In a large animal model of single-lung atelectasis, regional lung tissue proteomics and transcriptomics analyses revealed: (1) substantial differences in protein abundance patterns occurred in the early stages of lung atelectasis with and without systemic LPS exposure; (2) decreased inflammatory proteomic signatures in atelectasis in the absence of systemic LPS; (3) increased inflammatory signatures with high levels of acute lung injury marker RAGE, chemokine CCL5 and signaling factors BTK and STAT1 during LPS exposure; (4) alveolar-capillary barrier dysfunction in atelectasis with TGF-β, Hippo signaling and extracellular matrix components such as THBS1 as potential targets; and (5) an overlapped subset of gene expression and protein abundance in atelectatic lung regions.
Evidence is limited and controversial on the biological effects of atelectasis5,8,9,10. We have recently indicated that gene expression is remarkably distinct in atelectatic versus aerated lung either in the absence or presence of systemic LPS exposure16. Yet, it is unknown whether and which specific components of that transcriptomics response correlate or result in ultimate protein level changes during the early stages of atelectasis.
Our one-lung ventilation model allowed for direct visual and computed tomography documentation of atelectasis, and produced expected physiological impairment of pulmonary mechanics and gas exchange, supporting the reliability of tissue samples. Resulting proteomics were consistent with a hypoxemic atelectatic lung, as expressed by the increased glycolytic enzyme PGAM1 and positively enriched HIF-1 pathway, related to hypoxic stress and relevant for epithelial, endothelial, and immune/myeloid cell response in ALI/ARDS19. Endotoxemia promoted further metabolic changes in atelectasis with increased glycolytic enzymes, carbon metabolism and glycolysis pathways. Previous work has validated the used aptamer-based proteomics methodology20,21. Such observations support the reliability of our atelectasis model and proteomics measurements.
The proteomics response was remarkably distinct in atelectatic versus aerated regions from the same lung both with and without LPS. Such differential response was relevant in two key components of lung injury: inflammation and alveolar-capillary barrier function. Without LPS, atelectasis by itself was associated with decreased immune function, i.e., neutrophil migration. This finding is consistent with reduced leukocyte signaling and inflammation in the dependent lung of premature lambs18. Of note, no significant difference in blood volume per tissue was found between atelectasis and aerated lung regions, indicating comparable blood cell numbers in vessels per tissue in both lung regions. This finding implies that the observed decreased immune function in atelectasis is likely associated with other factors rather than immune cell numbers. Indeed, our proteomics data helped explain and suggested the downregulated cellular activities, as the production of interleukin-12, with important role in neutrophil migration, was negative enriched in atelectasis.
LPS produced a switch of that proteomic inflammatory response in atelectasis from repressed towards increased. Inflammatory factors, such as IL-6, IL-10, and RAGE, all established biomarkers of ARDS22,23, and MPO, a neutrophil marker24, were significantly increased in atelectasis with LPS. Such findings are consistent with previous in vivo inflammatory assessments of regional lung inflammation25,26 and support the plausibility of our model and findings. The pro-inflammatory proteomics response in atelectasis was characterized by positively enriched leukocyte-associated processes including leukocyte migration. Interestingly, blood volume normalized by tissue was found lower in atelectasis than in aerated lung during LPS exposure potentially due to the effects of gravity and inhibited hypoxic pulmonary vasoconstriction. Such finding implies that our observation of enhanced inflammatory activity in atelectasis regions with LPS exposure cannot be merely assigned to a change in regional blood volume, as lower not higher inflammatory response would be expected. Indeed, our proteomics findings indicated the increased cellular responses even in the setting of reduced blood volume per tissue. It can be understood by the positively enriched interleukin-8 production and chemokines signaling, and is consistent with previous genomics data with higher levels of chemokines (e.g., CXCL10 and CCL5) in atelectasis with LPS16. Our findings are also consistent with increased inflammatory response of isolated atelectatic versus expanded lungs in patients undergoing esophagectomy5, a condition related to systemic inflammation27. It may be additionally important given the correlation between increased inflammatory cytokine TNF-α in atelectatic lung and worsened pulmonary outcomes after lung resection6.
Additionally, our proteomics data suggest potential mechanistical and treatment protein targets for such atelectasis-induced inflammatory response. These included: (a) BTK, involved with neutrophil and macrophage response in infection-induced ARDS such as influenza28, SARS-CoV-229 and LPS/immunocomplex30, promotes a pro-inflammatory neutrophilic response through TLR4 signaling30; and (b) STAT1, an interferon-induced stimulator with function in CCL5 and CXCL11 regulation31. STAT1 may not only have contributed to the observed increases in CCL5 and CXCL11, but could also explain the transcriptomic findings of up-regulated interferon-stimulated genes in LPS-exposed atelectasis16. Overall, our results demonstrate inflammatory and corresponding mechanistic targets in atelectatic relative to aerated lung when acutely (8 h) exposed to endotoxemia.
We also found proteomic evidence for alveolar-barrier structure dysregulation in atelectasis. Of note, without LPS such response was associated to cellular function impairment in epithelium, endothelium and fibroblast growth signaling described in influenza and SARS-CoV-2-induced ARDS32. In contrast, dysfunction in cell junction assembly and cell–matrix/substrate adhesion, processes found in lung injury33,34, predominated in atelectasis with LPS. Negatively enriched TGF-β signaling, important in alveolar type-2 cell regeneration during LPS-induced murine lung injury35, further suggested barrier dysfunction in atelectasis. Our results are consistent with previous observations in the dependent lung of preterm lambs and recent transcriptomics findings in atelectasis16,18. The rate of imaging tracer influx from blood into lung tissues (K1) larger in atelectasis versus aerated lung regions could partially reflected ultrastructural or functional changes occurring in the alveolar-capillary barrier in atelectasis. This possibility is also consistent with the larger proportional increase of K1 in atelectatic than aerated tissue with LPS exposure. In line with such findings, RAGE, an injury marker of type-1 cell in acute lung injury36, was significantly increased in atelectasis during LPS exposure. Overall, given that our tissue samples derived from a non-ventilated lung, our findings imply that atelectasis, either with or without LPS exposure, could contribute to alveolar-capillary barrier dysfunction even in the absence of biomechanical injurious forces.
Of note, the Hippo pathway was significantly enriched in atelectatic versus aerated lungs. This pathway is associated with epithelial proliferation after lung injury37, and consequently a potential factor for alveolar-capillary barrier dysfunction. As an important effector of Hippo pathway, Yes-associated protein (YAP) was lower in atelectasis, validated by immunofluorescent staining and supported by the lower expression of YAP-responsive gene THBS1. Recently, YAP has been reported as involved in barrier function via regulation of cytoskeleton dynamics38. Our PCR validations confirmed the lower expression for YAP regulated cytoskeleton organization-associated genes (e.g., ACTN1, FLNA, FLNC, SRF and RHOD). Together these data reveals the consistency of proteomics with recent transcriptomics in barrier dysfunction, and provides further support to the link of Hippo-YAP signaling with atelectasis16. Such findings suggest the role of the Hippo pathway not only as a possible factor in the mechanism of atelectasis-related lung injury, but also as a potential therapeutic target aiming at barrier function improvement.
Additionally, extracellular matrix glycoproteins (e.g., THBS1 overlapped in transcriptomics and proteomics as well as validated in gene and protein levels) were lower in atelectatic than aerated lung with and without LPS. THBS1, associated with ventilator-induced lung injury39, and THBS2, with function in extracellular matrix assembly40, are both related to lung remodeling41. FSTL3, involved with epithelial regulation, is lower in patients with asthma or bronchoconstriction than in healthy individuals42. VEGFA, a major glycoprotein inducer of integrins and downstream angiogenesis processes, has been involved in the pathogenesis of ARDS during oesophagectomy43 and decreased in acute models of LPS-induced lung injury and early ARDS44,45. These findings support the ability of our model to reproduce previous lung injury observations, and present novel findings suggesting the presence of dysregulated barrier function by atelectasis regardless of LPS exposure.
Of note, there has been substantial controversy on the use of ventilatory management strategies directed to minimization of atelectasis both in surgical and critically-ill patients1,4. Biological information has been scant, with experimental studies reflecting controversial results in humans46,47, and the debate centered on results of specific clinical trials and global physiological studies. Using large animal models mimicking clinical settings and representing well the human condition structurally and physiologically, our demonstration of substantial transcriptomics and proteomics differences in atelectasis provides biological plausibility to optimal recruitment techniques beyond their specific effect on biomechanical forces.
There are limitations in our study. The sample size limits power to detect significant differences in regional protein concentrations. Our large animal model focused on the early stages of atelectasis with timeframe of 8 h, which is relevant to the perioperative period involving surgeries using one-lung ventilation such as lung resection, pneumonectomy, and thoracoabdominal aortic aneurysm repair. Thus, results are expected to relate to mechanisms of lung injury in the early stages of atelectasis and did not address the long-term effects of atelectasis. Our analysis does not identify the specific cells contributing to the proteomics findings. While functional analysis suggested cell subtypes, further studies will be required to determine cell-specific responses. Our large animal model represents clinical conditions of one-lung ventilation with mild-moderate histological lung injury16. Accordingly, endothelial and epithelial cell death following atelectasis has not been detected at this early time point. Further investigation extending phenotypic characterization and molecular biology is necessary to gain full understanding in atelectasis-associated alveolar barrier dysregulation. Our findings are restricted to the utilized model and do not describe all mechanisms and insults found in patients developing lung injury.
In conclusion, atelectatic tissue presented proteomics patterns consistent with a dysregulated immune response. Additional systemic endotoxin shifted this into an enhanced regional inflammatory response, e.g., increased acute lung injury marker RAGE, chemokine CCL5 and signaling factors BTK and STAT1 in atelectatic versus aerated lung. Alveolar-capillary barrier dysfunction in atelectatic lung regions was independent of LPS exposure. TGF-β signaling, Hippo pathway and extracellular matrix components (e.g., THBS1) can be potential culprits for the dysregulated barrier function in atelectasis.
