An Exploratory Study Based On Asthma Phenotypes And Serum Cytokine Profiles

Background and Objection:Asthma is a chronic inflammatory airways disease with distinct heterogeneity characterized by various clinical manifestations and disease mechanisms. The term "asthma phenotype" was put forward based on this heterogeneity of the disease. Many cytokines involved in asthma occur and progress, and the cytokine networks play a critical role in heterogeneity of asthma. A panoramic view of asthmatic serum cytokine expression will not only contribute to better knowledge of asthma heterogeneity, but also contribute to selection of potential target for asthma treatment.We put forward the hypothesis that (1) inflammatory phenotypes of asthma were shaped by serum cytokines partly;(2) serum cytokine profiles have the ability to identify relatively homogeneous subgroup of patients with distinct clinical characteristics so that we could partly elucidate the asthma heterogeneity. To test our hypothesis we measured serum concentrations of twenty cytokines that were recently described to play an important role in asthma pathogenesis:INF-y, TSLP, IL-4, IL-5, IL-13, TNF-a, IL-9, IL-33, IL-6, TGF-p1, IL-23, IL-10, IL-17, IL-8, VEGF, EGF, sRAGE, IL-10, GM-CSF and leptin. We analyzed the serum cytokine profiles in asthma patients stratified by sputum granulocytes (inflammatory phenotypes), and attempted to distinguish asthma cytokine phenotypes by unsupervised hierarchical clustering analysis.Chapter I Clinical data collection and microarray analysis of serum cytokine profilesObjectives:To collect clinical data and to detect the serum cytokines concentrations.Methods:We recruited50asthmatic patients in the non-acute episode phase (asthma group) and20normal control subjects (control group). Lung function data, demographic characteristics,5-item asthma control questionnaire (ACQ-5) and symptom score (daytime and nighttime) of asthmatics were collected before sputum induction, and the sputum was collected for cell differential. Venous blood samples of all subjects were collected at the same visit for cell differential, and serum was separated. Serum total IgE concentration and cytokine profiles were detected using electrochemiluminescence and custom Quantibody array, respectively.Results:Serum concentrations of IL-6and IL-8in asthmatics were statistically higher than in control subjects[2.50(1.08-4.08) pg/ml,1.07(0.86-1.07) pg/ml vs0.67(0.00-2.43) pg/ml,0.74(0.59-0.98) pg/ml; P=0.002,0.002]. Serum concentrations of sRAGE in asthmatics were statistically lower than in control subjects[368.94±182.37pg/ml vs368.94±182.37pg/ml, P=0.004]. The serum concentrations of INF-y, TSLP, IL-4, IL-5, IL-13, TNF-a, IL-9, IL-33, TGF-p1, IL-23, IL-10, IL-17, VEGF, EGF, IL-1βp, GM-CSF and leptin showed no significant difference between the asthma group and the control group.Conclusion:(1) Clinical data collection and microarray analysis of serum cytokine profiles were successful.(2) Comparison of serum cytokines between asthmatcs and normal controls irrespective of asthma heterogeneity might mask the molecular characteristics of certain asthma subtypes.Chapter Ⅱ Correlation analysis between asthma clinical data and serum cytokine concentrationsObjectives:To analyze the relationship between asthma clinical data and serum cytokine concentrations irrespective of asthma phenotypes.Methods:Subjects were the50asthmatics of the chapter I. The correlations between asthma clinical parameters and serum cytokine concentrations were analyzed. These clinical parameters included asthma duration (years), forced expiratory volume in1 second percentage predicted (FEV1%pred), ACQ-5, total serum IgE concentrations, induced sputum eosinophil count, induced sputum neutrophil count and peripheral eosinophil count.Results:FEV1%pred correlated inversely with serum concentrations of sRAGE (r=-0.316; P=0.025). ACQ-5correlated positively with serum concentrations of IL-8and leptin respectively (r=0.299,0.295; P=0.035,0.038, respectively). Sputum neutrophil count correlated positively with serum concentrations of GM-CSF (r=0.358; P=0.011). Peripheral eosinophil count correlated positively with serum concentrations of IL-33(r=0.295; p=0.038). No association was found between sputum eosinophil count, serum total IgE level and serum cytokine concentrations.Conclusion:Without considering asthma subtypes, we could not interpret the result of correlation between asthma clinical parameters and serum cytokine concentrations very well.Chapter III Analysis of cytokine profiles in asthma subjects stratified by sputum granulocytesObjectives:To analyze cytokine profiles in asthma subjects stratified by sputum granulocytes.Methods:(1) Stratified by sputum eosinophils:Comparison of cytokine profiles between the normal controls, eosinophilic asthma (eosinophils≥3%) and non-eosinophilic asthma (eosinophils<3%) was conducted without considering percentage of neutrophils.(2) Stratified by sputum neutrophils:Comparison of cytokine profiles between the normal controls, neutrophilic asthma (neutrophils≥64%) and non-neutrophilic asthma (neutrophils<64%) was conducted without considering percentage of eosinophils.(3) Stratified by sputum eosinophils (<3%or≥3%) and neutrophils (<64%or≥64%):Comparison of cytokine profiles and clinical parameters between eosinophilic asthma (eosinophils≥3%+neutrophils<64%), neutrophilic asthma (eosinophils<3%+neutrophils≥64%) and mixed-granulocytic asthma (eosinophils>3%+neutrophils≥64%) was conducted. Results:(1) Stratified by sputum eosinophils:Statistical differences were found in serum concentrations of IL-6, IL-8, leptin, sRAGE among the three groups. Pairwise comparisons:1) Serum IL-6concentrations of eosinophilic asthmatics and non-eosinophilic asthmatics were statistically higher than that in normal controls.2) Serum IL-8concentrations of eosinophilic asthmatics were statistically higher than that in normal controls.3) Serum leptin concentrations of eosinophilic asthmatics were statistically higher than that in non-eosinophilic asthmatics.4) Serum sRAGE concentrations of eosinophilic asthmatics and non-eosinophilic asthmatics were statistically lower than that in normal controls.(2) Stratified by sputum neutrophils: Statistical differences were found in serum concentrations of GM-CSF, IL-6, IL-8, IL-10, sRAGE among the three groups. Pairwise comparisons:1) There was an overall significant difference in serum GM-CSF concentrations between three groups, but none reached significance in pairwise comparison.2) Serum IL-6concentrations of neutrophilic asthmatics were statistically higher than that in normal controls.3) Serum IL-8concentrations of neutrophilic asthmatics were statistically higher than that in normal controls.4) Serum IL-10concentrations of neutrophilic asthmatics were statistically higher than that in non-neutrophilic asthmatics and normal controls.(3) Stratified by sputum eosinophils and neutrophils:Statistical differences were found in lung function, ACQ-5score, daytime symptom score, serum GM-CSF, IL-8, leptin and VEGF level among the three groups. Mixed-granulocytic asthmatics had the lowest lung function and increased symptoms.Conclusion:Different inflammatory phenotypes of asthma were associated with distinct serum cytokine profiles. Mixed-granulocytic asthmatics were a subset of patients with the lowest lung function, which were associated with increased serum level of GM-CSF, IL-8, leptin and VEGF.Chapter IV Preliminary study of asthma molecular phenotypesObjectives:We sought to discriminate asthma phenotypes on the basis of cytokine profiles in serum.Methods:Subjects were the50asthmatics of the chapter I. The normalized data were subjected to statistical and informatics analysis.Results:Data of cytokine profiles were processed with principal component analysis, and we extracted the six largest principal components which were able to capture80.113%of the information from the original data. Three groups of asthmatics with distinct clinical characteristics could be identified on the basis of principal component hierarchical cluster analysis. Phenotype1had not only high level of proinflammatory cytokines including EGF, GM-CSF, INF-y, IL-4, IL-5, IL-6, IL-9, IL-17, IL-23and TNF-a but also high level of anti-inflammatory cytokines including IL-10, TGF-P and sRAGE. Phenotype2had low level of IL-10and sRAGE. Phenotype3had high level of Leptin, VFGF and low level of sRAGE. To determine whether the patients within these groups represented clinically distinct subgroups of asthma, the clinical features of the3groups were compared with one another. Phenotype1was enriched in male patients with relatively good lung functions and relatively slight symptoms. Phenotype2was enriched in patients with relatively good lung functions and relatively slight symptoms. Phenotype3was enriched in female patients with low lung functions and severe symptoms.Conclusion:(1) Serum cytokine profiles can be applied to identifing distinct types of asthma. Further biomarker discovery in serum may be informative.(2) Principal component analysis can be applied to analyzing the microarray data for reducing noise and removing redundance.(3) The severity of asthma may depend mainly on the balance between pro-inflammatory cytokines and anti-inflammatory cytokines.(4) The applications of microarray provide for distinguishing asthma phenotypes and identifying suitable populations of molecular targeted therapy.