ATS 2021 will feature presentations of original research from accepted abstracts. Mini Symposia and Thematic Poster Sessions are abstract based sessions.
Please use the form below to browse scientific abstracts and case reports accepted for ATS 2021. Abstracts presented at the ATS 2021 will be published in the Online Abstract Issue of the American Journal of Respiratory and Critical Care Medicine, Volume 203, May 3, 2021.
Search Tips:
- Use the keyword search to search by keyword or author's name.
- Filter your search results by selecting the checkboxes that apply.
- Click on "Clear" to clear the form and start a new search. .
Search results will display below the form.
Lung Proteomics Biomarkers Associated with COPD
Session Title
TP112 - TP112 PROTEOMICS/GENOMICS/METABOLOMICS IN LUNG DISEASE
Abstract
A4342 - Lung Proteomics Biomarkers Associated with COPD
Author Block: Y. Zhang1, M. Hoopmann2, P. Castaldi1, K. Simonsen2, M. Midha2, M. H. Cho3, G. J. Criner4, R. Bueno5, J. Liu1, R. Moritz2, E. K. Silverman3; 1Channing Division of Network Medicine, Brigham and Women's Hospital, Boston, MA, United States, 2Institute for Systems Biology, Seattle, WA, United States, 3Channing Division of Network Medicine, Brigham and Women’s Hospital, Boston, MA, United States, 4Pulm & Crit Care Med, Temple Univ Hosp, Philadelphia, PA, United States, 5Division of Thoracic Surgery, Brigham and Women’s Hospital, Boston, MA, United States.
Introduction: Molecular biomarkers could assist in chronic obstructive pulmonary disease (COPD) diagnosis, risk evaluation, and progression monitoring. Identifying protein biomarkers for COPD has been challenging, especially in lung samples. Most previous proteomics studies have analyzed individual proteins or pre-selected protein panels in blood samples or small numbers of lung tissue samples. We analyzed a larger sample size of lung tissue specimens comprehensively to identify COPD-associated proteins.
Methods: We analyzed lung tissue samples of 100 COPD subjects and 52 controls obtained during clinical thoracic surgery procedures. All subjects were ex-smokers with at least five pack-years of smoking. Proteins from lung tissues were extracted and analyzed in triplicate using nano liquid chromatography high resolution QExactive HF tandem mass spectrometry (LC-MS/MS). The raw proteomics mass spectrometry data were averaged, filtered, normalized, and imputed. Linear regression models including age, sex, batch effects, and surrogate variables as covariates and three machine learning models (Random Forest, Elastic Net and Naïve Bayes) using top COPD-associated proteins as features were applied to identify potential COPD-associated protein biomarkers. Functional enrichment analyses and protein quantitative trait loci (pQTL) analyses on the top 5% of COPD-associated proteomics biomarkers were also applied.
Results: After quality control, 150 subjects and 4407 proteins were used for COPD biomarker identification. Using linear regression models, twenty-five protein biomarkers were significantly associated with COPD at FDR<0.05; the most significantly associated proteins were agrin (AGRN), annexin A2 (ANXA2), and retinoic acid-induced protein 3 (GPRC5A). Previously reported plasma protein biomarkers were not significantly associated in lung tissue, although RAGE was borderline significant (FDR=0.063). Machine learning models using Random Forests with the top 5% of protein biomarkers demonstrated reasonable accuracy (0.766) and AUC (0.702) for COPD prediction. The top 5% of proteins were enriched in multiple gene ontology (GO) terms potentially associated with COPD, including GO: 0044331 (cell-cell
adhesion mediated by cadherin). pQTL analyses were performed for the previously reported 82 COPD-associated SNPs from genome-wide association analysis and the top 5% of COPD-associated proteins. For rs2070600, a COPD-associated SNP in the AGER gene (which encodes RAGE), we found cis-pQTLs with AGER and CO2 at FDR < 0.05.
Conclusion: Twenty-five COPD-associated protein biomarkers in lung tissue were identified. Machine learning models demonstrated reasonable performance for COPD prediction. With pQTL analyses, we identified suggestive evidence of local gene regulatory effects of the AGER SNP rs2070600. These results may assist in understanding the complex pathobiological mechanisms of COPD.
Author Block: Y. Zhang1, M. Hoopmann2, P. Castaldi1, K. Simonsen2, M. Midha2, M. H. Cho3, G. J. Criner4, R. Bueno5, J. Liu1, R. Moritz2, E. K. Silverman3; 1Channing Division of Network Medicine, Brigham and Women's Hospital, Boston, MA, United States, 2Institute for Systems Biology, Seattle, WA, United States, 3Channing Division of Network Medicine, Brigham and Women’s Hospital, Boston, MA, United States, 4Pulm & Crit Care Med, Temple Univ Hosp, Philadelphia, PA, United States, 5Division of Thoracic Surgery, Brigham and Women’s Hospital, Boston, MA, United States.
Introduction: Molecular biomarkers could assist in chronic obstructive pulmonary disease (COPD) diagnosis, risk evaluation, and progression monitoring. Identifying protein biomarkers for COPD has been challenging, especially in lung samples. Most previous proteomics studies have analyzed individual proteins or pre-selected protein panels in blood samples or small numbers of lung tissue samples. We analyzed a larger sample size of lung tissue specimens comprehensively to identify COPD-associated proteins.
Methods: We analyzed lung tissue samples of 100 COPD subjects and 52 controls obtained during clinical thoracic surgery procedures. All subjects were ex-smokers with at least five pack-years of smoking. Proteins from lung tissues were extracted and analyzed in triplicate using nano liquid chromatography high resolution QExactive HF tandem mass spectrometry (LC-MS/MS). The raw proteomics mass spectrometry data were averaged, filtered, normalized, and imputed. Linear regression models including age, sex, batch effects, and surrogate variables as covariates and three machine learning models (Random Forest, Elastic Net and Naïve Bayes) using top COPD-associated proteins as features were applied to identify potential COPD-associated protein biomarkers. Functional enrichment analyses and protein quantitative trait loci (pQTL) analyses on the top 5% of COPD-associated proteomics biomarkers were also applied.
Results: After quality control, 150 subjects and 4407 proteins were used for COPD biomarker identification. Using linear regression models, twenty-five protein biomarkers were significantly associated with COPD at FDR<0.05; the most significantly associated proteins were agrin (AGRN), annexin A2 (ANXA2), and retinoic acid-induced protein 3 (GPRC5A). Previously reported plasma protein biomarkers were not significantly associated in lung tissue, although RAGE was borderline significant (FDR=0.063). Machine learning models using Random Forests with the top 5% of protein biomarkers demonstrated reasonable accuracy (0.766) and AUC (0.702) for COPD prediction. The top 5% of proteins were enriched in multiple gene ontology (GO) terms potentially associated with COPD, including GO: 0044331 (cell-cell
adhesion mediated by cadherin). pQTL analyses were performed for the previously reported 82 COPD-associated SNPs from genome-wide association analysis and the top 5% of COPD-associated proteins. For rs2070600, a COPD-associated SNP in the AGER gene (which encodes RAGE), we found cis-pQTLs with AGER and CO2 at FDR < 0.05.
Conclusion: Twenty-five COPD-associated protein biomarkers in lung tissue were identified. Machine learning models demonstrated reasonable performance for COPD prediction. With pQTL analyses, we identified suggestive evidence of local gene regulatory effects of the AGER SNP rs2070600. These results may assist in understanding the complex pathobiological mechanisms of COPD.
