Accepted Manuscript Inflammatory phenotypes in severe asthma are associated with distinct airway microbiology Steven L. Taylor, BSc, Lex E.X. Leong, PhD, Jocelyn M. Choo, PhD, Steve Wesselingh, FRACP PhD, Ian A. Yang, FRACP PhD, John W. Upham, FRACP PhD, Paul N. Reynolds, MBBS PhD, Sandra Hodge, PhD, Alan L. James, FRACP PhD, Christine Jenkins, MBBS FRACP, Matthew J. Peters, MD FRACP, Melissa Baraket, PhD, Guy B. Marks, MBBS PhD, Peter G. Gibson, MBBS FRACP, Jodie L. Simpson, PhD, Geraint B. Rogers, PhD PII:
S0091-6749(17)30743-1
DOI:
10.1016/j.jaci.2017.03.044
Reference:
YMAI 12790
To appear in:
Journal of Allergy and Clinical Immunology
Received Date: 16 October 2016 Revised Date:
28 February 2017
Accepted Date: 15 March 2017
Please cite this article as: Taylor SL, Leong LEX, Choo JM, Wesselingh S, Yang IA, Upham JW, Reynolds PN, Hodge S, James AL, Jenkins C, Peters MJ, Baraket M, Marks GB, Gibson PG, Simpson JL, Rogers GB, Inflammatory phenotypes in severe asthma are associated with distinct airway microbiology, Journal of Allergy and Clinical Immunology (2017), doi: 10.1016/j.jaci.2017.03.044. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT 1
Inflammatory phenotypes in severe asthma are associated with distinct airway
2
microbiology
3 4
Steven L. Taylor BSc1,2*, Lex E.X. Leong PhD1,2*, Jocelyn M. Choo PhD1,2, Steve
5
Wesselingh FRACP PhD1,2, Ian A. Yang FRACP PhD3,4, John W. Upham FRACP PhD
6
Paul N. Reynolds MBBS PhD6,7, Sandra Hodge PhD6,7, Alan L. James FRACP PhD8,9,
7
Christine Jenkins MBBS FRACP10,11, Matthew J. Peters MD FRACP11,12, Melissa Baraket
8
PhD13,16, Guy B. Marks MBBS PhD13,14,16, Peter G. Gibson MBBS FRACP14,15, Jodie L.
9
Simpson PhD15†, Geraint B. Rogers PhD1,2†
M AN U
10
,
SC
RI PT
3,5
11
1
12
Australia
13
2
14
Adelaide, South Australia, Australia
15
3
School of Medicine, The University of Queensland, St Lucia, QLD, Australia.
16
4
Department of Thoracic Medicine, The Prince Charles Hospital, Chermside, QLD, Australia.
17
5
Translational Research Institute, Princess Alexandra Hospital, Woolloongabba, QLD,
18
Australia.
19
6
20
Hanson Institute, Adelaide, SA, Australia.
21
7
School of Medicine, The University of Adelaide, Adelaide, SA, Australia.
22
8
Department of Pulmonary Physiology and Sleep Medicine, Sir Charles Gairdner Hospital,
23
Nedlands, WA, Australia.
24
9
25
Australia.
South Australian Health and Medical Research Institute, Adelaide, South Australia,
EP
TE D
SAHMRI Microbiome Research Laboratory, School of Medicine, Flinders University,
AC C
Department of Thoracic Medicine, Royal Adelaide Hospital and Lung Research Laboratory,
School of Medicine and Pharmacology, University of Western Australia, Crawley, WA,
1
ACCEPTED MANUSCRIPT 10
Respiratory Trials, The George Institute for Global Health, NSW, Australia,
27
11
Australian School of Advanced Medicine, Macquarie University, NSW, Australia
28
12
Department of Thoracic Medicine, Concord General Hospital, NSW, Australia,
29
13
Respiratory Medicine Department and Ingham Institute, Liverpool Hospital, NSW,
30
Australia
31
14
Woolcock Institute of Medical Research, Glebe, NSW, Australia.
32
15
Respiratory and Sleep Medicine, Priority Research Centre for Healthy Lungs, The
33
University of Newcastle, Callaghan, NSW, Australia.
34
16
35
Australia
36
* Both authors contributed equally to this work
37
†
SC
RI PT
26
M AN U
South Western Sydney Clinical School, University of New South Wales, Sydney. NSW
Both authors contributed equally to this work
38
Corresponding author: Associate Professor Geraint Rogers, School of Medicine, Flinders
40
University, University Drive, Bedford Park, Adelaide SA 5042, Australia.
41
Email:
[email protected], Tel: +61 (0)8 8204 7614
44
EP
43
Keywords: Asthma, microbiome, neutrophil, eosinophil
AC C
42
TE D
39
45
Declaration of interests
46
No author has a conflict of interest to declare in relation to manuscript.
2
ACCEPTED MANUSCRIPT Abstract
2
Background
3
Asthma pathophysiology and treatment responsiveness are predicted by inflammatory
4
phenotype. However, the relationship between airway microbiology and asthma phenotype is
5
poorly understood. We aimed to characterise airway microbiota in patients with symptomatic
6
stable asthma, and relate composition to airway inflammatory phenotype and other
7
phenotypic characteristics.
8
Methods
9
The microbial composition of induced sputum specimens collected from adult patients
10
screened for a multicenter randomized controlled trial was determined by 16S rRNA gene
11
sequencing. Inflammatory phenotypes were defined by sputum neutrophil and eosinophil cell
12
proportions. Microbiota were defined using alpha and beta diversity measures, and inter-
13
phenotype differences identified using SIMPER, network analysis, and taxon fold change.
14
Phenotypic predictors of airway microbiology were identified using multivariate linear
15
regression.
16
Results
17
Microbiota composition was determined in 167 participants, classified as eosinophilic (n=84),
18
neutrophilic (n=14), paucigranulocytic (n=60), or mixed neutrophilic-eosinophilic (n=9)
19
phenotypes of asthma. Airway microbiology was significantly less diverse (p=0.022) and
20
more dissimilar (p=0.005) in neutrophilic compared to eosinophilic participants. Sputum
21
neutrophil proportion, but not eosinophil proportion, correlated significantly with these
22
diversity measures (alpha-diversity: Spearman’s r=-0.374, p3% eosinophils, yellow= 3% eosinophils (mixed). Statistical significance was assessed by A) Kruskal-Wallis one-
M AN U
way ANOVA with Dunn’s post hoc test or B) and C) Spearman’s rank correlation.
Figure 2: Microbiota dispersion grouped by asthma phenotype. A) Principal Coordinate Analysis (PCoA). The first two principal coordinates are plotted on the x- and y-axes, respectively (representing 59.9% of the total variation). B) Distance from centroid. Statistical
TE D
significance was assessed by Kruskal-Wallis one-way ANOVA with Dunn’s post hoc test.
Figure 3: Relative abundance of discriminant taxa among asthma phenotypes. The thirteen
EP
taxa that collectively contribute to approximately 50% of variance among phenotypes, as
AC C
determined by SIMPER analysis. The clustering shows the similarity relationship of genera based on Bray-Curtis similarity distance and single linkage hierarchical clustering method. * Actinomyces sp. uncultured bacteria, # Actinomyces sp. oral clone DR002.
Figure 4: Bacterial network analysis of asthma cohort. Each edge represents a significant correlation coloured by either positive (blue) or negative (red). Edge width and transparency are proportional to the absolute value of the correlation coefficient. Node size is proportional to mean relative abundance. Node hue is proportional to the difference in taxon relative
26
ACCEPTED MANUSCRIPT abundance between the neutrophilic phenotype group and the eosinophilic phenotype group. Correlations performed by SparCC with a correlation cut-off of R>0.25 or 61% neutrophils, green= >3% eosinophils, yellow= 3% eosinophils (mixed).
M AN U
Dotted line at 61% neutrophils indicates phenotype cut-off point. Statistical significance was assessed by A) Kruskal-Wallis one-way ANOVA with Dunn’s post hoc test and B)
AC C
EP
TE D
Spearman’s rank correlation.
27
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Inflammatory phenotypes in severe asthma are associated with distinct airway microbiology Steven L. Taylor*, Lex E.X. Leong*, Jocelyn M. Choo, Steve Wesselingh, Ian A. Yang, John W. Upham, Paul N. Reynolds, Sandra Hodge, Alan L. James, Christine Jenkins, Matthew J. Peters, Melissa Baraket, Guy B. Marks, Peter G. Gibson, Jodie L. Simpson†, Geraint B. Rogers†
RI PT
Supplementary Methods
M AN U
SC
Exclusion criteria Asthma diagnosis was established using American Thoracic Society guidelines based on current episodic respiratory symptoms, clinical diagnosis and evidence of variable airflow obstruction.1 Participants with asthma were included if stable but symptomatic, despite being prescribed maintenance inhaled corticosteroid (ICS) and long acting bronchodilator treatment with an Asthma Control Questionnaire 6 (ACQ6) score >0.75.2 Participants with an FEV1 10 pack year history and DLCO/VA 10 pack years and exhaled carbon monoxide >10ppm) were also excluded. This study was conducted in accordance with the amended Declaration of Helsinki. Local institutional review boards approved the protocol and written informed consent was obtained from all participants.
EP
TE D
Institutional centres Sputum samples were collected from eight Australian centres: Hunter Medical Research Institute, Newcastle NSW Australia The Prince Charles Hospital, Chermside, QLD, Australia Princess Alexandra Hospital, Woolloongabba QLD, Australia Royal Adelaide Hospital, Adelaide SA, Australia Sir Charles Gairdner Hospital, Nedlands WA, Australia Woolcock Institute of Medical Research, Glebe NSW, Australia Concord Repatriation General Hospital, Concord NSW, Australia Liverpool Hospital, Liverpool NSW, Australia
AC C
Sample collection All study participants attended a single visit that included the assessment of lung function asthma symptoms, asthma specific quality of life,3 medication use and smoking status. Sputum induction with hypertonic saline (4.5%) was performed as described previously.4 Sputum aliquots were stored at -80°C for DNA extraction, or dispersed using dithiothreitol for sputum cell count assessment and inflammatory subtype determination.5 Patient inflammatory phenotyping Patient sputum was dispersed using dithiothreitol and inflammatory cells were counted as a percentage of total sputum cells. Inflammatory subtype was determined as described below. Neutrophilic cut-off values were age-dependent as described previously.6, 7 Neutrophilic phenotype Neutrophil % (< 20 years old) Neutrophil % (20-40 years old)
≥75.57% ≥61.61% 1
ACCEPTED MANUSCRIPT ≥63.25% ≥67.25%
≤3% ≤75.57% ≤61.61% ≤63.25% ≤67.25% ≥3% ≥75.57% ≥61.61% ≥63.25% ≥67.25%
RI PT
≥3%
SC
Neutrophil % (40-60 years old) Neutrophil % (>60 years old) Eosinophilic phenotype Eosinophil % Paucigranulocytic phenotype Eosinophil % Neutrophil % (< 20 years old) Neutrophil % (20-40 years old) Neutrophil % (40-60 years old) Neutrophil % (>60 years old) Mixed granulocytic phenotype Eosinophil % Neutrophil % (< 20 years old) Neutrophil % (20-40 years old) Neutrophil % (40-60 years old) Neutrophil % (>60 years old)
AC C
EP
TE D
M AN U
DNA extraction DNA extraction was performed on sputum sample aliquots of approximately 100 µl. Following the addition of 300 µl of phosphate buffered saline, samples were vortexed for 10 seconds and placed on ice for 2 min. Bacterial cells were then pelleted by centrifugation at 13,000 x g for 10 min. Following removal of supernatant, 300 µl of Tris-EDTA solution (10 mM Tris-HCl, 1 mM EDTA; pH 8.0; Ambion, ThermoFisher Scientific, Victoria, Australia), 200 µg of silica: zirconium beads (1:1 of 0.1 µm and 1.0 µm; Biospec Products, Inc., OK, USA), and a single chrome bead (3.2 mm, Biospec Products, Inc., OK, USA) were added to the tube containing the cell pellet. Samples underwent bead-beating at 6.5 m/s for 60 sec in a FastPrep®-24 Instrument (MP Biomedicals, CA, USA). Homogenised sample was heated to 90 °C for 5 min, before being cooled on ice for 5 min. Lysozyme (ROCHE, ThermoFisher Scientific, Victoria, Australia) and lysostaphin (Sigma-Aldrich, MO, USA) were then added to a final concentration of 2 mg/mL and 0.1 mg/mL, respectively, and samples incubated at 37 °C for 1 hr. Proteinase K (Fermentas, ThermoFisher Scientific, Victoria, Australia) and sodium dodecyl sulphate (Sigma-Aldrich, MO, USA) were then added to a final concentration of 1.2 mg/mL and 1.5 %, w/v, respectively. Following incubation at 30 min at 56 °C, 40 µl of 5M sodium chloride and 450 µl of phenol:chloroform:isoamyl alcohol (25:24:1; saline buffered at pH8.0; Sigma-Aldrich, MO, USA) were added and samples vortexed for 30 sec. The aqueous-organic layers were separated by centrifugation at 13,000 x g for 10 min and 400 µl of the aqueous layer was transferred to a new microfuge tube. DNA was recovered using an EZ-10 Spin column in accordance with manufacturer’s instructions (Bio Basic, Inc., Ontario, Canada), following precipitation by the addition of 10 M ammonium acetate and 99% ethanol (Sigma Aldrich, MO, USA) in a 1:10 and 1:1 ratio with sample volume, respectively. DNA was eluted in 50 µl UltraPure DNase/RNase-free distilled water (Gibco, ThermoFisher Scientific, Victoria, Australia) and stored at -80 °C prior to analysis. 16S rRNA gene amplicon sequencing The V1-3 hypervariable region of the bacterial 16S rRNA gene was amplified from sputum DNA using modified primers 27F (5'TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAGRGTTTGATCMTGGCTCAG-3') and 519R (5'GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGTNTTACNGCGGCKGCTG-3'), with Illumina adapter overhang sequences as indicated by underline. Amplicons were generated, cleaned, indexed and sequenced according to the Illumina MiSeq 16S Metagenomic Sequencing Library Preparation protocol 2
ACCEPTED MANUSCRIPT
RI PT
(http://support.illumina.com/downloads/16s_metagenomic_sequencing_library_preparation.ht ml) with certain modifications. Briefly, an initial PCR reaction contained at least 12.5 ng of DNA, 5 µL of forward primer (1 µM), 5 µL of reverse primer (1 µM) and 12.5 µL of 2× KAPA HiFi Hotstart ReadyMix (KAPA Biosystems, Wilmington, MA, USA) in a total volume of 25 µL. The PCR reaction was performed on a Veriti 96-well Thermal Cycler (Life Technologies) using the following program: 95 °C for 3 min, followed by 25 cycles of 95 °C for 30 sec, 55 °C for 30 sec and 72 °C for 30 sec and a final extension step at 72 °C for 5 min. Samples were multiplexed using a dual-index approach with the Nextera XT Index kit (Illumina Inc., San Diego, CA, USA) according to the manufacturer’s instructions. The final library was paired-end sequenced at 2 × 300 bp using a MiSeq Reagent Kit v3 on the Illumina MiSeq platform. Sequencing was performed at the David R Gunn Genomics Facility, South Australian Health and Medical Research Institute.
M AN U
SC
16S rRNA gene qPCR Approximate 16S rRNA gene copy number was assessed by quantitative PCR (qPCR) using the 16S rRNA universal primers B331F (5'-TCCTACGGGAGGCAGCAGT-3') and B797R (5'-GGACTACCAGGGTATCTAATCCTGTT-3') using Platinum SYBR Green 8 (ThermoFisher scientific, Vic, Australia) as previously described. Reactions were performed in duplicate and averages taken. Sample total bacterial copy number was calculated per µL of DNA eluate against a standard curve of a known bacterial copy number.
AC C
EP
TE D
Sequence data processing The Quantitative Insights Into Microbial Ecology (QIIME, v1.8.0)9 software was used to analyse the 16S rRNA sequence generated from paired-end amplicon sequencing using bioinformatics pipeline as previously described.10 Briefly, barcoded forward and reverse sequencing reads were quality filtered and merged using Paired-End reAd mergeR (PEAR v0.9.6).11 Chimeras were detected and filtered from the paired-end reads using USEARCH (v6.1)12 against the 97% clustered representative sequences from the Greengenes database (v13.8).13 Operational taxonomic units (OTUs) were assigned to the reads using an open reference approach with UCLUST algorithm (v1.2.22q) against the SILVA database release 111 (July 2012)14 that was clustered at 97% identity. Spurious OTUs were then removed systematically using previous reports of common laboratory sequencing contaminants.15 A minimum subsampling depth of 1,732 reads was then selected for all samples. Where taxa assignment failed to classify to the Family or Genus level, OTU reference sequences (accounting for >99% of OTU reads) were separately aligned using SILVA Incremental Aligner (SINA) (https://www.arb-silva.de/) which uses SILVA, RDP, Greengenes, LTP and EMBL sequence collections. If the alignments identified taxa to a genus level, and at >99% similarity, they replaced the previous taxon assignment. This occurred for Streptococcus II which was previously incorrectly assigned as Clostridiales;Other;Other. Streptococcus I refers to the OTU cluster which was assigned as Streptococcus during initial assignment. Diversity measurements and statistical analyses Bray-Curtis matrix was calculated based on sample-normalised, square root transformed relative taxon abundance. Principal coordinate analysis (PCoA) was used to visualize clustering of samples based on their similarity matrices with PCO1 and PCO2 coordinates and group centroids plotted using ggplot2 package of R statistical software.16 Distance from centroid was calculated as previously described, using PRIMER.17 Permutational multivariate analysis of variance (PERMANOVA)18 on the beta-diversity matrices was used to test the null hypothesis of no difference amongst a priori-defined groups using PERMANOVA + addon package for PRIMER. The test was computed using unrestricted permutation of raw data with 9,999 random permutations and at a significance level of 0.01. 3
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Taxon dispersion Variation in microbiota composition at the genus-level was assessed using multiple approaches. First, taxa that contributed to the overall variation between the asthma phenotypes were identified using SIMilarity of PERcentages (SIMPER) analysis in PRIMER. Subsequently, the abundance of the 13 highest ranked taxa (accounting for 50% of the dissimilarity between neutrophilic and eosinophilic groups) were used to generate a heatmap using ggplot2 package of R statistical software.16 Hierarchical clustering of the taxa was performed on Bray-Curtis dissimilarity and clustered using single linkage method. Dominance of Haemophilus and Moraxella was determined when the relative abundance of each taxa exceeded 40%. This cut-off was selected based to the distribution of the relative abundance, where a clear distinction between samples with >40% and 40% of total reads, their relative abundance was adjusted to the mean value for the study cohort and the remaining relative abundance measures rescaled, as described previously.20 PERMANOVA analyses were then performed on the rescaled data. Second, pairwise comparisons between neutrophilic and eosinophilic samples were performed using the phyloseq R package21 with DEseq222 extension, based on count data. p values were corrected using the Benjamini-Hochberg false discovery rate procedure and a corrected alpha value cut-off of
