Identification of early and intermediate biomarkers for ARDS mortality by multi-omic approaches

Abstract

The lack of successful clinical trials in acute respiratory distress syndrome (ARDS) has highlighted the unmet need for biomarkers predicting ARDS mortality and for novel therapeutics to reduce ARDS mortality. We utilized a systems biology multi-“omics” approach to identify predictive biomarkers for ARDS mortality. Integrating analyses were designed to differentiate ARDS non-survivors and survivors (568 subjects, 27% overall 28-day mortality) using datasets derived from multiple ‘omics’ studies in a multi-institution ARDS cohort (54% European descent, 40% African descent). ‘Omics’ data was available for each subject and included genome-wide association studies (GWAS, n = 297), RNA sequencing (n = 93), DNA methylation data (n = 61), and selective proteomic network analysis (n = 240). Integration of available “omic” data identified a 9-gene set (TNPO1, NUP214, HDAC1, HNRNPA1, GATAD2A, FOSB, DDX17, PHF20, CREBBP) that differentiated ARDS survivors/non-survivors, results that were validated utilizing a longitudinal transcription dataset. Pathway analysis identified TP53-, HDAC1-, TGF-β-, and IL-6-signaling pathways to be associated with ARDS mortality. Predictive biomarker discovery identified transcription levels of the 9-gene set (AUC-0.83) and Day 7 angiopoietin 2 protein levels as potential candidate predictors of ARDS mortality (AUC-0.70). These results underscore the value of utilizing integrated “multi-omics” approaches in underpowered datasets from racially diverse ARDS subjects.

The genomic landscape of Acute Respiratory Distress Syndrome: a meta-analysis by information content of genome-wide studies of the host response

“…These 40 studies produced 44 unique gene lists (22 transcriptomic, 13 proteomic, and 9 based on genome-wide association studies (GWAS); see Table 1). Three studies reported results from multiple methodologies 33,38,44 , and several used more than one tissue type 18,21,32 . Excluding GWAS, 14 gene lists (40%) were derived from lung or airways samples, and 21 (60%) from blood.…”

Postoperative bioactive adrenomedullin is associated with the onset of ARDS and adverse outcomes in patients undergoing open thoracoabdominal aortic surgery

“…Early detection of the patients at risk for ARDS remains of paramount importance, since it is associated with relevant mortality rates of 30–40% 34 . Given the multitude of stimuli and their heterogeneity, that may contribute to the onset of ARDS, the need of biomarkers as an extension of clinical scores for the prediction of ARDS-related mortality has been highlighted in the literature 35 , 36 . Johnsson et al recently reported a significant elevation of bio-ADM levels at the time of admission in ICU in patients that were diagnosed with ARDS within 72 h after admission 37 .…”

Identification of a pediatric acute hypoxemic respiratory failure signature in peripheral blood leukocytes at 24 hours post-ICU admission with machine learning

“…For example, HDACs repress the production of proinflammatory cytokines in alveolar macrophages in chronic inflammatory lung conditions such as chronic obstructive pulmonary disease (44,45). HDAC class I and chromatin modifying enzymes were identified as top pathways associated with ARDS mortality in a study using a multi-omics approach with multiple publicly available datasets (46). DNA methylation is another epigenetic modification that emerged in the Reactome analysis that distinguished moderate/ severe from mild pediatric ARDS.…”