INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nocturnal worsening of pulmonary function is a common and potentially serious complication of asthma as 70% of deaths from asthma occur during the nighttime hours (1). We have shown that subjects with asthma who experience nocturnal worsening have greater airway inflammation at 4:00 A.M. than at 4:00 P.M. (4). The mechanisms for heightened nocturnal inflammation are not well understood. One possibility is the evening reduction in cortisol, which may allow the inflammatory cascade to proceed. However, serum cortisol levels do not differ between subjects with nocturnal asthma (NA) and nonnocturnal asthma (NNA), and normal control subjects (5).

We have previously shown that NA is associated with decreased glucocorticoid receptor (GR) binding affinity (6). However, this abnormality occurs at night only in NA, with normal GR binding affinity exhibited during the day (6). In addition, this reduced GR binding affinity is associated with reduced steroid responsiveness, as shown by the inability of dexamethasone (DEX) to significantly reduce peripheral blood mononuclear cells (PBMC) proliferation at night (6). Reduced GR binding affinity may be a potential mechanism of nocturnal inflammation, as glucocorticoids must bind to their receptor followed by receptor translocation into the nucleus in order to bind to specific DNA sequences called glucocorticoid-responsive elements (GREs). Many of the glucocorticoid-inducible genes that have been identified are characterized by a cluster of multiple GREs located in their promoter and enhancer regions. The induction or repression of GR target genes ultimately results in the modulated expression of glucocorticoid-regulated proteins (7, 8). We hypothesize that the functional consequence of reduced GR binding affinity is a state of reduced steroid responsiveness, where endogenous and exogenous corticosteroids are unable to reduce cellular proliferation and production of mediators effectively.

However, we do not know if these events occur in the airway, and, if so, the mechanisms driving this alteration in steroid responsiveness. Interestingly, inflammation itself, particularly cytokines, have been shown to induce reductions in GR binding in PBMC (9, 10). These cytokines affect GR binding affinity in specific cells, as interleukin (IL)-2 and IL-4 have been shown to reduce GR binding affinity of T lymphocytes, whereas IL-13 induces a reduction in the GR binding affinity of monocytes (9, 10). In addition, IL-2 and IL-4 have been shown to induce the expression of an alternative form of the GR, GR (11). Although IL-13 has been shown to reduce GR binding affinity particularly in the peripheral blood monocyte population, its ability to induce GR has not been studied, nor is it known whether IL-13 exerts effects on alveolar macrophages. This form of the receptor, which arises from alternative splicing of the GR pre-mRNA transcript, is unable to bind to steroid and be transcriptionally active due to replacement of the last 50 amino acids in its COOH terminus with a unique 13 amino acid sequence (12, 13). Because GR is felt to compete and antagonize the effects of the active receptor, GR (14), its presence in asthma may explain the reduced steroid responsiveness seen in these subjects at certain times of day. The presence of daytime GR has been demonstrated in patients with glucocorticoid-insensitive asthma (11, 15). These patients demonstrated significantly higher expression of GR as compared with glucocorticoid-sensitive patients with asthma and normal control subjects. The expression of the active receptor, GR, by these patients is not known.

In this study, we tested the hypothesis that in NA, steroid responsiveness in the airway is decreased at night and associated with increased GR expression. Unexpectedly, the steroid insensitivity and increased GR expression in NA were localized to the airway macrophage.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

A group of 10 subjects with NA and 7 subjects with NNA was recruited from the general Denver community. All subjects met diagnostic criteria for asthma (16), exhibiting a methacholine PC₂₀ of less than 8 mg/ml and were maintained on inhaled ₂-agonists only. The group with nocturnal symptoms demonstrated a fall in peak expiratory flow rate (PEFR) of greater than 15% from bedtime to morning on at least four nights over a 7-d period of testing at home prior to study. The NNA group experienced an overnight fall in PEFR of 10% or less. For consistency, subjects were placed on a regimen of two puffs of inhaled ₂-agonists four times daily. Exclusion criteria included use of cromolyn, nedocromil, or inhaled or oral corticosteroids within the previous 6 wk, an upper respiratory tract infection within the previous 6 wk, immunotherapy within the previous 3 mo, cigarette use, and significant nonasthma pulmonary disease or other medical illnesses. Informed consent was obtained from all patients for this institutional review board-approved protocol.

Protocol

Subjects underwent two bronchoscopies with bronchoalveolar lavage (BAL) at 4:00 P.M. and 4:00 A.M., in a random order 1 wk apart as previously described (17). The site of lavage was randomized to either the right middle lobe or lingula for the first bronchoscopy, with the other segment lavaged during the second bronchoscopy. This method has been employed in several of our previous studies (4, 17) as bronchoscopy with lavage itself can produce visual changes in the airway, particularly edema. This edema may affect the following bronchoscopy, so we prefer an approach in which the first location is determined randomly, followed by the second lavage in the lung that has previously not undergone lavage. Prior to the 4:00 A.M. bronchoscopies, subjects spent the night in our clinical research laboratory and underwent spirometry prior to bedtime (10:00 P.M.). Prior to each bronchoscopy, spirometry was performed before and after 0.18 mg of albuterol from a metered dose inhaler (MDI) and 0.4 mg atropine were administered intravenously. Xylocaine (4%) was used to anesthetize the upper airway and xylocaine (1%) was applied to the laryngeal area, trachea, and orifice of the right lower or left lower lobe bronchi via the bronchoscope. The amount of lidocaine used to anesthetize the upper and lower airway was recorded, and serum was sampled 30 min after anesthesia to obtain lidocaine levels. Bronchoalveolar lavage was performed using five 60-ml aliquots of sterile normal saline at 37° C. Lavage fluid was obtained by immediate gentle hand suction applied to each instilling syringe. Supplemental oxygen was administered throughout the procedure along with monitoring of heart rate and oxygen saturation. Subjects' vital signs were monitored in our laboratory after the procedure. Each subject underwent two bronchoscopies only.

BAL Cell Protocol

The lavage fluid was put on ice immediately, and the aliquots were combined and centrifuged for 10 min at 1500 rpm at 4° C after 10 µl was removed for cell count. After the supernatant was aspirated, cells were resuspended in a serum-free media at a concentration of 1 × 10⁶/ml. These cells were then plated at 5 × 10⁴ cells/well in a 48-well plastic culture plate and allowed to adhere for 1 h at 37° C/5% CO₂. After this adherence, the nonadherent fraction was aspirated; the adherent cells (macrophage enriched) were washed three times with cold phosphate-buffered saline (PBS) and 0.5 ml of new media (RPMI/0.1% bovine serum albumin) was then added. Cells were stimulated with zymosan (100 particles/cell) in the presence and absence of dexamethasone (DEX) at concentrations of 10⁵-10⁹ M (Sigma Chemical Co., St. Louis, MO) and incubated for 24 h at 37° C/5% CO₂. Supernatant was aspirated and frozen at 70° C until measurement of tumor necrosis factor (TNF)- and interleukin (IL)-8 by enzyme-linked immunoassay (R&D Diagnostics, Minneapolis, MN). Data are expressed in picograms per milliliter. The lower limit of detection for both ELISAs is < 15 pg/ml.

The nonadherent lymphocyte-enriched cell population was plated at a concentration of 5 × 10⁴ cells/well and stimulated with phytohemaglutinin (5 µg/ml) in the presence and absence of DEX at concentrations of 10⁵-10⁸ M (Sigma). After incubation at 37° C/5% CO₂ for 48 h, [³H]thymidine (5 mCi/ml) was then added to each well and incubated for 12 h. The cells were then harvested onto glass fiber filter plates and [³H]thymidine incorporation was measured. Proliferation results are expressed as percentage of positive control.

GR and GR Expression

GR expression was determined by fixation of acetone/methanol-fixed cytospins of BAL cells using a GR-specific polyclonal rabbit antibody raised against human GR at a dilution of 1:100. This antibody has previously shown to be specific for GR (20) and demonstrates no cross-reactivity against GR, the active GR. The primary antibody was replaced with nonspecific rabbit immunoglobulin as the negative control. Cytospins were prepared by removing 10 ml of whole BAL prior to the adherence step and resuspending BAL cells in PBS at a concentration of 7.5 × 10⁵/ml. Then 75 µl was placed in a Shandon cytocentrifuge tube and centrifuged at 1,000 rpm for 1 min. This technique resulted in a concentration of approximately 56,250 cells/ slide. The percentage of cells positive for GR was determined by a blinded assessor counting a minimum of 1,000 cells. GR expression was also determined in the nocturnal asthma group at 4:00 P.M. and 4:00 A.M. using methanol-fixed BAL cytospins as described above. The antibody used was a rabbit anti-human polyclonal antibody (Affinity Bioreagents, Inc., Golden, CO).

Double immunostaining of acetone/methanol-fixed airway cells from subjects with NA and NNA and control subjects was performed as previously described in tissue by Minshall and colleagues (21). A mixture of primary antibodies CD68 (macrophages; Pharmacia, Uppsala, Sweden) at 1:50 dilution, CD3 (T cells; Pharmacia) at 1:100 dilution, neutrophil elastase (Pharmacia) at 1:100 dilution, and EG2 (eosinophils; Pharmacia) at 1:100 dilution was applied to cytospins after endogenous peroxidase activity in tissue and airway was blocked with 1% H₂O₂ (plus 0.02% sodium azide in tris-buffered saline [TBS]) for 30 min. After incubating with the appropriate secondary biotinylated antibodies, a tertiary layer of streptavidin peroxidase conjugate was then applied and developed by diaminobenzidine to produce brown staining. The GR-positive cells were developed with an alkaline phosphatase system and visualized with fast red, which produces red staining. The percentage of cells positive for both markers on each slide were counted by two blinded assessors counting a minimum of 1,000 cells. The results were expressed as a percentage of the particular cell expressing GR, as well as the percentage of the total number of cells expressing GR.

IL-13 In Situ Hybridization and ELISA

In situ hybridization was performed in parformaldehyde-fixed cytospins from all subjects as previously described (22). Briefly, ³⁵S-UTP-labeled RNA probes were prepared from cDNA for IL-13. To avoid nonspecific binding of ³⁵S-labeled RNA probes, incubation with N-ethyl maleimide, iodoacetamide, and triethanolamine was included in prehybridization steps, and dithiothreitol (150 mM was included in the hybridization mixture [Sigma]). As a negative control, preparations were hybridized with a sense IL-13 probe (having identical sequence to IL-13 mRNA). To further ensure the specificity of the signal, separate sets of preparations were pretreated with RNase solution (Promega, Southampton, UK) prior to hybridization with antisense probe.

Counting was performed in a blinded fashion without knowledge of asthma status. Hybridization between cytokine mRNA and cRNA probes was identified as dense collections of sliver grains overlying cells. Hybridization signals were assessed by counting mRNA-positive cells/1,000 total BAL cells on at least two slides for each cytokine probe and counts were performed in duplicate. The hybridization signal was considered to be specific when a positive signal was obtained with antisense probe, but not with the sense probe, and when the positive signal was abolished with the pretreatment with RNase.

Measurement of IL-13 protein levels was performed using enzyme-linked immunosorbent assay (ELISA) (R&D Diagnostics). The sensitivity is 32 pg/ml.

Neutralizing Antibody Experiments

Given the preferential effect of IL-13 on GR binding affinity of the monocyte population, BAL macrophages from five subjects with NA at 4:00 A.M. were incubated with neutralizing antibody to IL-13 (anti-IL-13, R&D Diagnostics) at 1 µg/ml for 48 h at 37° C/5% CO₂ using four-well microtiter slides (LabTek, Naperville, FL). The slides were converted to cytospin preparations and double immunostained with CD68 and GR. Again, the number of cells positive for CD68 and GR/1,000 BAL cells on at least two slides and counts were performed in duplicate.

Statistical Analysis

The FEV₁ at 4:00 P.M., 4:00 A.M., the percentage overnight fall in FEV₁, the expression of GR and the expression of IL-13 mRNA were compared within and between the two groups using the paired and unpaired t tests as the data were normally distributed as determined by the Shapiro-Wilk test (25). The steroid dose-response curves for BAL lymphocyte proliferation and macrophage production of IL-8 and TNF- at each time point were compared for all groups using a mixed effects model with fixed effects for time, group, and the interaction between time and group, plus a random subject effect (26- 28). If the interaction was significant, pairwise comparisons of least-squares means were made within groups and within times. All tests were two sided and conducted at the 5% significance level. Data are expressed as mean ± SEM.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Characteristics

The subject characteristics of our two groups are shown in Table 1. The percentage predicted and absolute FEV₁ were similar between the two groups at 4:00 P.M. (p = 0.81 and 0.50 for percentage and absolute FEV₁, respectively). At 4:00 A.M., the percentage predicted FEV₁ decreased in the NA group to 66.3 ± 3.4% as compared with 82.7 ± 4.8% in the NNA group (p = 0.02). The median lidocaine level at 4:00 P.M. in the NA group was 1.1 (IQ: 0.6-2.4) mg/L and 1.1 µg/ml (IQ: 0.52-1.18) in the NNA group, p = 0.72. The median amount of lidocaine used was 520 mg (IQ: 500-700) in the NA group and 520 mg (IQ: 525-565) in the NNA group, p = 0.85. At 4:00 A.M., the median lidocaine level was 0.4 mg/L (IQ: 0.4-0.6) mg/L in the NA group versus 0.6 mg/L (IQ: 0.5-0.7) in the NNA group, p = 0.32. The median amount of lidocaine used in each group was 500 mg (400-600) in the NA group versus 600 (480-720) in the NNA group, p = 1.00. 

BAL Return and Cell Counts

The BAL return was significantly different between the two groups (57.1 ± 3.6% versus 45.7 ± 4.9% in the NNA and NA groups at 4:00 P.M.; p = 0.037; 65.5 ± 6.2% versus 34.7 ± 4.9% in the NNA and NA groups at 4:00 P.M.; p = 0.002). There were no significant differences in the total white cell and white cell count/ml between the two groups at either time point (4:00 P.M.: 9.0 ± 2.0 × 10²/ml versus 11 ± 2 × 10² cells/ml in the NNA and NA groups, respectively, p = 0.74; 4:00 A.M.: 8.0 ± 2 × 10² cells/ml versus 11 ± 3 × 10² cells/ml in the NNA and NA groups, p = 0.14). In addition, the percentages of macrophages, lymphocytes, neutrophils, and eosinophils did not differ between the two groups at 4:00 P.M. and 4:00 A.M., or within each group at each time point. The numbers of cells obtained by BAL were adequate for the experiments, as the mean total white cell count when the two groups were combined was 16.3 ± 0.1 × 10⁶ at 4:00 P.M. and 12.7 ± 0.1 × 10⁶ at 4:00 A.M. Macrophages composed a mean percentage of 86.7 ± 1.5% at 4:00 P.M. and 85.5 ± 1.2% at 4:00 A.M. Lymphocytes composed 11.0 ± 1.3% at 4:00 P.M. and 10.7 % at 4:00 A.M.

Lymphocyte Proliferation Studies

The suppression of proliferation by DEX was not significantly different at 4:00 A.M. as compared with 4:00 P.M. within both groups (p = 0.28 and 0.72 within the NNA and NA groups, respectively).

Macrophage Products

Within the NA group, the production of IL-8 and TNF- by BAL macrophages was not suppressed as effectively by DEX at 4:00 A.M. and compared with 4:00 P.M. (p = 0.0001 for IL-8 and TNF-, respectively) (Figure 1). Within the NNA group, there was no significant difference in steroid responsiveness between the two time points with DEX (p = 0.82 and 0.81 for IL-8 and TNF-, respectively) (Figure 1). Suppression of TNF- and IL-8 at 4:00 P.M. was greater than 4:00 A.M. specifically in the NA group, and greater at 4:00 A.M. in the NNA group compared with NA. There were no differences within the NNA group, and between the groups at 4:00 P.M.

View larger version (26K): [in this window] [in a new window]   Figure 1.   Suppression of BAL macrophage production of TNF- and IL-8 with dexamethasone (DEX) in nocturnal asthma (NA) and nonnocturnal asthma (NNA) is shown for both groups at 4:00 P.M. and 4:00 A.M. Within the NA group, p = 0.0001 for IL-8 and TNF-, respectively, with decreased suppression of cytokine production by dexamethasone observed at 4:00 A.M. as compared with 4:00 P.M.; within the NNA group, p = 0.68 and 0.73 for IL-8 and TNF-, respectively, with no change in suppression of cytokine production observed at 4:00 P.M. as compared with 4:00 A.M.

When the analysis is performed with only the physiologically relevant concentrations of dexamethasone (10⁸ M and 10⁹ M) where the lines diverge within the NA group, the results are consistent with a time effect in the NA group, such that suppression is decreased with DEX at 4:00 A.M. as compared with 4:00 P.M. (p = 0.0001). This difference is not significant within the NNA group for IL-8 (p = 0.97) or TNF- (p = 0.70).

GR and GR Expression

The percentage of airway cells expressing GR was significantly higher at night in the NA group (48.0 ± 4.5% versus 32.0 ± 3.2% at 4:00 A.M. and 4:00 P.M., p = 0.008) (Figure 2). There was no difference in GR expression in the NNA group at either time point (21.3 ± 2.8% versus 20.2 ± 2.9% at 4:00 A.M. and 4:00 P.M., p = 0.10) (Figure 2). There was no difference in GR expression between the NNA and NA groups at 4:00 P.M. (p = 0.13).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Percentages of bronchoalveolar lavage (BAL) cells expressing glucocorticoid receptor-beta (GR) in nocturnal asthma (NA) and nonnocturnal asthma (NNA). The NA are represented by the closed bars, and the NNA are represented by the hatched bars. *p < 0.05 within the NA group at 4:00 P.M. and 4:00 A.M.  p < 0.05 between the NA and NNA group at 4:00 A.M.

A photomicrograph illustrating BAL cells from subjects with NA and NNA at 4:00 A.M. double immunostained for GR and CD68 expression is shown in Figure 3. Double immunostaining revealed that BAL macrophages exhibited the greatest expression of GR, and that their expression increased at night only in the NA group (NNA: 57.0 ± 2.3% at 4:00 P.M. versus 53.0 ± 2.2% at 4:00 A.M., p = 0.10; NA: 54.0 ± 1.9% at 4:00 P.M. versus 62.0 ± 1.9% at 4:00 A.M., p = 0.008) (Figure 4). The expression of GR was higher in the NA group at 4:00 A.M. as compared with the NNA group (p = 0.03), but there was no difference in macrophage GR expression between the two groups at 4:00 P.M. (p = 0.36). The eosinophils in the NA group, but not the NNA group, also increased their expression of GR, although the change was small (6.3 ± 0.8% at 4:00 P.M. versus 9.0 ± 1.2% at 4:00 A.M., p = 0.05) (Figure 4). There was no change in GR expression from 4:00 P.M. to 4:00 A.M. in the T cells and neutrophils of either asthma group.

View larger version (97K): [in this window] [in a new window]   Figure 3.   Photomicrographs of BAL cells from a subject with NA (A) and NNA (B) at 4:00 A.M. double immunostained with GR and CD68. The small arrow in A denotes a cell positively staining for CD68, and negative for GR; the large arrow in B denotes a cell positively staining for both CD68 and GR. (C ) BAL stained with GR and CD68 from a subject with NA at 4:00 A.M. Original magnification for each photomicrograph: ×400.

View larger version (26K): [in this window] [in a new window]   Figure 4.   Percentages of bronchoalveolar lavage (BAL) neutrophils (neuts), eosinophils (eos), T cells, and macrophages (macs) expressing GR at 4:00 P.M. and 4:00 A.M. within the nonnocturnal asthma group (NNA, hatched bars) and the nocturnal asthma group (NA, closed bars). *p < 0.05 for GR expressed by eosinophils within the NA group at 4:00 P.M. and 4:00 A.M.  p < 0.05 for GR expression by macrophages within the NA group at 4:00 P.M. and 4:00 A.M.  p < 0.05 for GR expression by macrophages between the NA and NNA groups at 4:00 A.M.

Expression of GR in BAL did not change significantly from 4:00 P.M. to 4:00 A.M. in subjects with nocturnal asthma (GR expression [%]: 55.7 ± 2.9% at 4:00 P.M. versus 55.8 ± 0.8% at 4:00 A.M., p = 0.99).

Role for IL-13 in Macrophage Expression of GR

Because IL-13 has been reported to induce glucocorticoid insensitivity in monocytes (10), we examined IL-13 gene expression in NA versus NNA and the ability of IL-13 neutralization to decrease GR expression. The number of BAL cells/1,000 expressing IL-13 mRNA was increased in the NA group at 4:00 A.M. as compared with 4:00 P.M. (80.6 ± 9.3 versus 102.3 ± 15.2 at 4:00 P.M. and 4:00 A.M., p = 0.02) (Figure 5). There were no differences in expression of IL-13 mRNA expression within the NNA group (44.7 ± 6.5 versus 47.8 ± 5.6 at 4:00 P.M. and 4:00 A.M., p = 0.43) (Figure 6). IL-13 was not detected in either group by ELISA in BAL. In the presence of IL-13 neutralizing antibody at 4:00 A.M. in the NA group, the expression of GR was initially 48.3 ± 3.4% prior to incubation, and decreased to 31.3 ± 3.8% in the presence of anti-IL-13 (p = 0.0012).

View larger version (158K): [in this window] [in a new window]   Figure 5.   Representative example of in situ hybridization for interleukin (IL)-13 mRNA in the bronchoalveolar lavage preparation from a subject with nocturnal asthma (NA) at 4:00 A.M. (a), a subject with NA at 4:00 P.M. (b), and a subject with nonnocturnal asthma at 4:00 A.M. (c) using the antisense probe. The arrows indicate positive IL-13 mRNA signals. (d ) In situ hybridization for IL-13 mRNA using the sense or control probe. Original magnification: ×400.

View larger version (17K): [in this window] [in a new window]   Figure 6.   The number of bronchoalveolar (BAL) cells expressing interleukin-13 mRNA (IL-13), reported as number per 1,000 BAL cells, in nocturnal asthma (NA) and nonnocturnal asthma (NNA). The NNA are represented by the hatched bars, and the NA are represented by the closed bars. *p < 0.05 between NA and NNA groups at 4:00 P.M.  p < 0.05 within the NA group from 4:00 P.M. to 4:00 A.M.  p < 0.05 between the NA and NNA groups at 4:00 A.M.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we have shown for the first time that airway cells from subjects with NA exhibit a circadian variation in steroid responsiveness. This was illustrated by the decreased ability of corticosteroids to suppress production of cytokines from BAL macrophages at 4:00 A.M. as compared with 4:00 P.M. This alteration in steroid responsiveness was associated with increased expression of GR by BAL macrophages and to a lesser extent BAL eosinophils. In addition, blockade of IL-13 resulted in decreased expression of GR by BAL macrophages. These results are distinctly different from what we have shown in peripheral blood, where DEX did not suppress all PBMC effectively, which includes lymphocytes and monocytes (6). In the airway at night, the alveolar macrophage appears to be the least steroid responsive cell. These results are also different from those seen in glucocorticoid insensitive asthma, where there was significantly greater airway expression of GR by the T cell (95-100%) as compared with the macrophage, where 15-30% of macrophages expressed GR (15).

The association of increased GR expression with reduced steroid responsiveness suggests that GR may be involved in producing a state of relative steroid resistance. However, the mechanism by which GR may produce this state is not well understood. Previous studies have demonstrated that increased expression of GR by transfection of the pRShGR plasmid into various cell lines induces glucocorticoid insensitivity (11, 14). There are several mechanisms for this action that have been hypothesized: its role as a dominant negative inhibitor of GR function via formation of GR-GR heterodimers, occupation of GRE binding sites by nontransactivating GR/GR homodimers, or squelching, that is, titration of limiting amounts of accessory proteins or coactivators of GR. The latter mechanism has been suggested by several investigators (29), but not shown by others (14). In regard to the formation of GR-GR heterodimers and the dominant negative function of GR, Oakley and colleagues have shown in multiple cell types that GR does form heterodimers with GR and inhibits GR-mediated activation of a mouse mammary tumor virus promoter (32). GR also inhibits GR-mediated repression of an NF-B responsive promoter, but does not interfere with homologous down-regulation of GR (32). They also demonstrated that GR can associate with the heat shock protein 90 in the cytosol, but with lower affinity than GR and bind to the GRE with greater affinity than GR (32). Bamberger and coworkers also suggested that a dominant negative function of GR was its mechanism of action, although no binding to the GRE was seen when GR and GR were coexpressed (14).

Despite data that the dominant negative function may indeed be the mechanism of action for GR (14, 32), three reports (29) argue against this concept. de Lange and colleagues demonstrated that human GR did not exert a dominant negative effect, but only a nonspecific repression of transcriptional activity in general (31). Hecht and colleagues illustrated that GR levels are much lower than GR, and that GR does not repress GR-mediated activation of the MMTV promoter of the COS-7 cells (30). Brogan and coworkers reported that GR does not affect GR-mediated repression of the IL-2 promoter and AP-1 or NF-B activities (29). Thus, these investigators did not find a dominant negative action of GR. However, these discrepant results could be explained by the use of different expression vectors, differences in cell/tissue specificity, or inadequate GR expression connected to the transient transfection systems they used. The generation of transgenic animals and stable cell lines expressing varying levels of GR isoforms are needed to resolve these issues.

Additionally, for a dominant negative inhibitor to exert its effect, the concentration must be adequate to perform this duty. de Castro and colleagues showed that the level of GR protein was between being equal to GR or up to five times higher and when not exposed to dexamethasone, is located in the cytosol (20). However, Oakley and coworkers demonstrated that GR concentration was lower than GR in whole lung tissue, but varied significantly according to the cell type studied (33). They also showed accumulation of the GR primarily in the nucleus but also in the cytosol (33). As shown in Figure 3, GR immunoreactivity detected appeared to be in the cytosol, as shown by de Castro and coworkers (20).

Inflammation, particularly IL-13 and the combination of IL-2 and IL-4, has been shown to induce reductions in GR binding affinity and stimulate GR expression in peripheral blood (9, 10). The finding of increased GR in the airway at night in NA may be a marker of reduced steroid responsiveness, which was functionally exhibited by the inability of DEX to decrease TNF- and IL-8 production by macrophages as effectively at night as compared with daytime. In this study, GR was expressed in the highest numbers by macrophages and the expression was modulated by IL-13. We speculate that this is due to the preferential effect of IL-13 on the macrophage at night, and this cell drives the reduction in steroid responsiveness we see in NA.

The function of the macrophage in the normal airway is to be involved in several roles in its defense of the local environment. Evidence supports the hypothesis that macrophages exert a protective effect on the local milieu by preventing an immunological overreaction to the large amounts of inhaled antigens (34). This suppressor activity is thought to be reduced in patients with asthma, where macrophages initiate and prolong allergic and inflammatory processes (35, 36). Furthermore, heterogeneity of the macrophage population has been shown to exist in normal subjects and patients with asthma (37, 38) but little is known about the steroid responsiveness of these populations. The observation that macrophages are perhaps the least steroid responsive cell at night in NA is an attractive hypothesis, in that the ability of the macrophage to participate and perpetuate the inflammatory response is greater at night than the daytime.

IL-13 has received attention recently due to its potentially important role in asthma pathogenesis (39). Zhou and colleagues recently demonstrated that pulmonary IL-13 expression in a transgenic murine model resulted in a complex phenotype that included many of the features of chronic asthma: eosinophilic tissue inflammation, epithelial hypertrophy, mucus hypersecretion, subepithelial airway fibrosis, Charcot-Leyden crystal-like deposition, airway obstruction, and airway responsiveness to methacholine (41). IL-13 has been shown in two recent publications to be critical to the formation of the asthma phenotype in a murine model (39, 40). Furthermore, the development of the asthma phenotype was mediated through the alpha subunit of the IL-4 receptor, IL-4R. Characteristics of the asthma phenotype (airway hyperresponsiveness, eosinophil recruitment, and mucus overproduction) did not occur in mice deficient in IL-4R (40). This shared receptor may explain why IL-13 also shares several biological activities with IL-4, but acts primarily in the monocyte population to regulate surface antigen expression, antibody-dependent cellular cytotoxicity, and cytokine synthesis (42). We hypothesize that IL-13 acts preferentially on the alveolar macrophage, and may contribute to the circadian change in lung function seen in asthma. We realize that in situ hybridization is a semiquantitative tool, but we feel our results reflect a change in the number of cells expressing IL-13, thus new production of IL-13 mRNA. We do not feel that the lack of detection of IL-13 protein refutes our hypothesis. BAL may not be the best sampling method for detection as it is very dilute and much more easily measured in allergen challenge models, for which nocturnal asthma is not a surrogate.

Although Turner-Warwick demonstrated that nocturnalworsening of asthma is a common aspect of asthma in her survey of over 7,000 outpatients (46), the question frequently arises as to whether patients with NA simply have more severe asthma. In the NA subjects recruited for our studies, lung function during the day is sometimes reduced or in the normal range and exhibit predictable nocturnal declines in lung function (4, 47). In this study, daytime lung function between the NA and NNA groups was similar (Table 1). The changes appreciated only at night in NA, such as increased airway inflammation and decreased steroid responsiveness, can also be seen in patients with severe asthma. However, these immune alterations are observed during the day in this latter group (48, 49). In addition, Wenzel and colleagues compared airway and parenchymal inflammation in patients with severe asthma as compared with a subset of our NA group, and found an increase in alveolar tissue neutrophils in the severe group (48). This was not appreciated in the NA, where increased macrophages and eosinophils were noted. It is of interest that these cell types increase their expression of GR at night. We hypothesize that in patients with asthma who have a significant nocturnal component, physiological triggers such as airway cooling and/or allergen exposure, coupled with the evening fall in cortisol, results in enhanced expression of IL-13. IL-13 induces GR expression and a state of relative steroid insensitivity. As cortisol increases overnight, IL-13 decreases, along with GR expression. Thus, patients with asthma with nocturnal worsening provide a unique "model" in which to study physiological inflammatory mechanisms in humans.

In regard to steroid responsiveness in general, there may be other mechanisms that could be involved in reducing steroid responsiveness in our subjects with NA that were not studied. These include interaction of GR-GR or GR-GR complexes with transcription factors AP-1 or NF-B (50), receptor modification, as has been shown by nitrosylation via inducible nitric oxide (54) or export of the corticosteroid out of the cell via membrane transport proteins (54). The effects of the transcription factors AP-1 and NF-B are particularly relevant, as both have been shown to regulate TNF- and IL-8 expression (55, 56).

We conclude that altered steroid responsiveness may lead to the presence of NA. This altered steroid responsiveness occurs through the airway macrophage and is modulated by IL-13. This state of reduced steroid responsiveness appears to be present even in the patients with asthma not considered severe by clinical criteria during daytime hours. These observations may explain why a spectrum of responses of antiinflammatory therapy occurs in asthma, and how we can increase our knowledge of inflammatory mechanisms in asthma pathogenesis by studying this form of asthma in humans.