Interferon γ (IFN-γ) (top) and interleukin (IL) 12p40/IL-10 (bottom) production by sinus lavage (SL), bronchial lavage (BL), and peripheral blood mononuclear (PBMN) cells with or without stimuli. Each data point represents a median (horizontal bar) and range (vertical bar). Asterisk indicates P<.05 compared with values obtained in group 2 patients; Con A, concanavalin A; PHA, phytohemagglutinin; group 1, patients with high production of INF-γ by SL cells; and group 2, patients with low production of INF-γ by SL cells.
Jyonouchi H, Sun S, Rimell FL. Cytokine Production by Sinus Lavage, Bronchial Lavage, and Blood Mononuclear Cells in Chronic Rhinosinusitis With or Without Atopy. Arch Otolaryngol Head Neck Surg. 2000;126(4):522-528. doi:10.1001/archotol.126.4.522
Chronic sinus inflammation may be determined partly by a balance of proinflammatory and counterregulatory cytokines and other mediators in the sinus. However, their mechanistic roles in chronic rhinosinusitis (CRS) are not well understood.
To evaluate production of proinflammatory (interferon γ [IFN-γ] and interleukin [IL] 12) and counterregulatory cytokines (IL-10 and IL-4) by sinus lavage (SL), bronchial lavage (BL), and peripheral blood mononuclear (PBMN) cells in patients with CRS.
We analyzed SL, BL, and PB samples obtained at surgery from 26 patients with CRS. Cytokine production was determined by culturing cells with or without stimuli. The results were evaluated in comparison with other inflammatory variables (cytologic findings, total protein, IgG, and lactose dehydrogenase), bacterial cultures, and clinical features.
Production of IFN-γ by SL cells was variable and did not correlate with other inflammatory variables, microbes grown, IL-10/IL-12p40 production by SL cells, or IFN-γ production by BL or PBMN cells. Production of IL-4 by lavage cells was undetectable. None of 10 patients with elevated IFN-γ production (>800 pg/106 SL cells with mitogen stimuli) had allergic rhinitis, whereas 12 of 16 patients with low IFN-γ production (<500 pg/106 SL cells) had allergic rhinitis with positive reactivity to common aeoroallergens. There was no significant difference in other variables measured between low and high IFN-γ production groups.
Elevated IFN-γ production by SL cells may indicate much less possibility of allergic rhinitis in patients with CRS, but other variables measured did not differ in patients with high or low IFN-γ production by SL cells.
CHRONIC rhinosinusitis (CRS) is a common but often debilitating disease, consuming a great deal of medical resources.1 Results of routine immune workups are generally normal in patients with CRS; causative pathogens are those commonly found in otitis media and bronchitis; and CRS seldom develops into systemic infection.1- 5 Nevertheless, sinus inflammation persists in spite of aggressive treatments, including prolonged courses of antibiotic therapy and surgical procedures. There may be abnormal immune defense and/or disregulated inflammatory responses occurring in the sinus of patients with CRS, but pathogenesis of CRS is poorly understood.
Recent progress in basic immunology revealed an intricate immune network centered on subsets of T helper (Th) cells6- 9; Th as well as cytotoxic T-cell subsets are characterized by their distinguished cytokine production patterns that are pivotal in determining the subsequent immune responses, ie, type 1 and type 2 (T1 and T2) responses. The T1 responses induce phagocytic cell-mediated immune responses by producing T1 cytokines (interferon γ [IFN-γ] and interleukin [IL] 2) and IgG1/IgG3 antibodies (Ab) that enhance opsonization.6- 9 The T2 responses induce eosinophil-mediated inflammatory responses and counterregulate T1 responses by production of T2 cytokines (IL-4, IL-5, IL-13, etc) and IgG4/IgE Abs.6- 9 The differentiation of T1 and T2 cells on antigen (Ag) stimuli is partly determined by microenvironmental factors, including concentrations of cytokines (IFN-γ, IL-4, etc), the kind of adjuvants (microbial products), and Ag doses.6,10- 15 A balance of T1 and T2 responses appears to determine the status and outcome of many diseases.8,9 Imbalance of T1 and T2 responses and resultant disregulated inflammatory responses may also be associated with pathogenesis of CRS. In fact, others reported up-regulated messenger RNA expression of proinflammatory cytokines in patients with CRS and nasal polyposis,16,17
This study examines our hypothesis that in patients with CRS, disregulated proinflammatory responses in the sinus are reflected in changes in production of T1 and T2 and their regulatory cytokines by sinus lavage (SL) cells. We determined production of IFN-γ (a T1 cytokine), IL-4 (a T2 cytokine), IL-10 (a counterregulatory cytokine), and IL-12p40 (a key cytokine for IFN-γ production) by SL cells obtained at the time of sinus surgery from patients with CRS along with other nonspecific inflammatory variables (cytologic findings, cell number, and levels of total protein [TP], total IgG, and lactose dehydrogenase [LDH] in lavage samples). We also evaluated the results in comparison with these variables in bronchial lavage (BL) samples and cytokine production by peripheral blood mononuclear (PBMN) cells. Our purpose was to determine whether sinus inflammation affects inflammation in the lower airway or cytokine production by BL and PBMN cells.
The study population included 26 patients (age, 2-44 years; 15 female and 11 male; 7 adults and 19 children) who underwent sinus surgery or adenoidectomy plus sinus tap (children aged <2 years) at the Fairview University Medical Center, Minneapolis, Minn. Children younger than 12 years underwent adenoidectomy at the time of sinus surgery. All patients had clinical signs of rhinosinusitis (pain, nasal congestion, rhinorrhea, cough, etc) for longer than 3 months, had undergone more than 2 failed courses of antibiotic therapy (>14 days), and had positive findings of sinus computed tomographic scan. Antibiotic therapy was discontinued 1 week before the surgery. The study protocol was approved by the Institutional Review Committee, University of Minnesota, Minneapolis, and a signed written consent form was obtained before the surgery. Presence of asthma and allergic rhinitis was evaluated by history, results of physical examination, skin test (ST) reactivity, and pulmonary function tests, including responses to β2 agonist and metacholine challenge (aged >12 years), and/or IgE Ab levels against common aeroallergens.18 Positive reactivity on ST results was defined as a reaction to common aeroallergens. Patients with CRS and nasal polyposis, aspirin sensitivity, known primary or secondary immunodeficiency, or illness involving major organs were excluded. Five adult volunteers (2 healthy adults and 3 with allergic rhinitis) without history of CRS provided control SL samples. Control peripheral blood samples were obtained from 10 adult volunteers without atopy or CRS.
Samples of SL and BL were obtained by flushing the maxillary sinus through an 18-gauge spinal needle and a rigid bronchoscope with sterile isotonic sodium chloride solution (10-15 mL) before surgery. The BL samples were obtained from patients who consented or whose parents consented to the procedure. The SL samples from adult controls were obtained through sinus tap under local anesthesia (lidocaine hydrochloride [Xylocaine]).
The SL samples were sent to the clinical microbiology laboratory at the University of Minnesota for bacterial culture in 23 of 26 patients; culture results were expressed as light, moderate, and heavy growth when the bacterial colony was detected on the first, second, or third agar plate, respectively, on which samples were streaked consequently without reapplying the sample to an applicator.
To obtain levels of LDH in lavage samples, SL and BL samples were spun down, and the supernatant was harvested and frozen at −20°C until the day of measurement. Levels of LDH were measured using an LDH kit (EC220.127.116.11 UV-test; Sigma-Aldrich Corporation, St Louis, Mo),19 and results were expressed as units per gram of protein. Total protein levels were measured using the Lowry method (sensitivity, 1 U/L for LDH and 0.005 g/L for TP).
The cell number of SL and BL samples was measured using trypan-blue dye exclusion in a hemocytometer. Then 1 × 105 to 2 × 105 cells in 200 µL of phosphate-buffered saline solution (PBS) were cytospinned, dried, and stained (Giemsa stain [Diff-Quick]; Baxter, MacGaw Park, Ill). A single person (H.J.) throughout the study evaluated cytologic findings of cytospinned samples.
Total IgG levels in the lavage samples were determined as a marker of exudation and increased permeability of mucosa. Lavage samples diluted serially were tested for total IgG levels using enzyme-linked immunosorbent assay (ELISA) as reported previously.20 Purified human IgG (Sigma-Aldrich Corporation) was used as a standard (sensitivity, 5 µg/L; interassay and intra-assay variations of LDH, TP, and IgG levels, <5%).
For cytokine production, SL and BL cells were filtered through coarse gauze to remove mucins and tissue debris, spun down briefly, and washed once with PBS. The PBMN cells were obtained by centrifuging cells with density gradient (Ficoll-Hypaque; Life Technologies, Rockville, Md) at 1500 rpm for 30 minutes at room temperature. Lavage cells (2 × 105 cells/mL) as well as PBMN cells (106 cells/mL) were cultured with or without stimuli in RPMI 1640 supplemented with 2.5% fetal calf serum (HyClone, Logan, Utah), sodium pyruvate (1 mmol/L), L-glutamine (2 mmol/L), HEPES (25 mmol/L), streptomycin (100 mg/L), penicillin G (105 U/L), and 2-mercaptoethanol (10−6 mol/L) for 3 days in a 5% carbon dioxide incubator at 37°C in 5-mL sterile, capped tubes. We used mitogens (phytohemagglutinin, 2 mg/L, and concanavalin A, 1 mg/L [Sigma-Aldrich Corporation]) and dust mite extract (a mixture of Dermatophagoides farinae and Dermatophagoides pteronyssinus, 5 mg/L of each [Greer, Lenoir, NC]) as stimuli. Dust mite Ag was selected as a recall Ag, since all STs included dust mite, since there is no seasonal variation of dust mite reactivity, and since most subjects demonstrate responses to dust mite Ag in the assays of cytokine production by PBMN cells (H.J. and S.S., unpublished data, 1999).21 Levels of IFN-γ and IL-4 were measured as representative T1 and T2 cytokines, respectively, using ELISA (Endogen, Cambridge, Mass). Levels of IL-10 and IL-12p40 were also measured using ELISA (R & D, Minneapolis, Minn) as representative regulatory cytokines; IL-10 suppresses immune responses, whereas IL-12 augments T1 responses.13
Cytokines that function as autocrine growth factors, such as IL-2, are consumed rapidly by T cells. Interleukin 12p70 is rapidly degraded into IL-12p40 and IL-12p35. Certain cytokines, such as IFN-γ, stimulate endogenous production. Thus, it is difficult to evaluate recovery of cytokines added to the culture. Because of these problems, it is a common procedure to conduct a time-course study to determine an optimal time point at which the highest levels of cytokines are obtained. We conducted the time-course study using PBMN cells and determined the optimal time point for each cytokine (3-4 days for IFN-γ, IL-12, and IL-10 and 2-3 days for IL-4). It was difficult to perform such a time course study using lavage cells, owing to limited number of cells recovered. Cytokines (IFN-γ, IL-10, and IL-12p40) added to the culture medium without cells are fairly stable, and we recovered more than 90% of cytokines following 3 days' incubations.
Duplicate samples appropriately diluted with the culture medium or standards (50 µL/well) were added to the precoated ELISA plate and incubated at room temperature for 2 hours, washed, incubated with biotinylated second Ab at room temperature for 1 hour, washed again, and incubated with streptavidin–horse radish peroxidase conjugate (100 µL/well) at room temperature for 30 minutes. After washing, the color was developed by adding substrate solution (tetramethylbenzidine, 100 µL/well [DAKO, Carpinteria, Calif]), and optical density at 450 nm was read using optic density at 650 nm as a reference value (sensitivity ELISA, 3.9 ng/L [IFN-γ and IL-10], 15.6 ng/L [IL-12p40], and 0.25 ng/L [IL-4]). Intra-assay and interassay variation of cytokine levels were less than 5%.
Equality of 2 sets of data values was evaluated using Mann-Whitney test (2 sets of independent samples) or Wilcoxon weighed ranks test (2 sets of related samples). Comparison of multiple values was performed using Kruskal-Wallis test. Correlation of 2 variables was assessed using Kendall τ-b test. Differences with P<.05 were considered to be significant.
The SL and BL samples were evaluated to determine if there was any association between sinus and bronchial inflammation. The SL and BL cell numbers from patients with CRS varied considerably, whereas the number of SL cells recovered by sinus tap in controls was very low (Table 1). Major cellular components were neutrophils and epithelial cells in SL and BL samples obtained from patients with CRS (Table 1). In 5 controls, they were neutrophils, lymphocytes, and epithelial cells (Table 1). In patients with CRS, percentage of neutrophils was higher in SL than in BL samples, whereas the percentage of epithelial cells was higher in the BL samples (Table 1). No significant eosinophilia was observed, but percentage of eosinophils was higher in the SL cells (Table 1). The LDH values were low in the BL and SL fluid from patients with CRS. Total IgG level was higher in the SL than in the BL fluid samples (Table 1). Levels of IgG were very low in the SL fluid from 5 controls. Levels of IgG in the SL samples from adults were higher than those in children, but there was no statistical difference between children and adults in all of the other variables tested.
Bacterial cultures were positive in 19 of 25 patients undergoing testing. Light and moderate growth of a mixed bacterial flora (coagulase-negative Staphylococcus, Streptococcus pneumoniae, α-hemolytic streptococcus, and Haemophilus influenzae) were found in 9 and 7 patients, respectively, of the 17 with positive cultures. Cultures were all negative for bacteria in healthy adult controls.
Highly variable IFN-γ levels were detected in the supernatant of cultured SL and BL cells, but IFN-γ levels were higher in SL cells than in BL cells (Figure 1). Spontaneous IFN-γ production by SL cells was detected in 11 of 26 patients with CRS (Figure 1). Production of IFN-γ by SL cells was undetectable in 5 controls, irrespective of the stimuli. The PBMN cells from the patients with CRS produced equivalent amounts of IFN-γ and IL-4 compared with healthy adult volunteers (n=10) (Table 2). There was no correlation between IFN-γ production by PBMN cells and that by SL or BL cells (P>.05). Interleukin 4 was undetectable (<5 ng/106 cells) in the culture of SL and BL cells (Table 2). Lavage cells produced a fair amount of IL-10 and IL-12p40, regulatory cytokines, in most patients but not in controls (Table 2). Interleukin 12p70 (functional IL-12) was detected in only 5 of 24 of the SL supernatants; IL-12p70 was generally detectable only when we stimulated isolated monocytes and macrophages with lipopolysaccharide. Levels of IL-12p40/IL-10 in the SL or BL cell culture supernatants did not correlate with IFN-γ levels in all the culture conditions tested. Peripheral blood mononuclear cells produced a fair amount of IL-12p40 with stimuli but not much IL-10 in most patients with CRS (23/26); similar results were found in healthy adult controls.
When we compared all the IFN-γ data, it became clear that there was a subset of patients who produced IFN-γ at significantly high levels in at least 1 of the culture conditions (n=10). All of these patients produced IFN-γ levels of greater than 1000 pg/106 SL cells, except 1 patient (810 pg/106 SL cells), whereas others produced IFN-γ levels of less than 500 pg/106 cells. We thus evaluated the value of each variable in patients with high (group 1) or low (group 2) IFN-γ production by SL cells (Figure 1 and Table 3). Spontaneous IFN-γ production by SL cells was detected in 4 of 10 and 6 of 16 group 1 and 2 patients, respectively. Interleukin 10/IL-12p40 and the ratio of IL-10/IL-12p40 produced by SL cells did not differ significantly between groups (P>.05). Production of IFN-γ, IL-4, and IL-12p40 by PBMN cells did not differ between groups. Likewise, there was no difference in cytologic findings or levels of LDH, TP, and IgG in the BL and SL samples between groups.
Cultures were positive for bacteria in 5 of 10 and 12 of 16 patients with high and low IFN-γ production, respectively (Table 3). In group 1, no patients were diagnosed as having allergic rhinitis; however, 2 were diagnosed as having asthma. In group 2, 12 patients received a diagnosis of allergic rhinitis; 9, asthma.22 In group 1, all subjects undergoing testing (n=7) showed no reactivity to any of the aeroallergens tested, but all 12 patients in group 2 reacted to more than 3 aeroallergens, confirmed by results of skin-prick testing or presence of allergen-specific IgE antibodies. There was no significant difference in cytokine production with dust mite Ag between patients with or without reactivity to dust mites in results of skin testing.
We determined proinflammatory and counterregulatory cytokine production pattern by SL, BL, and PBMN cells in patients with CRS and evaluated the results in comparison with other nonspecific inflammatory variables and clinical features. Our results indicate that there may be at least 2 distinct subsets of patients with CRS, one with elevated IFN-γ production in the sinus, and the other with low IFN-γ production. The latter group appears more likely to suffer from allergic rhinitis.
The etiology of CRS appears heterogeneous, and a subset of patients with CRS undergoes multiple courses of antibiotic therapy as well as surgical procedures despite appropriate medical treatment.1 It is desirable to identify patients with CRS who are resistant to conventional treatment in the early stage of the disease to prevent disease persistence and complications. Postulated risk factors for CRS include atopic disorders, asthma, peripheral eosinophilia, aspirin sensitivity, specific Ab deficiency, and age (>50 years), but risk factors may vary depending on the study population.1- 5,16,17,23,24 This may well be associated with heterogeneity in patients with CRS per se. It is critical to identify objective variables that enable us to distinguish subsets of CRS for clinical management.
Chronic rhinosinusitis with eosinophilic nasal polyposis composes a distinct subset of CRS, characterized by massive eosinophil infiltration into the sinus and nasal mucosa and recurrent, extensive nasal polyposis.16,17 In these patients with CRS and nasal polyposis, others have reported the following 2 possible subsets on the basis of cytokine messenger RNA and protein expression in the sinus mucosa: one with elevated T1 cytokine expression and higher frequency of aspirin sensitivity, and the other with increased T2 cytokine expression and atopic predisposition.16,17 Given these findings, we postulated that in patients with CRS without nasal polyposis, T1 and T2 cytokine production pattern by SL might help differentiate subsets of patients with CRS also. Type 1 and T2 cytokine levels in the SL fluid were very low or undetectable in our preliminary studies; this may result partly from degradation of cytokines by enzymatic activities of other factors secreted. Thus, in our study, we determined cytokine production by SL cells, reasoning that SL cells are easier to obtain than a biopsy specimen of sinus mucosa, and the effects of other microenvironmental factors can be removed in part by washing cells. The cytokine production by SL cells may be less affected by the biopsy site, unlike cytokine expression of biopsy specimen. Since extensive CRS in the adults was postulated as a T2 disease,24 we initially determined IFN-γ and IL-4 production by SL cells as representative T1 and T2 cytokines, respectively.
Our results showed that IFN-γ production by SL cells was markedly variable, and IL-4 production was minimal in our study subjects. However, IFN-γ and IL-4 production by PBMN cells from patients with CRS was equivalent to that in healthy adult controls. When culturing SL cells, we did not separate mononuclear cells from the lavage cells because of the small cell number. It is therefore possible that IFN-γ produced may have derived from cells other than monocytes, macrophages, and lymphocytes. Although most cells surviving at the end of culturing were lymphocyte and monocyte or macrophage lineage cells, the source of IFN-γ needs to be further addressed in future studies using other methods.
In our study, we used a relatively small number of lavage cells (2 × 105 cells/mL) for culture, and IFN-γ was undetectable in the culture supernatants of SL cell from controls. Thus we were surprised to find that SL cells from a few patients with CRS produced a large amount of IFN-γ. These apparently augmented IFN-γ responses may be associated in part with pathogens grown in the tissue.7,10,13 However, bacteria grown in the sinus were common pathogens, including coagulase-negative staphylococcus strains, S pneumoniae, and H influenzae. Most patients with cultures positive for these pathogens had light to moderate bacterial growth. Moreover, bacterial cultures in 7 of 25 patients had no growth of bacteria, and there was no apparent correlation between bacterial growth and high IFN-γ production. We thus hypothesized that IFN-γ production by SL cells in patients with CRS more likely is determined by microenvironmental factors, including cytokine levels and other mediators produced by SL cells and the sinus mucosa.
To examine our hypothesis, we divided our study population into 2 groups with high or low IFN-γ production by SL cells and compared other inflammatory variables and clinical features between groups. Among regulatory cytokines, we determined levels of IL-12 and IL-10; IL-12 augments IFN-γ production, whereas IL-10 suppresses its production.12- 15 However, we did not find any significant correlation between IL-12p40/IL-10 levels and IFN-γ levels. Moreover, there was no significant difference between groups in IL-12p40/IL-10 levels or the ratio of IL-12p40/IL-10. Levels of IFN-γ are also affected by various other mediators, including IL-18, transforming growth factor β, nitric oxide, lipid mediators, reactive oxygen species, etc.9,10,12,13 It is now generally agreed that optimal IFN-γ production requires the presence of IL-12 and IL-18.12 Had we measured levels of IL-12p40 and IL-18, we may have found the better correlation, but we were unable to measure IL-18 levels owing to lack of a commercially available ELISA kit. We also found that increased IFN-γ levels may be negatively correlated with IL-12p40 production by PBMN cells in other studies (H.J., S.S., and F.L.R., unpublished data, 1999-2000). In the patient group with elevated IFN-γ production (group 1), IL-12p40 production by SL cells may have already been suppressed, resulting in an apparent lack of correlation between IFN-γ and IL-12p40 production by SL cells.
When clinical features were compared between groups, we found a strikingly high frequency of allergic rhinitis in group 2. Namely, 12 of 12 group 2 patients undergoing ST for allergy or radioallergosorbent testing had positive reactions to common aeroallergens and had symptoms of allergic rhinitis. Six of them were also diagnosed as having asthma. In contrast, none of patients in group 1 received a diagnosis of allergic rhinitis, and results in 7 patients undergoing ST or radioallergosorbent testing were considered nonreactive. Taken together, low IFN-γ production by SL cells may indicate presence of allergic rhinitis in patients with CRS. However, we found no significant eosinophilia in the SL and BL samples in this group, despite clinical features of allergic rhinitis. This may partly be attributed to the frequent use of nasal steroid inhalers. However, a steroid nasal inhaler does not seem to deliver steroid into the sinus lumen.1 Alternatively, cytologic features of lavage specimen may not be closely associated with tissue eosinophilia in patients with CRS, as reported by others.25
In group 1, we found moderate to severe bacterial growth in 4 of 5 patients with positive bacterial cultures, whereas in group 2, light growth of bacteria was found in 7 of 12 patients with positive bacterial cultures. In group 1 patients with negative bacterial cultures, persistent bacterial growth may have triggered excessive T1 inflammatory responses mediated by T1 cytokines in the sinus. In group 1 patients with CRS and cultures negative for bacteria, proinflammatory T1 responses may not be properly down-regulated, perhaps in association with dysregulated mucosal immune defense. However, a much larger number of patients with CRS need to be studied to firmly establish an association between clinical features and/or bacterial culture results and cytokine production pattern.
Our results indicate that there may be distinguished subsets of patients with CRS without nasal polyposis on the basis of IFN-γ production by SL cells. However, such change is unlikely to be reflected in IFN-γ production by PBMN or BL cells. Levels of IFN-γ in the sinus are unlikely to be regulated solely by IL-10 and by IL-12p40, but are likely to be influenced by other various microenvironmental factors. High or low IFN-γ production by SL cells is likely associated with absence or presence, respectively, of atopic disorders (especially allergic rhinitis). The classification of patients with CRS in this way may help to evaluate further the effects of microenvironmental factors in the sinus and the efficacy of therapeutic agents.
Accepted for publication November 16, 1999.
This study was partly supported by grants from Lion's Multiple 5M Hearing Foundation and Minnesota Medical Foundation, Minneapolis.
Reprints: Harumi Jyonouchi, MD, Department of Pediatrics, University of Minnesota, Box 610 UMHC, 420 Delaware St SE, Minneapolis, MN 55455 (e-mail: firstname.lastname@example.org).