Comments Abstract
Rationale: The role of airway microbiome in corticosteroid response in asthma is unknown.
Objectives: To examine airway microbiome composition in patients with corticosteroid-resistant (CR) asthma and compare it with patients with corticosteroid-sensitive (CS) asthma and normal control subjects and explore whether bacteria in the airways of subjects with asthma may direct alterations in cellular responses to corticosteroids.
Methods: 16S rRNA gene sequencing was performed on bronchoalveolar lavage (BAL) samples of 39 subjects with asthma and 12 healthy control subjects. In subjects with asthma, corticosteroid responsiveness was characterized, BAL macrophages were stimulated with pathogenic versus commensal microorganisms, and analyzed by real-time polymerase chain reaction for the expression of corticosteroid-regulated genes and cellular p38 mitogen-activated protein kinase (MAPK) activation.
Measurements and Main Results: Of the 39 subjects with asthma, 29 were CR and 10 were CS. BAL microbiome from subjects with CR and CS asthma did not differ in richness, evenness, diversity, and community composition at the phylum level, but did differ at the genus level, with distinct genus expansions in 14 subjects with CR asthma. Preincubation of asthmatic airway macrophages with Haemophilus parainfluenzae, a uniquely expanded potential pathogen found only in CR asthma airways, resulted in p38 MAPK activation, increased IL-8 (P < 0.01), mitogen-activated kinase phosphatase 1 mRNA (P < 0.01) expression, and inhibition of corticosteroid responses (P < 0.05). This was not observed after exposure to commensal bacterium Prevotella melaninogenica. Inhibition of transforming growth factor-β–associated kinase-1 (TAK1), upstream activator of MAPK, but not p38 MAPK restored cellular sensitivity to corticosteroids.
Conclusions: A subset of subjects with CR asthma demonstrates airway expansion of specific gram-negative bacteria, which trigger TAK1/MAPK activation and induce corticosteroid resistance. TAK1 inhibition restored cellular sensitivity to corticosteroids.
Culture-independent microbiota profiling based on sequence polymorphisms in the 16S rRNA present in all bacteria is widely applied in human microbiome studies. To date, limited knowledge about the airway microbiome and its role in health and disease is available.
This study provides evidence that a subset of subjects with corticosteroid-resistant asthma demonstrate expansion of specific gram-negative bacteria in the airways. Moreover, airway cell stimulation by bacteria in the airway microbiome can inhibit corticosteroid responses, potentially influencing the efficacy of corticosteroid treatment.
Current guidelines (1, 2) recommend the use of corticosteroids to control airway inflammation in persistent asthma. Clinical studies, however, demonstrate highly variable responses to corticosteroid therapy, with up to 45% of patients responding suboptimally to inhaled corticosteroids (3–5), and up to 25% of patients do not respond to oral corticosteroids (6). These corticosteroid-resistant (CR) patients demonstrate airway inflammation (6, 7) and remodeling (8), which contribute to the severe asthma phenotype often observed in these patients. For those patients with variable corticosteroid response, alternative treatment strategies are needed to improve asthma control (9).
Recently, 16S rRNA gene sequence analyses of airway specimens have demonstrated diversified microbial communities in the airways of subjects with asthma (10, 11). When compared with healthy patients, subjects with asthma manifest significantly higher bacterial burden and diversity caused by expansion of pathogenic bacteria. Moreover, airway microbiome composition and diversity correlate with bronchial hyperresponsiveness (11). The functional and pathobiologic impact of the airway microbiome in asthma remains unknown. In the current study, we hypothesized that expanded pathogenic microbial communities residing in the airways of subjects with CR asthma may inhibit cellular responses to corticosteroids.
Some of the results of these studies have been previously reported in the form of an abstract (12, 13).
Thirty-nine adult subjects with asthma and 12 healthy control participants were enrolled. All subjects with asthma had baseline FEV1 percent predicted less than or equal to 85% with either airway hyperresponsiveness (PC20 methacholine, <10 mg/ml) or bronchodilator responsiveness (>12% improvement in FEV1% predicted after inhalation of 180 μg albuterol), as previously described (8, 14–16). The Juniper Asthma Control Questionnaire was used to assess asthma control. All subjects with asthma underwent bronchoscopy, followed by prednisone, 20 mg by mouth twice daily for 7 days. Subjects with asthma were categorized as corticosteroid sensitive (CS) if the FEV1% predicted value improved greater than or equal to 15%, and as CR if the FEV1% predicted value improved less than 10%. Healthy participants underwent bronchoscopy but did not receive prednisone. Smokers were excluded from participation in this study. Informed consent was obtained from all study participants before enrollment in this study. The Institutional Review Board at National Jewish Health approved this study. Details of patient characteristics are presented in Table 1.
| Normal Control Subjects (n = 12) | CR Asthma (n = 29) | CS Asthma (n = 10) | |
|---|---|---|---|
| Age, yr (mean ± SD) | 31.1 ± 9.2 | 34.2 ± 11.1 | 37.9 ± 10.8 |
| Sex, male/female | 4/8 | 14/15 | 2/8 |
| Race, white/black/other | 11/0/1 | 19/4/6 | 9/0/1 |
| Body mass index, kg/m2 (mean ± SD) | 24.1 ± 4.7 | 26.3 ± 6.4 | 32.2 ± 8.5 |
| IgE, U/ml (mean ± SD) | 75 ± 108 | 253 ± 289 | 177 ± 211 |
| Number of positive skin tests | 0.1 ± 0.3 | 6 ± 4 | 6 ± 4 |
| PC20, methacholine, mg/ml (mean ± SD) | >25 | 1.5 ± 2.2 | 0.5 ± 0.6 |
| ACQ score (mean ± SD) | NA | 1.5 ± 0.8 | 1.9 ± 0.6 |
| eNO, ppm (mean ± SD) | NA | 34.9 ± 24.8 | 47.3 ± 29.2 |
| Baseline FEV1% predicted (mean ± SD) | 98.7 ± 11.7 | 76.3 ± 10.3 | 61.9 ± 16.1 |
| FEV1% reversal with albuterol (mean ± SD) | 5.1 ± 2.4 | 15.0 ± 9.4 | 37.2 ± 24.7 |
| FEV1% change after prednisone burst (mean ± SD) | NA | −0.2 ± 5.9* | 31.0 ± 23.5 |
| Corticosteroid medications† | |||
| ICS/LABA | NA | 5 | 4 |
| ICS | 7 | 0 | |
| None | 17 | 6 |
Fiberoptic bronchoscopies with bronchoalveolar lavage (BAL) were performed according to the guidelines of the American Thoracic Society (17–19). BAL cells were prepared as described (14). BAL cell suspensions (250 μl containing 1 × 106 cells/ml) were preserved at −80°C before DNA extraction. Presence of LPS in the BAL fluid samples was analyzed by chromogenic limulus amebocyte lysate test as previously described (20).
BAL differentials were obtained on cytospin preparations using a Diff-Quick (Scientific Products, McGraw Park, IL) stain, counting a minimum of 500 cells.
Details of bacterial DNA preparation from BAL samples, sequencing, and sequencing data analysis are presented in the online supplement.
Haemophilus parainfluenzae (ATCC number 9796) and Prevotella melaninogenica (ATCC number 25845) were obtained from ATCC (Manassas, VA) and cultured according to ATCC instructions (see online supplement).
In a separate group of subjects with asthma (n = 7) (see Table E1 in the online supplement) BAL and peripheral blood monocytes were collected to examine functional interaction with bacteria. Details of the cells culture conditions and bacterial stimulation are shown in the online supplement.
The effect of bacteria on p38 mitogen-activated protein kinase (MAPK) activation and cellular response to corticosteroids and the influence of selected pathway inhibitors on corticosteroid responses were assessed (see online supplement). Statistical data analyses are described in the Methods section of the online supplement.
In the current study, we recruited 39 subjects with asthma and 12 healthy control subjects for lung microbiome evaluation. Study subject characteristics are summarized in Table 1. Subjects with asthma were divided into CR and CS groups based on FEV1% predicted responses after a 1-week burst with oral prednisone. Both groups had comparable baseline FEV1% predicted. Patients in the CR group did not show any improvement in FEV1 after exposure to prednisone; in contrast, patients in the CS group showed significant improvement in lung function after steroid burst (ΔFEV1% after oral prednisone burst was −0.2 ± 5.9% and 31.0 ± 23.5% for subjects with CR and CS asthma, respectively; P < 0.0001) (Table 1).
BAL sample differentials are presented in Table E2. BAL samples consisted mostly of macrophages (>80%), with significantly higher levels of eosinophils in subjects with asthma as compared with normal control subjects (P < 0.01) (see Table E2). No significant differences were noted in the number of macrophages, lymphocytes, and neutrophils in BAL samples between groups (see Table E2).
Bacterial 16S rRNA sequencing of BAL from 29 subjects with CR asthma, 10 subjects with CS asthma, and 12 control subjects was performed. Figure 1 summarizes the strategic approach for analysis of the BAL microbiome, which involved (1) bacterial community metrics analysis (bacterial load, richness, evenness, and diversity); (2) overall comparison of major bacterial phylum distribution between subjects with CR asthma, subjects with CS asthma, and normal control subjects; (3) detailed analysis of 16S rRNA sequencing data within bacterial phyla (i.e., bacterial families and bacterial genera); and (4) evaluation for expansions in bacterial genera as compared with normal control subjects and analysis for unique bacterial expansions in CR and CS asthma. Step 4 was introduced into analysis, because significant variations in the percentage of sequences for different bacterial genera within CR and CS asthma groups were noted. 16S rRNA sequencing data obtained from BAL samples from normal control subjects were used to establish the ranges for normal distribution of the bacterial genera in the airways. Microbial genera were considered as expanded in asthmatic airway microbiome if they represented more than 5% of the total 16S rRNA sequences and the percentage of sequences was increased at least twofold over control subjects for the genera present both in the airways of subjects with asthma and normal control subjects; for the genera found only in subjects with asthma they were considered as expanded if they represented more than 5% of the total 16S rRNA sequences. Bacterial expansion was considered as unique if the bacterial expansion was only found in one group of patients with asthma (only subjects with CR asthma or only subjects with CS asthma), but not the other group of subjects with asthma.
Figure 1. A schematic diagram of the analysis approaches for the bronchoalveolar lavage (BAL) 16S rRNA sequencing data to characterize airway microbiome of corticosteroid-resistant (CR) and corticosteroid-sensitive (CS) subjects with asthma.
[More] [Minimize]No significant differences in airway bacterial load were observed between subjects with asthma and normal control subjects as assessed by 16S rRNA copy number (Figure 2A). The medium (range) number of sequences was 727 (335–941), 1,127 (167–2,821), and 1,475.5 (397–2,777) for BAL samples from normal control subjects and patients with CR and CS asthma, respectively, represented by 24 (18–40), 36 (11–47), and 28 (19–49) bacterial genera, respectively. No significant differences in the microbiome richness, evenness, and diversity between normal control subjects and subjects with CR and CS asthma were observed (Table 2).
Figure 2. Bacterial load and composition of the airway microbiome based on 16S rRNA sequencing of the bacterial DNA isolated from bronchoalveolar lavage samples from normal control subjects, and corticosteroid-resistant (CR) and corticosteroid-sensitive (CS) subjects with asthma. (A) Bacterial load based on 16S RNA copy number in bronchoalveolar lavage samples from normal control subjects, and patients with CR and CS asthma. (B) The taxonomic composition of the airway microbiome in normal control subjects, and subjects with CR and CS asthma. A mean % of sequences for the major bacterial taxonomic groups is presented. The following phyla are shown: Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria, and Fusobacteria. The remaining microorganisms that do not belong to these phyla are presented as “Other.” *P < 0.05 as compared to normal control.
[More] [Minimize]| Alpha diversity metric | Normal Control Subjects (n = 12) | CR Asthma (n = 29) | CS Asthma (n = 10) |
|---|---|---|---|
| Good’s coverage, % | 97.4 (95.1–98.6) | 95.8 (93.1–98.6) | 97.0 (93.0–98.4) |
| Richness (Sobs) | 14.7 (99.5–25.3) | 16.8 (12.3–28.4) | 16.8 (12.3–28.4) |
| Estimated richness (Chao1) | 19.2 (12.3–34.1) | 27.4 (11.0–42.0) | 21.8 (14.6–41.4) |
| Shannon diversity index | 2.8 (1.7–3.8) | 2.7 (1.7–3.9) | 2.9 (2.3–3.7) |
| Shannon evenness | 0.71 (0.44–0.83) | 0.68 (0.51–0.80) | 0.74 (0.52–0.80) |
We first compared main bacterial phylum composition of BAL samples of normal control subjects and patients with CR and CS asthma. No significant difference in bacterial phylum composition was noted between normal control subjects and subjects with CR and CS asthma (Figure 2B; see Table E3). Patients with CR asthma had significantly greater number of sequences of phylum Actinobacteria as compared with normal control subjects (P < 0.05). Differences in Proteobacteria phylum composition in subjects with CR asthma as compared with normal control subjects were observed (see Figure E1). The summaries of BAL samples bacterial composition at bacterial family and bacterial genus levels are shown in Tables E4 and E5, respectively. Again, no significant differences in BAL bacterial composition between subjects with CR and CS asthma were observed at bacterial family and bacterial genus level.
However, significant variations in the percentage of sequences for different bacterial genera within CR and CS asthma groups were noted, with some patients demonstrating either expansions or reductions in the number of sequences. This prompted us to further subgroup subjects with asthma based on bacterial expansions or lack of bacterial expansions. A total of 33 out of 39 subjects with asthma studied had expansions of specific groups of microorganisms. Subjects with CR asthma with microbial expansions (24 of 29 patients) had significantly increased proportion of sequences of microorganisms in phyla Actinobacteria (P < 0.001) and Proteobacteria (P < 0.05), and significantly reduced number of sequences for genera Prevotella (P < 0.05) and Veillonella (P < 0.05) and phylum Fusobacteria (P < 0.05) as compared with normal control subjects (Figure 3). Subjects with CS asthma with bacterial expansions (9 of 10 patients) had significant increased proportion of sequences for bacteria from phylum Proteobacteria (P < 0.05) and significantly reduced number of sequences for genera Prevotella (P < 0.05) and Veillonella (P < 0.001) (Figure 3).
Figure 3. The taxonomic composition of the airway microbiome in corticosteroid-resistant (CR) and corticosteroid-sensitive (CS) subjects with asthma with (bact. exp.) and without (no bact. exp.) bacterial expansions. Microbial genera were considered as expanded in asthmatic airway microbiome if they represented more than 5% of the total 16S rRNA sequences and the % of sequences was increased at least twofold over control subjects for the genera present both in the airways of subjects with asthma and normal control subjects; for the genera found only in subjects with asthma they were considered as expanded if they represented more than 5% of the total 16S rRNA sequences. Twenty-four subjects with CR asthma and nine subjects with CS asthma were found to have bacterial expansions as compared with normal control subjects; five CR subjects with asthma and one CS subject with asthma did not have bacterial expansions. A mean % sequences of major bacterial phyla per group is shown. *P < 0.05, ***P < 0.001 as compared to normal control.
[More] [Minimize]Subjects with asthma without bacterial expansion (n = 6; five patients with CR asthma and one patient with CS asthma) did not vary significantly from normal control subjects based on bacterial phylum distribution, but had a significantly reduced bacterial diversity and bacteria evenness compared with normal control subjects (P < 0.05 and P < 0.05) and subjects with asthma with bacterial expansions (P < 0.001 and P < 0.01) (data not shown). No difference in bacterial evenness was observed between normal control subjects and subjects with asthma with bacterial expansions (data not shown).
Fourteen of 29 subjects with CR asthma had unique bacterial expansions in the airways that were not expanded in subjects with CS asthma (genera Neisseria, Haemophilus, Simonsiella, and Campylobacter, all phylum Proteobacteria; genus Leptotrichia [phylum Fusobacteria]; genus Tropheryma [phylum Actinobacteria]; and genera Leuconostoc and Megasphaera [phylum Firmicutes]) (P < 0.01 as compared with patients with CS asthma) (Table 3). Four out of 10 patients with CS asthma had bacterial expansions not present in subjects with CR asthma (genera Bradyrhizobium, Aquabacterium, Limnobacter, and Pasteurella, all phylum Proteobacteria; genus Fusobacterium [phylum Fusobacteria]; and genus Streptophyta [phylum Cyanobacteria]) (Table 4) (P < 0.01 as compared with patients with CR asthma).
| Number of Patients with Bacterial Expansions (n) | Expanded Microorganisms Present in the Airways of Normal Control Subjects, Y/N (Mean % Sequences) | |||
|---|---|---|---|---|
| Types of organisms | CR Asthma†‡ | CS Asthma | ||
| Phylum | Genus | |||
| Actinobacteria | 1 | 0 | ||
| Tropheryma | 1 | 0 | N | |
| Firmicutes | 2 | 0 | ||
| Leuconostoc | 1 | 0 | N | |
| Megasphaera | 1 | 0 | Y (1.2%) | |
| Fusobacteria | 4 | 0 | ||
| Leptotrichia | 4 | 0 | Y (4.2%) | |
| Proteobacteria | 9 | 0 | ||
| β-Proteobacteria | Neisseria | 5 | 0 | Y (6.0%) |
| Simonsiella | 1 | 0 | Y (0.5%) | |
| γ-Proteobacteria | Haemophilus | 2 | 0 | Y (2.9%) |
| ε-Proteobacteria | Campylobacter | 1 | 0 | Y (3.2%) |
| Number of Patients with Bacterial Expansions (n) | Expanded Microorganisms Present in the Airways of Normal Control Subjects, Y/N (% Sequences) | |||
|---|---|---|---|---|
| Types of organisms | CR Asthma | CS Asthma†‡ | ||
| Phylum | Genus | |||
| Cyanobacteria | 0 | 1 | ||
| Streptophyta | 0 | 1 | Y (0.4%) | |
| Fusobacteria | 0 | 1 | ||
| Fusobacterium | 0 | 1 | Y (6.3%) | |
| Proteobacteria | 0 | 4 | ||
| α-Proteobacteria | Bradyrhizobium | 0 | 1 | N |
| β-Proteobacteria | Aquabacterium | 0 | 1 | N |
| Limnobacter | 0 | 1 | N | |
| γ-Proteobacteria | Pasteurella | 0 | 1 | Y (0.6%) |
The distinct bacteria expanded in the airways of these subjects with CR asthma were mainly gram-negative organisms (phyla Proteobacteria and Fusobacteria and genus Tropheryma [phylum Actinobacteria]) (Table 3). Significantly higher levels of LPS (P < 0.05) were detected in the BAL fluid of these patients (see Figure E2A). BAL cells from these participants expressed significantly higher IL-8 mRNA levels (P < 0.05 compared with subjects with CR asthma with no bacterial expansions or subjects with CR asthma with gram-negative bacteria expansions that were common between subjects with CR and CS asthma) (see Figure E2B).
Haemophilus parainfluenzae (phylum Proteobacteria), one of the organisms identified as uniquely expanded in the airways of subjects with CR asthma, was used to examine the effects of this microorganism on cellular responses to corticosteroids. The effects of H. parainfluenzae were compared with the airway commensal organism, P. melaninogenica. For these experiments, a separate group of subjects with asthma (see Table E1) was recruited to collect BAL and peripheral blood monocytes to examine functional interaction with bacteria. Because the numbers of BAL macrophages from subjects with asthma were limited, we initially assessed interactions of these microbes with peripheral blood monocytes (see online supplement), and then used established protocols to assess BAL macrophages interactions with these bacteria. As shown in Figure E3, H. parainfluenzae, but not P. melaninogenica, induced a dose-dependent and time-dependent activation of the p38 MAPK pathway, which persisted even after 3 hours since addition of H. parainfluenzae (see Figure E3). Cells cultured with H. parainfluenzae had significantly reduced responses to corticosteroids (see Figure E3).
P38 MAPK activation in asthmatic BAL macrophages was only observed in response to H. parainfluenzae, but not P. melaninogenica, after 3 hours of stimulation (Figure 4A). H. parainfluenzae, but not P. melaninogenica, significantly up-regulated IL-8 mRNA (Figure 4B) and mitogen-activated kinase phosphatase 1 (MKP-1) mRNA (Figure 4C) expression in BAL macrophages. In the presence of H. parainfluenzae MKP-1 induction by dexamethasone (DEX) was significantly reduced (Figure 4D) (P < 0.05). BAL cells remained steroid sensitive in the presence of P. melaninogenica (Figure 4D).
Figure 4. Effects of bacteria from the airways of subjects with asthma on bronchoalveolar lavage (BAL) macrophages activation and response to corticosteroids in vitro. Incubation of asthmatic BAL macrophages with Haemophilus parainfluenzae (H. para) results in p38 mitogen-activated protein kinase activation in the cells as detected by Western blot (A), up-regulation of IL-8 (B) and mitogen-activated kinase phosphatase 1 (MKP-1) mRNA production (C), and reduced responsiveness to corticosteroids in vitro (D) as shown by real-time polymerase chain reaction. Cells cultured with airway commensal organism Prevotella melaninogenica (P. melan) do not activate p38, do not up-regulate IL-8 mRNA and MKP-1 mRNA expression, and remain sensitive to corticosteroid treatment. (B–D) For IL-8 mRNA and MKP-1 mRNA production the cells were cultured overnight in X-Vivo 15 medium, incubated with bacteria for 15 minutes followed by 3 hours of treatment with 10−6M dexamethasone (DEX) or medium, and analyzed by real-time polymerase chain reaction. 0.25 × 106 cells was used per condition (bacterium to cell ratio 0.1:1 and 1:1). The amount of bacteria shown was calculated per 1 × 106 cells/ml per condition. The responses of BAL macrophages from five subjects with asthma were examined. *P < 0.05, **P < 0.01 as compared with medium-treated cells.
[More] [Minimize]Toll-like receptor (TLR) engagement by bacteria is a major pathway of cell activation by bacteria (21). Transforming growth factor-β–associated kinase-1 (TAK1) phosphorylation represents a key downstream TLR signaling branching point that control subjects’ MAPK and nuclear factor-kB pathway activation in the cells (22). We examined the role of MAPK and TAK1 activation in the alteration of cellular response to corticosteroids by H. parainfluenzae.
Peripheral blood monocytes from subjects with asthma cultured in the presence of H. parainfluenzae had significant up-regulation of MKP-1 mRNA (P < 0.05) (Figure 5A) and IL-8 mRNA (P < 0.05) (Figure 5B) expression. Cells cultured with H. parainfluenzae had reduced responses to corticosteroids as shown by significant inhibition of MKP-1 mRNA induction by DEX (Figure 5C). Pretreatment of monocytes with the TAK1 inhibitor resulted in significant inhibition of MKP-1 mRNA (Figure 5A) and IL-8 mRNA (Figure 5B) induction by H. parainfluenzae and restoration of cellular sensitivity to corticosteroids in vitro (Figure 5C). P38 MAPK inhibitor and the combination of p38 MAPK/ERK/JNK inhibitors also reduced MKP-1 mRNA and IL-8 mRNA induction by H. parainfluenzae, but not as effectively as in the presence of the TAK1 inhibitor (Figures 5A and 5B). P38 and the combination of MAPK inhibitors failed to restore cellular sensitivity to corticosteroids in the presence of H. parainfluenzae (Figure 5C). As shown in Western blots the doses of the inhibitors used were sufficient to inhibit their respective targets as 2 μM TAK1 inhibitor and 10 μM p38 MAPK inhibitor fully suppressed activation of ERK and hsp27 phosphorylation by H. parainfluenzae as downstream read out targets for TAK1 and p38 MAPK activation, respectively (Figure 5D).
Figure 5. Influence of Toll-like receptor pathway inhibitors on cellular response to corticosteroids in the presence of bacteria. Pretreatment of asthmatic peripheral blood monocytes with transforming growth factor-β–associated kinase-1 (TAK1) but not p38 mitogen-activated protein kinase (MAPK) or p38 MAPK/ERK/JNK inhibitors results in significant inhibition of mitogen-activated kinase phosphatase 1 (MKP-1) mRNA (A) and IL-8 mRNA (B) induction by Haemophilus parainfluenzae (H. para) and restoration of cellular sensitivity to corticosteroids in vitro (C) as shown by real-time polymerase chain reaction. (D) A total of 2 μM TAK1 and 10 μM p38 inhibitors fully inhibited activation of downstream signaling targets in response H. para. Phosphorylation of ERK and hsp27 as a downstream read out targets for TAK1 and p38 MAPK activation, respectively, in response to 15 minutes of treatment of monocytes from subjects with asthma with H. para (bacterium to cell ratio 1:1) with and without corresponding inhibitor is shown by Western blot. For IL-8 mRNA and MKP-1 mRNA production the cells were cultured overnight in X-Vivo 15 medium, incubated with inhibitors for 1 hour, stimulated with bacteria for 15 minutes followed by 3 hours of treatment with 10−6 M dexamethasone (DEX) or medium, and analyzed by real-time polymerase chain reaction. Bacteria were added to 0.5 × 106 cells per condition (bacterium to cell ratio 1:1). The responses of monocytes from four subjects with asthma were examined.
[More] [Minimize]
In the current study, we evaluated the lung microbiome in 39 subjects with asthma and 12 healthy control subjects. In our study, all subjects with asthma were defined clinically as CR or CS based on the change in lung function after 1 week of an oral prednisone burst. The two main objectives of the study were comparison of the airway microbiome composition in subjects with CR versus CS asthma, and examination of the effects of bacteria found in BAL from subjects with CR asthma on cellular response to corticosteroids.
Based on BAL microbiome data distribution in the major bacterial phyla, subjects with CR and CS asthma were divided into subgroups with bacterial expansions or lack of bacterial expansions. Most subjects with asthma in both CR and CS groups (24 out of 29 subjects with CR asthma and 9 out of 10 subjects with CS asthma) had bacterial expansions. Subjects with CR and CS asthma with bacterial expansions had significant alteration in their airway microbiome composition as compared with normal control subjects, but not compared with each other. The differences were only observed at the bacterial genus level, with distinct bacteria found expanded in 14 subjects with CR asthma, and these were not present in subjects with CS asthma (P < 0.01 as compared with subjects with CS asthma) (Table 3). Four subjects with CS asthma had unique bacterial genera expanded not found in patients with CR asthma (P < 0.01 as compared with subjects with CR asthma) (Table 4).
As compared with normal control subjects, subjects with CR and CS asthma with bacterial expansions had alterations in bacterial community structure as manifested by reduction of the airway commensals, genera Prevotella and Veillonella, reduction in phylum Fusobacteria and expansions of phyla Actinobacteria and Proteobacteria. The airway microbiome of subjects with asthma without bacterial expansion was also abnormal with less diversified microbial communities as compared with normal control subjects. We did not find any unique clinical features for this group that would distinguish them from other patients. Contrary to prior reports (11), we did not observe differences in bacterial load between samples from subjects with asthma and control subjects, suggesting that the microbial expansions in subjects with asthma reflected changes in the qualitative structure of the bacterial communities that reside in the airways of subjects with asthma. Also in contrast to previous reports (10, 11) no significant differences in the microbiome diversity between the subjects with asthma and control subjects were observed. Microbial expansions in subjects with asthma were distinct from previously reported alterations in airway microbiome in patients with chronic obstructive pulmonary disease (23–25).
At the genus level 14 out of 29 subjects with CR asthma had expansion of microbial genera that were expanded in the airways of none of the subjects with CS asthma. These microorganisms were mainly gram-negative LPS-producing bacteria with the short acyl chains lipid A LPS structures with known high endotoxic activity (Table 3) (26, 27). These patients had significantly higher IL-8 expression by BAL cells and significantly elevated levels of LPS in their BAL fluid, suggesting microbial stimulation. These data are in agreement with a previous report demonstrating induction of LPS signaling pathways in BAL cells from subjects with CR asthma (20).
Four out of 10 subjects with CS asthma had distinct microbial genera expanded not present in subjects with CR asthma. Among five gram-negative organisms identified as uniquely expanded in subjects with CS asthma, Bradyrhizobium (28, 29), Aquabacterium (30), Limnobacter (31), and Fusobacterium (32) have long acyl chain lipid A and low endotoxic activity. We were unable to locate information about the structure of the LPS produced by Pasteurella.
In this study we used limulus amebocyte lysate assay to examine LPS levels in BAL fluid. However, limulus amebocyte lysate assay cannot distinguish the differences in LPS structures among various gram-negative microorganisms. Additional methods, like mass spectrometry (33, 34), are more suited for these purposes to further characterize the differences in LPS structure. Our laboratory previously used mass spectrometry for LPS characterization in BAL fluid and reported higher levels of LPS in CR than CS subjects with asthma (20).
To date, functional interactions of the microbiome with BAL macrophages, and its effects on cellular response to corticosteroids have not been investigated. We explored whether bacteria in the airway of subjects with CR asthma direct alterations in cellular responsiveness to corticosteroids. The effects of bacteria found to be uniquely expanded in CS asthma airways were not studied further. We chose one of the microorganisms from the Proteobacteria phylum that was uniquely expanded in the airways of subjects with CR asthma (H. parainfluenzae) and a commensal microorganism (P. melaninogenica) to examine functional responses to these bacteria in the presence of DEX. We introduced MKP-1 and IL-8 as gene targets to assess cellular responsiveness to corticosteroids in the presence of bacteria in vitro. Alterations in DEX suppression of these targets in the presence of bacteria were used as a read-out of cellular sensitivity to corticosteroids. High levels of p38 MAPK activation and reduced cellular responses to corticosteroids were observed in both peripheral blood monocytes and BAL macrophages in the presence of H. parainfluenzae, but not P. melaninogenica. The absolute increase in MKP-1 mRNA was much greater in cells incubated with H. parainfluenzae than with media or P. melaninogenica. An increase in MKP-1 mRNA expression is one of the cellular regulatory mechanisms that inhibits p38 MAPK activation, and reduces the inflammatory response (35, 36). However, this MKP-1 induction seemed to be insufficient. As shown in Figure E3, p38 MAPK activation persisted after 3 hours of treatment with H. parainfluenzae, and at the same time the cells expressed high IL-8 mRNA levels. Based on the data on the cellular responses to H. parainfluenzae versus P. melaninogenica, we suggest that the regulation of these two targets by DEX should be assessed concurrently, as likely both absolute levels of induction of these targets by bacteria and a proportional increase (for MKP-1) or decrease (for IL-8) by DEX are critical to assess the efficacy of corticosteroid responses in each setting.
The experiments suggest intrinsic differences in cell-stimulating properties between commensal versus expanded pathogenic bacteria associated with CR asthma airway microbiota. We hypothesize that the difference in LPS composition between LPS-producing commensal and pathogenic microorganisms in the airways (27, 37–39) accounts for airway cell activation only by pathogenic bacteria and results in altered cellular response to corticosteroids. Of note, microorganisms found to be uniquely expanded in CS asthmatic airways were also long acyl chain lipid A producers. We have studied the effects of commercial long acyl chain bacterial LPS on monocyte responses to corticosteroids and found that they are not immunostimulatory and do not subvert cellular responses to corticosteroids (data not shown). Figure 6 presents the proposed schematic diagram of monocyte/macrophage activation and cellular response to corticosteroids in the presence of H. parainfluenzae versus P. melaninogenica.
Figure 6. Proposed schematic diagram of monocyte-macrophage activation and cellular response to corticosteroids in the presence of Haemophilus parainfluenzae and Prevotella melaninogenica. H. parainfluenzae short length acyl chains lipid A LPS interacts with Toll-like receptor (TLR) 4 and activates transforming growth factor-β–associated kinase-1 (TAK1) by MyD88 pathway, resulting in p38 mitogen-activated protein kinase phosphorylation and nuclear factor-κB (NF-κB) activation, which activate transcription of the proinflammatory cytokines like IL-8 (21, 26, 27, 53). On the contrary, P. melaninogenica long acyl chains lipid A LPS is a poor agonist for TLR4 (54). On interaction cytoplasmic glucocorticoid receptor (GR) with dexamethasone receptor translocates to the cell nuclei and activates mitogen-activated kinase phosphatase 1 (MKP-1) mRNA production. MKP-1 dephosphorylates activated p38 mitogen-activated protein kinase. GR interacts with NF-κB and inhibits IL-8 transcription. Monocyte-macrophage activation by H. parainfluenzae results in reduced cellular responses to corticosteroids. TAK1 activation by H. parainfluenzae inhibits GR-mediated MKP-1 production and suppresses GR inhibition of NF-κB–induced IL-8 production. The cells remain corticosteroid sensitive in the presence of P. melaninogenica.
[More] [Minimize]Bacterial alteration in airway macrophage responses to corticosteroids is of great importance, because macrophages are the major cell type in the airway lumen, and produce an array of proinflammatory mediators that control responses of various cell types, including T cells and airway epithelium that are involved in host defense, airway inflammation, tissue remodeling, and repair (40–42). Lack of responses of pulmonary macrophages to corticosteroids leads to persistent airway inflammation in CR asthma (42). Corticosteroids are the most effective antiinflammatory drugs used in the treatment of eosinophilic disorders, including asthma (43). Failure of steroids to suppress macrophage responses in the presence of bacteria may result in failure to propagate apoptotic signals for eosinophils and may result in the deficiency of eosinophil apoptotic bodies clearance by macrophages (7, 43).
It is unclear whether expansions in Actinobacteria and Proteobacteria in BAL from subjects with CR and CS asthma resulted in reciprocal reductions in commensal organisms like Prevotella and Veillonella. It is possible that initial reduction in airway commensals resulted in overexpansion of other bacterial phyla. Recent literature suggests that commensal microbiota maintains and shapes normal mucosal immunity in the gut (44, 45). Similarly, it has been reported that Staphylococcus epidermidis, a commensal organism in the skin, can protect the host from development of injurious inflammation by tolerizing the response by TLR (46). By analogy, it is possible that commensal microbiota in the airways is protective from development of inflammatory responses, and loss of commensal organisms allows cellular inflammatory response. Restoration of the commensal microbiota in the airways of subjects with asthma should be evaluated for its protective role and alleviation of cellular steroid response in the airways of subjects with asthma.
Persistent infection and colonization with pathogenic bacteria may be difficult to eradicate with antibiotic treatment (47), therefore alternative approaches are needed. In this study, we evaluated whether inhibition of TAK1 and MAPK activation by bacteria via TLR by selective inhibitors would restore cellular responses to corticosteroids in the presence of pathogenic bacteria. We found that TAK1 inhibitor–treated cells were steroid sensitive despite incubation with H. parainfluenzae. MAPK inhibitors reduced IL-8 and MKP-1 mRNA induction by H. parainfluenzae but did not restore cellular sensitivity to corticosteroids in the presence of H. parainfluenzae. The data suggest that activation of TAK1 by bacteria is essential in alteration of cellular responses to corticosteroids.
Heterogeneity in asthma is well described, and prior studies have reported a wide variety of other factors that can contribute to corticosteroid insensitivity, such as obesity (48), allergen exposure (6, 15), vitamin D deficiency (49, 50), sex (51), and race (52). The novelty of our study is the identification of a distinct microbial trigger for CR asthma caused by unique mainly gram-negative expansions in the airways, and demonstration that these microorganisms can alter responses of airway macrophages to corticosteroids ex vivo.
In summary, we propose that airway cell stimulation by bacteria as a result of alterations in the airway microbiome composition and microbial expansions reduce cellular responses to corticosteroids and influences efficacy of corticosteroid treatment. TAK1 inhibition should be explored as a potential therapeutic pathway for modulation of corticosteroid responses in subjects with asthma with poor responses to corticosteroids as an alternative approach to their management.
| 1. | National Asthma Education and Prevention Program (National Heart Lung and Blood Institute) Third Expert Panel on the Management of Asthma. National Center for Biotechnology Information (US). Expert panel report 3 guidelines for the diagnosis and management of asthma. NIH publication no 07-4051. Bethesda, MD: National Institutes of Health National Heart Lung and Blood Institute; 2007. |
| 2. | Global strategy for asthma management and prevention. Global Initiative for Asthma (GINA) 2012 [accessed 2013 Apr 1]. Available from: http://www.ginasthma.org/ |
| 3. | Malmstrom K, Rodriguez-Gomez G, Guerra J, Villaran C, Piñeiro A, Wei LX, Seidenberg BC, Reiss TF; Montelukast/Beclomethasone Study Group. Oral montelukast, inhaled beclomethasone, and placebo for chronic asthma: a randomized, controlled trial. Ann Intern Med 1999;130:487–495. |
| 4. | Martin RJ, Szefler SJ, King TS, Kraft M, Boushey HA, Chinchilli VM, Craig TJ, Dimango EA, Deykin A, Fahy JV, et al.; National Heart, Lung, and Blood Institute’s Asthma Clinical Research Center. The predicting response to inhaled corticosteroid efficacy (PRICE) trial. J Allergy Clin Immunol 2007;119:73–80. |
| 5. | Barnes PJ, Adcock IM. Glucocorticoid resistance in inflammatory diseases. Lancet 2009;373:1905–1917. |
| 6. | Leung DY, Bloom JW. Update on glucocorticoid action and resistance. J Allergy Clin Immunol 2003;111:3–22, quiz 23. |
| 7. | Leung DY, Martin RJ, Szefler SJ, Sher ER, Ying S, Kay AB, Hamid Q. Dysregulation of interleukin 4, interleukin 5, and interferon gamma gene expression in steroid-resistant asthma. J Exp Med 1995;181:33–40. |
| 8. | Goleva E, Hauk PJ, Boguniewicz J, Martin RJ, Leung DY. Airway remodeling and lack of bronchodilator response in steroid-resistant asthma. J Allergy Clin Immunol 2007;120:1065–1072. |
| 9. | Drazen JM. Asthma: the paradox of heterogeneity. J Allergy Clin Immunol 2012;129:1200–1201. |
| 10. | Hilty M, Burke C, Pedro H, Cardenas P, Bush A, Bossley C, Davies J, Ervine A, Poulter L, Pachter L, et al. Disordered microbial communities in asthmatic airways. PLoS ONE 2010;5:e8578. |
| 11. | Huang YJ, Nelson CE, Brodie EL, Desantis TZ, Baek MS, Liu J, Woyke T, Allgaier M, Bristow J, Wiener-Kronish JP, et al. Airway microbiota and bronchial hyperresponsiveness in patients with suboptimally controlled asthma. J Allergy Clin Immunol 2011;127:372–381. |
| 12. | Goleva E, Harris JK, Martin RJ, Dakhama A, Alam R, Gelfand EW, Leung DY. Asthma control and disordered microbial communities in the lower airways of patients with poorly controlled asthma [abstract]. J Allergy Clin Immunol 2012;129:AB128. |
| 13. | Goleva E, Jackson LP, Harris JK, Martin RJ, Leung DYM. Microbiome of the lower airways alters corticosteroid responsiveness in asthma [abstract]. J Allergy Clin Immunol 2013;131:AB138. |
| 14. | Goleva E, Li LB, Eves PT, Strand MJ, Martin RJ, Leung DY. Increased glucocorticoid receptor beta alters steroid response in glucocorticoid-insensitive asthma. Am J Respir Crit Care Med 2006;173:607–616. |
| 15. | Sher ER, Leung DY, Surs W, Kam JC, Zieg G, Kamada AK, Szefler SJ. Steroid-resistant asthma. Cellular mechanisms contributing to inadequate response to glucocorticoid therapy. J Clin Invest 1994;93:33–39. |
| 16. | Hamid QA, Wenzel SE, Hauk PJ, Tsicopoulos A, Wallaert B, Lafitte JJ, Chrousos GP, Szefler SJ, Leung DY. Increased glucocorticoid receptor beta in airway cells of glucocorticoid-insensitive asthma. Am J Respir Crit Care Med 1999;159:1600–1604. |
| 17. | Summary and recommendations of a workshop on the investigative use of fiberoptic bronchoscopy and bronchoalveolar lavage in asthmatics. Am Rev Respir Dis 1985;132:180–182. |
| 18. | American Thoracic Society. Medical Section of the American Lung Association. Guidelines for fiberoptic bronchoscopy in adults. Am Rev Respir Dis 1987;136:1066. |
| 19. | Busse WW, Wanner A, Adams K, Reynolds HY, Castro M, Chowdhury B, Kraft M, Levine RJ, Peters SP, Sullivan EJ. Investigative bronchoprovocation and bronchoscopy in airway diseases. Am J Respir Crit Care Med 2005;172:807–816. |
| 20. | Goleva E, Hauk PJ, Hall CF, Liu AH, Riches DW, Martin RJ, Leung DY. Corticosteroid-resistant asthma is associated with classical antimicrobial activation of airway macrophages. J Allergy Clin Immunol 2008;122:550–559, e3. |
| 21. | Akira S, Takeda K. Toll-like receptor signalling. Nat Rev Immunol 2004;4:499–511. |
| 22. | Liew FY, Xu D, Brint EK, O’Neill LA. Negative regulation of toll-like receptor-mediated immune responses. Nat Rev Immunol 2005;5:446–458. |
| 23. | Pragman AA, Kim HB, Reilly CS, Wendt C, Isaacson RE. The lung microbiome in moderate and severe chronic obstructive pulmonary disease. PLoS ONE 2012;7:e47305. |
| 24. | Sze MA, Dimitriu PA, Hayashi S, Elliott WM, McDonough JE, Gosselink JV, Cooper J, Sin DD, Mohn WW, Hogg JC. The lung tissue microbiome in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2012;185:1073–1080. |
| 25. | Erb-Downward JR, Thompson DL, Han MK, Freeman CM, McCloskey L, Schmidt LA, Young VB, Toews GB, Curtis JL, Sundaram B, et al. Analysis of the lung microbiome in the “healthy” smoker and in COPD. PLoS ONE 2011;6:e16384. |
| 26. | Lodowska J, Wolny D, Weglarz L, Dzierzewicz Z. [The structural diversity of lipid A from gram-negative bacteria]. Postepy Hig Med Dosw (Online) 2007;61:106–121. |
| 27. | Miller SI, Ernst RK, Bader MW. LPS, TLR4 and infectious disease diversity. Nat Rev Microbiol 2005;3:36–46. |
| 28. | Komaniecka I, Choma A, Lindner B, Holst O. The structure of a novel neutral lipid A from the lipopolysaccharide of Bradyrhizobium elkanii containing three mannose units in the backbone. Chemistry 2010;16:2922–2929. |
| 29. | Komaniecka I, Zdzisinska B, Kandefer-Szerszen M, Choma A. Low endotoxic activity of lipopolysaccharides isolated from Bradyrhizobium, Mesorhizobium, and Azospirillum strains. Microbiol Immunol 2010;54:717–725. |
| 30. | Chen WM, Cho NT, Yang SH, Arun AB, Young CC, Sheu SY. Aquabacterium limnoticum sp. nov., isolated from a freshwater spring. Int J Syst Evol Microbiol 2012;62:698–704. |
| 31. | Lu H, Sato Y, Fujimura R, Nishizawa T, Kamijo T, Ohta H. Limnobacter litoralis sp. nov., a thiosulfate-oxidizing, heterotrophic bacterium isolated from a volcanic deposit, and emended description of the genus Limnobacter. Int J Syst Evol Microbiol 2011;61:404–407. |
| 32. | Asai Y, Makimura Y, Kawabata A, Ogawa T. Soluble CD14 discriminates slight structural differences between lipid as that lead to distinct host cell activation. J Immunol 2007;179:7674–7683. |
| 33. | Park JH, Szponar B, Larsson L, Gold DR, Milton DK. Characterization of lipopolysaccharides present in settled house dust. Appl Environ Microbiol 2004;70:262–267. |
| 34. | Reynolds SJ, Milton DK, Heederik D, Thorne PS, Donham KJ, Croteau EA, Kelly KM, Douwes J, Lewis D, Whitmer M, et al. Interlaboratory evaluation of endotoxin analyses in agricultural dusts—comparison of LAL assay and mass spectrometry. J Environ Monit 2005;7:1371–1377. |
| 35. | Chen P, Li J, Barnes J, Kokkonen GC, Lee JC, Liu Y. Restraint of proinflammatory cytokine biosynthesis by mitogen-activated protein kinase phosphatase-1 in lipopolysaccharide-stimulated macrophages. J Immunol 2002;169:6408–6416. |
| 36. | Liu Y, Shepherd EG, Nelin LD. MAPK phosphatases—regulating the immune response. Nat Rev Immunol 2007;7:202–212. |
| 37. | Netea MG, van Deuren M, Kullberg BJ, Cavaillon JM, Van der Meer JW. Does the shape of lipid A determine the interaction of LPS with Toll-like receptors? Trends Immunol 2002;23:135–139. |
| 38. | Erridge C, Bennett-Guerrero E, Poxton IR. Structure and function of lipopolysaccharides. Microbes Infect 2002;4:837–851. |
| 39. | Larsen JM, Steen-Jensen DB, Laursen JM, Søndergaard JN, Musavian HS, Butt TM, Brix S. Divergent pro-inflammatory profile of human dendritic cells in response to commensal and pathogenic bacteria associated with the airway microbiota. PLoS ONE 2012;7:e31976. |
| 40. | Murray PJ, Wynn TA. Protective and pathogenic functions of macrophage subsets. Nat Rev Immunol 2011;11:723–737. |
| 41. | Minnicozzi M, Sawyer RT, Fenton MJ. Innate immunity in allergic disease. Immunol Rev 2011;242:106–127. |
| 42. | Yang M, Kumar RK, Hansbro PM, Foster PS. Emerging roles of pulmonary macrophages in driving the development of severe asthma. J Leukoc Biol 2012;91:557–569. |
| 43. | Walsh GM, Sexton DW, Blaylock MG. Corticosteroids, eosinophils and bronchial epithelial cells: new insights into the resolution of inflammation in asthma. J Endocrinol 2003;178:37–43. |
| 44. | Ivanov II, Littman DR. Modulation of immune homeostasis by commensal bacteria. Curr Opin Microbiol 2011;14:106–114. |
| 45. | Artis D. Epithelial-cell recognition of commensal bacteria and maintenance of immune homeostasis in the gut. Nat Rev Immunol 2008;8:411–420. |
| 46. | Lai Y, Di Nardo A, Nakatsuji T, Leichtle A, Yang Y, Cogen AL, Wu ZR, Hooper LV, Schmidt RR, von Aulock S, et al. Commensal bacteria regulate Toll-like receptor 3-dependent inflammation after skin injury. Nat Med 2009;15:1377–1382. |
| 47. | Kong HH, Oh J, Deming C, Conlan S, Grice EA, Beatson MA, Nomicos E, Polley EC, Komarow HD, Murray PR, et al.; NISC Comparative Sequence Program. Temporal shifts in the skin microbiome associated with disease flares and treatment in children with atopic dermatitis. Genome Res 2012;22:850–859. |
| 48. | Sutherland ER, Goleva E, Strand M, Beuther DA, Leung DY. Body mass and glucocorticoid response in asthma. Am J Respir Crit Care Med 2008;178:682–687. |
| 49. | Sutherland ER, Goleva E, Jackson LP, Stevens AD, Leung DY. Vitamin D levels, lung function, and steroid response in adult asthma. Am J Respir Crit Care Med 2010;181:699–704. |
| 50. | Zhang Y, Leung DYM, Richers BN, Liu Y, Remigio LK, Riches DW, Goleva E. Vitamin D inhibits monocyte/macrophage pro-inflammatory cytokine production by targeting mitogen-activated protein kinase phosphatase 1. J Immunol 2012;188:2127–2135. |
| 51. | Zhang Y, Leung DY, Nordeen SK, Goleva E. Estrogen inhibits glucocorticoid action via protein phosphatase 5 (PP5)-mediated glucocorticoid receptor dephosphorylation. J Biol Chem 2009;284:24542–24552. |
| 52. | Federico MJ, Covar RA, Brown EE, Leung DY, Spahn JD. Racial differences in T-lymphocyte response to glucocorticoids. Chest 2005;127:571–578. |
| 53. | Wieland CW, Florquin S, Maris NA, Hoebe K, Beutler B, Takeda K, Akira S, van der Poll T. The MyD88-dependent, but not the MyD88-independent, pathway of TLR4 signaling is important in clearing nontypeable Haemophilus influenzae from the mouse lung. J Immunol 2005;175:6042–6049. |
| 54. | Hashimoto M, Asai Y, Tamai R, Jinno T, Umatani K, Ogawa T. Chemical structure and immunobiological activity of lipid A from Prevotella intermedia ATCC 25611 lipopolysaccharide. FEBS Lett 2003;543:98–102. |
Supported by NIAID and NHLBI of the National Institutes of Health under Award Numbers AI070140, 2R56AI070140, and HL37260. The authors acknowledge The Edelstein Family Foundation for their generous support of this work.
Author Contributions: E.G. aided in designing the research project; acquisition, analysis, and interpretation of data; and drafting and revising the article. L.P.J. performed in vitro experiments with bacteria, corticosteroids, and inhibitors. J.K.H. performed bacterial DNA sequencing analysis from bronchoalveolar lavage samples collected and assisted with sequencing data analysis. C.E.R. developed Explicet software used for microbiome data analysis in this manuscript and assisted with the sequencing data informatics. C.F.H. assisted with bacterial cultures. E.R.S., J.T.G., and R.J.M. performed bronchoscopies and assisted with clinical evaluation of subjects with asthma. E.W.G. assisted with patient characterization and revising of this manuscript. D.Y.M.L. aided in designing the research, interpretation of data, and drafting and revising the article.
Originally Published in Press as DOI: 10.1164/rccm.201304-0775OC on September 11, 2013
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Author disclosures are available with the text of this article at www.atsjournals.org.
