Expression of Leukotriene Biosynthetic Enzymes in Tonsillar Tissue of Children With Obstructive Sleep Apnea: A Prospective Nonrandomized Study

Importance  Cysteinyl leukotrienes (CysLTs) potentially promote adenotonsillar hypertrophy in children with obstructive sleep apnea (OSA). Previous studies have identified CysLTs and their receptors in tonsillar tissue from children with OSA.

Objective  To demonstrate expression of the leukotriene biosynthetic enzymes 5-lipoxygenase (5-LO), 5-lipoxygenase activating protein (FLAP), leukotriene A₄ hydrolase (LTA₄H), and leukotriene C₄ synthase (LTC₄S) in T and B tonsillar lymphocytes from pediatric patients with OSA. It was hypothesized that children with OSA have greater expression of biosynthetic enzymes for CysLTs (5-LO, FLAP, and LTC₄S) in their tonsillar tissue than do children with recurrent tonsillitis (RT), who were enrolled as controls.

Design, Setting, and Participants  This prospective, nonrandomized study was performed at a tertiary care university hospital on 13 children with OSA and adenotonsillar hypertrophy undergoing adenotonsillectomy and 12 children without OSA also undergoing tonsillectomy for RT. Tonsillar tissue from children with OSA or RT was examined for 5-LO, FLAP, LTA₄H, and LTC₄S expression under real time–quantitative polymerase chain reaction (RT-qPCR), flow cytometry (FC), and confocal laser scanning microscopy (CM).

Main Outcomes and Measures  Expression of biosynthetic enzymes for CysLTs (5-LO, FLAP, and LTC₄S) was the main outcome measure. Patients with OSA and control patients with RT were compared for numbers of copies of 5-LO, FLAP, and LTC₄S messenger RNA (by RT-qPCR) in T or B tonsillar lymphocytes and proportions of CD3⁺ or CD19⁺ tonsillar lymphocytes that expressed 5-LO, FLAP, and LTC₄S (by FC).

Results  Messenger RNA for all 4 enzymes was detected in T and B lymphocytes from both study groups, and expression of all biosynthetic enzymes was demonstrated in participants with OSA and RT by FC. Patients with OSA differed from controls in the proportions (median [10th-90th percentile]) of LTC₄S⁺ CD3⁺ T lymphocytes (23.31% [8.64%-50.07%] vs 10.81% [3.48%-23.32%], respectively) (P = .01) and LTC₄S⁺ CD19⁺ B lymphocytes (20.66% [14.62%-65.77%] vs 12.53% [2.87%-36.64%], respectively) (P = .01) detected by FC. Immunoreactivity for the 4 enzymes was detected by CM in B lymphocytes of mantle zones and T lymphocytes of extrafollicular areas.

Conclusions and Relevance  Leukotriene biosynthetic enzymes are expressed in tonsillar lymphocytes, and the previously reported detection of CysLTs in tonsillar tissue from children with OSA may be attributed to endogenous synthesis. Enhanced expression of LTC₄S is a potential target for pharmacologic interventions in OSA.

Increased upper airway resistance resulting from enlarged adenoid and tonsils is a common predisposing factor for obstructive sleep apnea (OSA) in childhood.¹ In children with snoring, adenotonsillar hypertrophy occurs during preschool years and persists beyond the eighth birthday,² and cysteinyl leukotrienes (CysLTs) have been implicated in its pathogenesis.^3-5 More specifically, tonsillar T and small B lymphocytes express CysLT receptors and the addition of leukotriene D₄ to tonsillar cell culture induces a proliferative response.^6,7 Administration of montelukast, an inhibitor of type 1 CysLT receptors, to children with mild OSA is accompanied by reduction in the size of adenoids and a decrease in the severity of intermittent upper airway obstruction during sleep.⁴ Moreover, increased numbers of CysLTs (leukotrienes C₄, D₄, and E₄) have been found in tonsillar tissue excised from children with OSA.⁸

Following stimulation of neutrophils and lymphocytes, 5-lipoxygenase (5-LO) localized in the cytoplasm translocates to the nucleus where it catalyzes the conversion of free arachidonic acid to leukotriene A₄.⁹ Arachidonic acid is released from the outer nuclear membrane and is presented to 5-LO by the 5-LO activating protein (FLAP).^10,11 Leukotriene A₄ can be transformed to leukotriene B₄ by leukotriene A₄ hydrolase (LTA₄H) or to leukotriene C₄ by leukotriene C₄ synthase (LTC₄S). Both FLAP and LTC₄S are proteins embedded in the nuclear membrane.¹² Leukotriene C₄ is further converted to leukotrienes D₄ and E₄ by extracellular enzymes.¹³

Previous studies in adults have detected the presence of 5-LO in tonsillar B cells, but conflicting data have been presented in relation to the presence of 5-LO in T lymphocytes.^9,11,14 To our knowledge, there are no published reports on the expression of biosynthetic enzymes for leukotrienes B₄ and C₄ in tonsillar tissue of children with OSA. Hence, the primary goal of this investigation was to demonstrate expression of enzymes related to the leukotriene biosynthetic pathway (5-LO, FLAP, LTA₄H, and LTC₄S) in tonsillar T and B lymphocytes. Furthermore, it was hypothesized that children with OSA have greater expression of biosynthetic enzymes for CysLTs (5-LO, FLAP, and LTC₄S) than children with recurrent tonsillitis (RT) recruited as controls. Detection of leukotriene biosynthetic enzymes in tonsillar cells is of clinical importance, since the LTC₄S catalytic architecture has been elucidated, and an LTC₄S inhibitor with potential therapeutic applications has been described recently.^12,15

The research protocol was approved by the institutional review board of the Aghia Sophia Children’s Hospital (Scientific Council approval No. 25930/19-11-10), and written informed consent for participation in the study was obtained from the parents of all participants.

Children with OSA and adenotonsillar hypertrophy who underwent adenotonsillectomy after preoperative polysomnography were recruited for the study. OSA was diagnosed when symptoms of a sleep-related breathing disorder (SRBD) were present, and the apnea-hypopnea index in polysomnography was greater than 1 episode per hour.¹ Children without symptoms of OSA who had tonsillectomy for RT (≥7 episodes over the past year) and an SRBD score lower than 0.33 (by the Pediatric Sleep Questionnaire¹⁶) were recruited as controls. Children with a diagnosis of asthma or history of respiratory infection during the previous 8 weeks were excluded.

The SRBD score was calculated for all participants, and a physical examination was completed. Size of tonsils was graded from 1+ to 4+ by direct inspection of the oropharynx, and tonsillar hypertrophy was diagnosed when tonsils were larger than 2+.¹⁷ The patients’ weight and standing height were measured, and body mass index (BMI) z-score was calculated.¹⁸ Adenoidal hypertrophy was diagnosed by lateral neck radiography.

Polysomnography was carried out for 1 night (9 hours) at the Sleep Disorders Laboratory using the Somnostar Cephalo Pro Amplifier and Software (Viasys Healthcare). A 4-channel electroencephalogram (C3/M2, O2/M1, O1/M2, and F4/M1), 2-channel electrooculogram, submental and tibial electromyograms, and an electrocardiogram were recorded. Airflow was detected by thermocouples at the nose and mouth and by nasal pressure transducer, and respiratory movements were monitored using inductive plethysmography with thoracic and abdominal belts (RespiTrace QDC, RIP module; Viasys Healthcare). The oxygen saturation of hemoglobin was measured by an oximeter. Sleep stages, arousals, and respiratory events were scored using the American Academy of Sleep Medicine recommendations.¹⁹

After surgical excision, tonsillar tissue was placed in phosphate-buffered saline and transferred rapidly to the pathology laboratory for further processing and use in real time–quantitative polymerase chain reaction (RT-qPCR), flow cytometry, and confocal laser scanning microscopy (for details, see the eAppendix in the Supplement).

Total RNA was extracted from T or B lymphocyte fractions with the NucleoSpin RNA/Protein kit (Macherey-Nagel GmbH & Co KG), and its concentration and quality were determined using a spectrophotometer (Genova; Jenway). A total of 1.0 µg of RNA was reverse-transcribed into cDNA by Moloney Murine Leukemia Virus Reverse Transcriptase using oligo(dT) as reverse transcription primer (PrimeScript RT-PCR kit; TaKaRa Bio Europe).

The cDNA equivalent to 5.0 ng of total RNA was subjected to RT-qPCR analysis in triplicates in an Mx3005P RT-qPCR system (Stratagene) according to the manufacturer’s protocol (KAPA SYBR Fast Universal qPCR kit, KAPA Biosystems). Expression of 5-LO, FLAP, LTA₄H, and LTC₄S were reported as ratios of the gene of interest value to the corresponding β-actin value; PCR results were analyzed using Mx3005P software (Stratagene). Details for the RT-qPCR methods are provided in the eAppendix in the Supplement.

A 3-color experiment was set up. Two-color staining with directly labeled antibodies was applied to identify subpopulations of T and B lymphocytes. More specifically, T and B lymphocytes were defined by a fluorescein isothiocyanate (FITC) mouse antibody against human CD3 and a PerCP-Cy5.5 mouse antibody against human CD19 (BD Biosciences). To quantify the expression of the intracellular leukotriene biosynthetic enzymes (5-LO, FLAP, LTA₄H, and LTC₄S) in tonsillar lymphocytes, indirect immunofluorescence staining was used. Lymphocytes were permeabilized and fixed with the BD Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer’s protocol. Cells that had undergone fixation and permeabilization were then incubated for 30 minutes at 4°C in the dark with rabbit polyclonal antibodies against 5-LO (dilution 1:50; Cayman Chemical), FLAP (dilution 1:50; Santa Cruz Biotechnology), LTA₄H (dilution 1:200; Cayman Chemical), or LTC₄S (dilution 1:100, Sigma-Aldrich). Finally, cells were incubated with fluorochrome R-phycoerythrin (PE)–conjugated F(ab′) secondary anti-rabbit antibody (BD Biosciences) for 30 minutes at 4°C in the dark.

During the experimental procedure and along with the test samples, various controls were prepared: (1) unstained cells to check the background fluorescence of the cells; (2) CD3 FITC–stained cells and CD19 PerCP-Cy5.5–stained cells to check the background fluorescence of the double-stained cells, the nonspecific antibody fixation, and the fluorescence spillover in the channel used for quantitation of the intracellular enzymes; and (3) CD3 FITC–, CD19 PerCP-Cy5.5–, and PE–conjugated F(ab′) secondary anti-rabbit antibody–stained cells (secondary control) to check the background fluorescence of the stained cells and the nonspecific antibody fixation. The secondary control sample was used to set the boundaries for the marker delimiting the negative expression for every patient.

Fluorescence was measured on a FACS Calibur cytometer (BD Biosciences) using BD CellQuest Pro software. The analysis of the results was performed with BD FACS Diva software (BD Biosciences).

Confocal laser scanning microscopy was used to localize T and B lymphocytes expressing leukotriene biosynthetic enzymes within the tonsillar tissue. After deparaffinization and antigen retrieval, a 2-day, double-stain protocol was applied in adjacent formalin-fixed, paraffin-embedded tissue sections. The same antibodies used for detection of biosynthetic enzymes by flow cytometry were also applied in the tonsillar tissue for examination by confocal laser scanning microscopy (anti-5-LO, dilution 1:100; anti-FLAP, dilution 1:100; anti-LTA₄H, dilution 1:100; and anti-LTC₄S dilution 1:100). The detailed protocol is described in the eAppendix in the Supplement.

Expression of biosynthetic enzymes for CysLTs (5-LO, FLAP, and LTC₄S) was the main outcome measure. Patients with OSA and control participants with RT were compared as follows: (1) for patient characteristics, the t test was used for continuous variables and the χ² test (with the Yates correction) for categorical variables; (2) for numbers of 5-LO, FLAP, and LTC₄S mRNA copies found by RT-qPCR in T or B lymphocytes, multivariable analysis of variance (MANOVA) was used, followed by univariate F tests; and (3) for proportions of CD3⁺ or CD19⁺ lymphocytes that expressed 5-LO, FLAP, and LTC₄S found by flow cytometry, MANOVA was used, followed by univariate F tests.

Thirteen children who underwent polysomnography and adenotonsillectomy for OSA were enrolled in the study. Twelve children without symptoms of OSA, with SRBD scores lower than 0.33, and who underwent tonsillectomy for RT were also enrolled in the study as controls. The 2 study groups were similar in terms of age at surgery, female-to-male ratio, and BMI z-score, but they differed significantly in SRBD score (P < .01) (Table 1). None of the participants had a history of physician-diagnosed allergic rhinitis or atopic dermatitis. Children with OSA had a median (10th-90th percentiles) apnea-hypopnea index of 10.6 (3.7-54.1) episodes per hour, a respiratory arousal index of 1.5 (0-6.8) episodes per hour, an oxygen desaturation of hemoglobin index of 11 (1.3-53.8) episodes per hour, and an oxygen saturation of hemoglobin nadir of 86% (74.2%-97.0%).

Messenger RNA for 5-LO, FLAP, LTA₄H, and LTC₄S was detected in both CD3⁺ T lymphocytes and CD19⁺ B lymphocytes isolated from children with OSA and controls with RT (Table 2). The 2 study groups did not differ in terms of 5-LO, FLAP, and LTC₄S mRNA copies obtained from CD3⁺ T or CD19⁺ B lymphocytes (Table 2).

Both tonsillar CD3⁺ T lymphocytes and CD19⁺ B lymphocytes in children with OSA or RT expressed all 4 enzymes of the biosynthetic pathway for leukotrienes B₄ and C₄ (5-LO, FLAP, LTA₄H, and LTC₄S) (eFigures in the Supplement). Children with OSA had significantly higher LTC₄S⁺ fractions of CD3⁺ T lymphocytes and CD19⁺ B lymphocytes than participants with RT (P = .01) (Table 3). The 2 study groups were similar regarding fractions of tonsillar CD3⁺ and CD19⁺ lymphocytes expressing 5-LO and FLAP (Table 3).

Tonsillar tissue samples from 4 patients with OSA and 3 control participants, randomly selected among children of the current study cohort, were examined by confocal microscopy to localize the expression of leukotriene biosynthetic enzymes. In patients from both groups, the biosynthetic enzymes (5-LO, FLAP, LTA₄H, and LTC₄S) were expressed mostly by CD3⁺ T lymphocytes located in the extrafollicular areas (Figure 1) and by CD20⁺ B lymphocytes found in the follicular mantle zones (Figure 2). The T and B lymphocytes visualized by confocal microscopy had the characteristic DAPI-stained large nucleus surrounded by a rim corresponding to the nuclear membrane, cytoplasm, and cellular membrane. Immunoreactivity detected in the rim results from expression of CD3 or CD20 antigens (cellular membrane) with or without coexpression of the biosynthetic enzymes (nuclear membrane or cytoplasm) (Figures 1 and 2).

In adults, circulating myeloid cells (neutrophils, eosinophils, basophils, monocytes) and B lymphocytes express 5-LO, and hence they have the potential to synthetize leukotrienes.^9,10,20 However, not all reports have provided consistent data about the presence of 5-LO in T lymphocytes.^9,11,21 To our knowledge, the present study is the first to demonstrate that both tonsillar T and B lymphocytes from children with OSA express enzymes for the biosynthesis of leukotrienes B₄ and C₄. This finding is of clinical importance because CysLTs (leukotriene C₄ and its product leukotrienes D₄ and E₄) have been implicated in the pathogenesis of tonsillar hypertrophy, and LTC₄S—a key enzyme for the biosynthesis of leukotriene C₄—is soon to become a therapeutic target for a recently developed LTC₄S inhibitor.^7,15

In the adenotonsillar tissue, germinal centers of the lymphoid follicles are surrounded by the mantle zones of the follicles and the extrafollicular areas. A previous study by our group⁶ has shown that type 1 and type 2 CysLT receptors are expressed by B lymphocytes in the tonsillar mantle zones and by T lymphocytes residing in the extrafollicular areas. In this investigation, leukotriene biosynthetic enzymes were identified in the same histologic areas and tonsillar cell subpopulations. In addition, greater expression of LTC₄S was found in children with OSA than in children with RT.

It is known that naive B lymphocytes move to the tonsillar mantle zones from the systemic circulation during early life and subsequently interact in the extrafollicular areas with T lymphocytes and antigen-presenting cells that have processed external antigens.²² Thereafter, these primed B lymphocytes are attracted to the germinal centers, where they evolve into plasma cells producing immunoglobulins.²² Therefore, we propose that during early childhood, cells of the immune system and external antigens interact in the tonsillar mantle zones and extrafollicular areas, promoting endogenous production of leukotrienes and proliferation of tonsillar lymphocytes and ultimately leading to tonsillar growth.⁷ This physiologic phenomenon might be exaggerated in certain subgroups of susceptible children such as those with increased urinary excretion of CysLTs or a history of wheezing requiring treatment with inhaled medications.^3,23

During the first 8 years of life, pharyngeal lymphoid tissue (adenoid and palatine tonsils) restricts the upper airway lumen in a variable degree in most children.² Thereafter, adenotonsillar tissue overgrowth resolves in nonsnorers but persists in children with snoring. As a result, the oropharyngeal airway lumen in snorers is steadily restricted by hypertrophic lymphoid tissue regardless of age.²

Only speculations can be made about the stimuli inducing overproduction of CysLTs and enhanced expression of their receptors. Under experimental conditions, 5-LO in tonsillar B lymphocytes from adults can be activated by oxidative agents such as hydrogen peroxide.¹⁰ Increased hydrogen peroxide levels measured in the exhaled breath condensate of children with OSA might promote biosynthesis of leukotrienes and proliferation of lymphocytes in the adenoid and tonsils.²⁴ Additionally, viral respiratory tract infections in infancy such as respiratory syncytial virus bronchiolitis have been associated with increased production of CysLTs.²⁵

Although adenotonsillectomy is the standard treatment for OSA in children with adenotonsillar hypertrophy, concerns have been raised regarding its adverse effects on regional immunity of the respiratory tract.^22,26 Pharmacologic inhibition of the leukotriene biosynthetic pathway in the adenoid and tonsils in early childhood could prove to be an alternative intervention for the treatment or even prevention of adenotonsillar hypertrophy.^5,23 It is unknown whether adenotonsillectomy, antibiotics, and systemically or intranasally administered corticosteroids reduce biosynthesis of leukotrienes in children with adenotonsillar hypertrophy and OSA. However, in adults with OSA, treatment with nasal continuous positive airway pressure or upper airway surgery is not accompanied by reduction in leukotriene B4 levels.^27,28

In accordance with previous studies, children undergoing tonsillectomy for RT were recruited as controls^4,29 in the present study, since it is unethical to obtain tonsillar tissue from healthy children. A potential limitation of the present study is that control participants did not undergo polysomnography owing to families’ practical difficulties. Instead, the Pediatric Sleep Questionnaire by Chervin et al¹⁶ was used to rule out OSA.

Tonsillar lymphocytes from children with OSA or RT express enzymes of the leukotriene biosynthetic pathway in the tonsillar mantle zones and the extrafollicular areas. Furthermore, enhanced expression of LTC₄S in children with OSA may contribute to the pathogenesis of palatine tonsil overgrowth and increased upper airway resistance. These novel findings suggest that inhibition of leukotriene biosynthetic enzymes should be explored as a potential therapeutic intervention for pediatric OSA.

Submitted for Publication: April 6, 2014; final revision received June 14, 2014; accepted July 12, 2014.

Corresponding Author: Athanasios G. Kaditis, MD, First University Department of Pediatrics, Aghia Sophia Children’s Hospital, Thivon and Papadiamantopoulou Streets, Athens 11527, Greece (kaditia@hotmail.com).

Published Online: September 11, 2014. doi:10.1001/jamaoto.2014.1936.

Author Contributions: Dr Kaditis had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Tsaoussoglou, Hatzinikolaou, Baltatzis, Lianou, Maragozidis, Balatsos, Kaditis.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Tsaoussoglou, Hatzinikolaou, Baltatzis, Lianou, Maragozidis, Balatsos, Kaditis

Critical revision of the manuscript for important intellectual content: Chrousos, Kaditis.

Statistical analysis: Tsaoussoglou, Hatzinikolaou, Baltatzis, Lianou, Maragozidis, Balatsos, Kaditis.

Obtained funding: Chrousos.

Administrative, technical, or material support: Balatsos, Kaditis.

Study supervision: Kaditis.

Conflict of Interest Disclosures: None reported.

Funding/Support: This research was supported by intramural funding from the University of Athens Research Committee.

Role of the Funder/Sponsor: The University of Athens Research Committee had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Information: Drs Tsaoussoglou and Hatzinikolaou have contributed equally to the preparation of this report.