Scientific Papers

Acellular ex vivo lung perfusate silences pro-inflammatory signaling in human lung endothelial and epithelial cells | Journal of Translational Medicine


Similar to what we have observed in human lung transplants studies, we found that IR simulation activated pathways related to inflammatory responses in human lung endothelial and epithelial cells. On the other hand, different from what we have seen in human lungs, our EVLP models did not involve inflammatory responses. The cell culture models will be helpful for interpreting the direct and indirect responses of these lung parenchymal cell types to IR and EVLP conditions in the human lung grafts.

IR condition activates inflammatory responses directly at cellular levels

Wong et al. analyzed the transcriptomic changes in human lung biopsies before and after transplantation, and found upregulated themes of inflammation, cell death, and heat stress and downregulated themes of metabolism and protein synthesis [12]. The lung allografts contain a wide range of cell types, such as donor endothelial and epithelial cells, and macrophages, monocytes, neutrophils, and natural killer cells from both donor and recipient. Various interactions among these cells have been reported to contribute to IR injury [34]. The contribution of lung endothelial and epithelial cells to inflammatory responses cannot be determined with the bulk RNA studies in lung tissues.

In both lung endothelial and epithelial IR models, we identified upregulation of inflammatory signals that were seen in human lung transplant samples, such as response to bacteria, response to TNF and IL-1, regulation of mitogen-activated protein kinase (MAPK) and Janus kinase (JAK)-signal transducer and activator of transcription protein (STAT) signaling, response to inflammatory cytokines, TCR signaling, human immunodeficiency virus (HIV)-negative regulatory factor (NEF) and TNF signaling, regulation of blood coagulation, and regulation of leukocyte chemotaxis. On the other hand, regulation of apoptosis signaling, cell death signaling, and IL-2 signaling were only seen in endothelial cells, whereas response to TLR/myeloid differentiation primary response 88 (MyD88) signaling and B-cell receptor (BCR) signaling were only seen in epithelial cells. Similarly, downregulation of pathways in amino acid metabolism, protein translation, oxidative phosphorylation, and DNA repair, seen in human lung transplants, were also detected in both human lung endothelial and epithelial cells (Fig. 6). These cell culture results indicate that human lung endothelial and epithelial cells may play an essential role in IR-induced inflammatory responses. Activation of inflammatory responses in these residential cells may prime the donor lung to further interact with donor and recipient immune cells for subsequent tissue injury. The pathways that were not detected or found to be significant in our cell culture models, such as regulation of reactive oxygen species (ROS) metabolism, IL-12 and IL-23 signaling, regulation of adaptive immunity, Sphingosine-1-phosphate (S1P) signaling and fatty acid beta-oxidation (Fig. 6) could be attributed to other cell types and/or to the interactions among different cell types.

Fig. 6
figure 6

Simulated ischemia–reperfusion (IR) in human lung endothelial and epithelial cell cultures induced similar pathways as reported in human lung transplants (11). Dark red nodes indicate upregulated pathway clusters, while blue nodes indicate downregulated clusters observed in human lung transplants. Blue-labeled tags are attached to the pathway boxes if identified in epithelial cells (BEAS-2B) and pink-labeled tags if identified in endothelial cells (HPMEC)

Acellular EVLP perfusate does not activate inflammatory responses in cell culture models

On the contrary, after cold preservation, the simulated EVLP did not show enriched pathways related to pro-inflammatory responses in both cell types, which was the opposite to what Wong et al. observed in the human lung biopsies after EVLP [12]. In animal models and clinical studies, inflammatory mediators are released from the donor lungs into the EVLP perfusate, and using cytokine filters to reduce the levels of inflammatory mediators ameliorated inflammation and improved lung graft function during EVLP and post-LTx [35,36,37,38]. In addition, Yeung et al. examined gene expression profiles of human donor lungs during EVLP, and found that EVLP may improve donor lung function through the washout of leukocytes and facilitate innate mechanisms of repair [39]. In our EVLP cell culture model, these inflammatory mediators and other types of cells are not included in the system, thus this may partially explain the lack of pathways in inflammation and cell death.

However, in our IR models, the inflammatory mediators and other cell types are also missing, but the serum and nutrients in the medium may be activating genes related to endothelial and epithelial cell functions and those involved in the pro-inflammatory responses. It has been shown that simulated reperfusion condition in cell culture can increase the release of pro-inflammatory cytokines and induce apoptosis and necroptosis, and these mechanisms identified in the cell culture models have been validated in animal models or human lung samples [17,18,19, 22, 23, 32]. The Steen solution, which only contains buffered electrolytes, glucose, albumin and dextran 40, may not fully activate genes related to endothelial/epithelial functions and pathways related to inflammation and cell death. The acellular Steen solution used for EVLP may provide a protective milieu for donor lung cells to recover from the stress induced by the hypothermic and ischemic condition. These results support a recent literature review on EVLP clinical trials reporting a reduced incidence of PGD following LTx for donor lungs assessed with EVLP [9].

Differential responses of endothelial and epithelial cells to IR and EVLP

In both IR and EVLP models, human pulmonary microvascular endothelial cells had nearly double the number of DEGs than that of human lung epithelial cells. Moreover, between the EVLP models, the number of gene sets downregulated in the endothelial cells was far more than that of the epithelial cells. These results suggest that endothelial cells might be more responsive to cold preservation and warm (re)perfusion conditions. On the contrary, Saren et al.’s study that directly compared between the two cell types undergoing cold ischemia and warm reperfusion has reported significantly different transcriptomic profiles for each cell type, and the epithelial cells were suggested to be more sensitive to IR conditions than endothelial cells, due to the disappearance of epithelial-specific clusters after IR [16]. However, the disappearance of epithelial-specific pathways may be due to the disorganized regulation of genes that did not reach statistical significance in pathway analyses, whereas the number of significant DEGs may reflect a better overview of the responsiveness of each cell type. Even though both endothelial cells and epithelial cells are abundant in the lung, their responses to cold donor lung preservation and warm reperfusion or EVLP can be distinctive. Therefore, cell-type-specific biomarkers and therapeutic targets should be explored in the future studies.

Limitations and future directions

Interestingly, a vascular process theme was observed in our epithelial IR model. Because of its bronchial epithelial origin, BEAS-2B cells have been widely used in lung cell studies. However, very few studies have reported their non-epithelial features, some of which are comparable with those seen in human mesenchymal stem cells [40]. Whether the gene set regulations in the vascular process theme of epithelial IR model can be found in vivo needs to be determined.

To compare with transcriptomic data from human lung transplant, we used GSEA and Cytoscape for pathway analyses. Ingenuity pathway analysis (IPA) is another commonly used method for transcriptomics studies. The bulk RNA sequencing data collected in this study can be analyzed with various methods and should be explored further.

In the present study, we used fetal bovine serum in cell culture medium. Even though this is a common practice in cell biology studies, considering the transplant setting, using bovine serum for human cell culture, may have different stimulatory effects from human blood. This should be considered as a potential cofounding factor.

Moreover, due to the nature of mono cell cultures of two-dimensions, there were no interactions among different cell types, which would have induced more diverse signals after warm (re)perfusion. Hence, whether our results can be reproduced in three-dimension cell cultures or co-cultures should be explored further. In addition, both IR and EVLP models did not include flow and ventilation, which are important factors for lung cell physiology [41, 42]. Advancements in development of the experimental cell culture models representing in vivo lung environment may provide a closer reflection of molecular responses in clinical scenarios.

Lastly, because the findings are at a transcriptional level, we are yet to investigate whether proteomics and metabolomics data would support or rebut our current findings. In future studies, a more developed version of cell culture models simulating IR and EVLP and a multi-omics approach would yield a more solid proposition on the mechanisms underlying IR and acellular EVLP.

In conclusion, using bulk RNA sequencing technique, we profiled transcriptomic changes of human lung endothelial and epithelial cells in simulated IR and EVLP models. The similarity of inflammatory responses observed in these cell cultures and in human lung transplants suggests that lung endothelial and epithelial cells can play a huge role in IR-induced injury of the lung grafts. The absence of inflammatory responses in our EVLP models suggests that the lack of serum components in the acellular Steen solution may limit the activation of pro-inflammatory signals, thus supporting the protective role of clinical EVLP.



Source link