: Mutual Spearman's rank correlations of inflammatory and clinical parameters characterizing our control group of 80 patients with aspirintolerant asthma and the correlations of those variables with induced sputum supernatant (ISS) or urine eicosanoids, Figure S2. Full set of mutual Spearman's rank correlations of inflammatory and clinical parameters characterizing a subgroup of 61 patients with aspirin-exacerbated respiratory disease having eosinophilic asthma phenotype and the correlations of those variables with induced sputum supernatant (ISS) or urine eicosanoids, Figure S3. Full set of mutual Spearman's rank correlations of inflammatory and clinical parameters characterizing a subgroup of 66 patients with aspirinexacerbated respiratory disease having non-eosinophilic asthma phenotype and the correlations of those variables with induced sputum supernatant (ISS) or urine eicosanoids.  Informed Consent Statement: All participants provided written informed consent.

Data Availability Statement:
The data presented in this work are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.