Single-cell proteomics delineates murine systemic immune response to blast lung injury

Explosions commonly occur during military conflicts, terrorist attacks, and civilian practices. Coal mine gas, fireworks, flammable and explosive chemical products, and other factors are commonly reported as causes of explosions. In a recent study, coal mine gas was blamed for nearly half (49.1%) of China’s explosions, resulting in 67% of deaths and 26.7% of injuries¹. Among all types of coal mine accidents in Chinese coal mining industry, gas explosions are the leading cause of fatalities^2,3,4,5. The prevention of coal mine accidents, especially gas explosions, and the protection of coal mine workers’ health have become key issues for the sustainable development of the coal mining industry and social property. Despite extensive engineering studies on safe production, there are few studies elucidating the pathogenic process following coal mine gas explosions.

Air-containing organs are most susceptible to shockwaves and high pressure⁶. As the lung exchanges air, it is particularly susceptible to blast injuries due to its direct exposure to external air⁷. Over 60% of civilian victims of conflicts experience blast lung injury (BLI), which accounts for 12.8-33% of military casualties^7,8. During BLI, systemic immune modulation plays a crucial role. Studies have evidenced that leukocytes infiltration and the expression levels of proinflammatory cytokines, including tumor necrosis factor-α, interleukin-6, and interleukin-1β significantly increased post-exposure to blast waves^9,10,11,12. Polymorphonuclear leukocytes in peripheral blood increase immediately after BLI, resulting in elevated systemic inflammation¹³. BLI is also associated with biological processes associated with inflammatory responses, including oxidative stress¹⁴, apoptosis^15,16, autophagy^16,17, and pyroptosis¹⁶. Inflammatory responses have also been reported to be regulated by T regulatory cells and macrophages during blast injury^18,19. These studies have emphasized that immune cells contribute to blast injuries and the recovery of blast-induced injuries. However, our comprehensive understanding of the systemic immune regulation in response to blast injuries remains limited.

Innate immunity relies heavily on neutrophils, which are types of phagocytes, to protect the host against potential pathogens²⁰. As well as maintaining immune homeostasis, they also play a role in the development of acute lung injury following major trauma or sepsis, chronic obstructive pulmonary disease, and inflammatory arthritis^21,22. Dysregulated neutrophil migration, activation, and survival are associated with these clinical conditions²³. Thus, restoring neutrophil homeostasis is essential for recovery from these diseases. Neutrophils originate from hematopoietic stem cells (HSCs) located within the bone marrow (BM). There is a large proportion of polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) and monocytic MDSCs (M-MDSCs) in the BM, which are progenitors of macrophages and granulocytes, respectively²⁴. Some extramedullary tissues, such as the spleen, can also produce neutrophils by their progenitors. Extramedullary organs undergo emergency granulopoiesis due to high demand for new neutrophils in a short period²⁵. Despite extensive studies of neutrophils in lung injury, their contributions to BLI and the dynamics of relevant immune cells in the BM and extramedullary organs during BLI remain to be determined.

Flow cytometry is widely used for immune phenotyping, but its limited detection channels and cumbersome compensation make it difficult to identify all immune cell types simultaneously²⁶. Mass cytometry, a cutting-edge technology combining inductively coupled plasma mass spectrometry and flow cytometry, permits up to 50 parameters to be simultaneously measured in a single cell with minimal or no compensation^27,28. This technology allows for accurate quantification of surface and intracellular proteins, as well as in-depth analysis of systemic immunity^29,30,31. It has been used to understand the complex processes in immune cellular development³², differentiation³³, and tumor immunology^34,35.

Considering the complex nature of systemic immunity, we employed mass cytometry to simultaneously evaluate 16 markers for immune phenotyping and 7 critical proteins involved in the activation of eight signaling pathways, including cell cycle, Jak/Stat, NF-κB, MAPK/ERK1/2, MAPK/p38, CREB, and mTOR pathways, in a mouse model that suffered from a gas explosion. After in-depth analyses of more than five million cells collected from the lung, blood, spleen, and BM, the changes of systemic immunity and signaling pathways in a variety of immune lineages in response to blast and during recovery from blast-induced injury were thoroughly elucidated. These data revealed dynamic changes in neutrophils, macrophages, T cells, and other immune cell populations post-blast injury, along with the involvement of key signaling pathways, underlying the pathogenesis and recovery mechanisms of BLI. Collectively, our work highlighted significant alterations in immune cells and pathways associated with BLI across tissues at single-cell resolution. These findings have important implications for understanding the systemic immune dysregulation in response to BLI and provide valuable insights for potential immune interventions for the disease.

Gas explosion disturbs the immune environment in the lungs in the acute phase

To understand the potential mechanisms underlying lung injury induced by explosion, we have established a mouse model that is able to accurately mimic real gas explosions. The experimental workflow and approaches for the present study are shown in Fig. 1a. After the explosion, tissues of peripheral blood, lung, spleen, and BM at designated time points were collected. Single cells from these tissues were stained with a cocktail of metal isotope-conjugated antibodies against 16 surface antigens and 7 intracellular proteins, and thereafter used for mass cytometry analysis (Fig. 1b). At 6 h post-gas explosion, all mice in the control group and placed at 2.2 meters (m), as well as one subgroup of mice at 1.6 m were killed for tissue collection and the lungs were subjected to hematoxylin and eosin (H&E) staining (Fig. 2a). Compared to control mice, severe pathological changes characterized by alveolar collapse and widened septa were observed in mice placed at 1.6 m by microscopic examination, but lung injuries were less obvious in those at 2.2 m (Fig. 2a). Quantitative analysis of lung pathology further confirmed the occurrence of blast-induced lung injuries and bestowed significantly higher scores in mice 1.6 m than those 2.2 m from explosion site (Fig. 2b). These results suggested the successful establishment of BLI in mice models.

a Experimental workflow and approaches used in this study. Created with BioRender.com/c02g856. b Markers and proteins used to characterize the immune and intracellular pathway atlas.

Full size image

Mice placed at 1.6 or 2.2 m from the explosion point were subjected to a gas explosion. After 6 h, the lung tissues were collected for H&E staining and mass cytometry-based single-cell analysis. Mice without explosion exposure were used as controls. a Representative images of H&E staining of the lungs 6 h post-explosion exposure (Scale bars: 200 µm). b Pathology quantification with lung injury (LI) score. c viSNE maps of CD45⁺ cells in the lung colored by the normalized expression of specific surface markers. d A heatmap showing the normalized median expression of 10 indicated markers in 13 CD45⁺ cell populations in the lung. e viSNE maps colored by 14 cell populations of CD45⁺ cells in the lung after clustering. f A heatmap showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the lung and the LI score. g, h Bar plots showing the frequencies of indicated populations in lung CD45⁺ cells of control mice (n = 10) and blast-treated mice at 1.6 (n = 10) or 2.2 (n = 10) m from the explosion site. i A heatmap showing the normalized expression of T cell markers in the indicated T cell SPADE clusters. j Bar plots showing the frequencies of the indicated T cell SPADE cluster in lung T cells of control mice and blast-treated mice 1.6 or 2.2 m from the explosion site. DNT double negative T, Mono monocyte, AMφ alveolar macrophage, iMφ interstitial macrophage. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

After the explosion, the total CD45-positive cells in the lung tissues increased, suggesting leukocytes infiltration to the lung (Supplementary Fig. 1a). To profile the immune response in the lungs after the gas explosion, we prepared single cells of the lungs and conducted mass cytometry analysis. The expression of 16 surface markers and two nuclear antigens in lung cells were simultaneously measured and analyzed (Fig. 1b and Supplementary Fig. 1b). We performed visualization of t-distributed stochastic neighbor embedding (viSNE) analysis to visualize high-dimensional data in two dimensions. Based on the expression of 10 surface markers in lung CD45⁺ cells, 14 cell populations were identified on the viSNE maps (Fig. 2c–e). Among them, eight populations, including CD4 T cells, CD8 T cells, neutrophils, B cells, resident monocytes (rMono), CD11b⁺ dendritic cells (DCs), CD11b^- DCs and eosinophils, were strongly and significantly correlated with lung injury (Fig. 2f). The proportion of neutrophils was significantly increased in the lungs after blast (Fig. 2g), while all the other populations, except for alveolar and interstitial macrophages and inflammatory monocytes (iMono), decreased in mice 1.6 m away from explosion site (Fig. 2g, h and Supplementary Fig. 1c). In mice 2.2 m away from explosion site, the proportions of CD4 T cells and total T cells were not affected by blast compared to the controls. Next, we used spanning-tree progression analysis of density-normalized events (SPADE) analysis to systematically assess changes in T cells. 50 minor lung T cell clusters (nodes) containing cells with similar phenotypes were identified (Supplementary Fig. 1d). Ten T cell clusters, including five CD8 T clusters, one CD4 T cell cluster, one Th17 cluster, and three double-negative (DN) T cell clusters, were strongly and significantly correlated with lung injury (Fig. 2i, Supplementary Fig. 1e). Among these clusters, the percentages of clusters 24, 14, 9, 47, and 35 in lung T cells decreased while clusters 23, 44, 4, 40, and 36 increased after 6 h of blast (Fig. 2j). Although gas explosion resulted in the same immune cell populations associated with BLI in mice positioned at both distances, the T cell components in mice at 2.2 m underwent milder changes than at 1.6 m, while distribution patterns of other influenced immune cells were similar. These results indicated that at the early stage of explosion-induced BLI, neutrophils quickly accumulated in conjunction with significant changes in T cell subsets in the lung tissues. Together, these findings demonstrate that our mass cytometry analysis captures differences in immune cell responses in the acute phase of explosion-induced BLI.

Granulocytes, monocytes, and Ly6C^high macrophages in peripheral blood are involved in the acute immune response to blast

Immune cells in the peripheral blood play a crucial role in the development of BLI. Thus, we systemically investigated the immune landscape in peripheral blood in response to blast. Using the viSNE analysis, we identified 13 immune cell populations (Fig. 3a and Supplementary Fig. 2a, b). The proportions of several cell populations, including granulocytes, monocytes, B cells, DCs and Ly6C^high macrophages exhibited a strong and significant correlation with BLI (Fig. 3b). The percentage of granulocytes, monocytes, and Ly6C^high macrophages in the peripheral blood significantly increased post-exposure to blast, while the proportion of CD4 T cells, CD8 T cells, DCs, and Ly6C^low macrophages decreased (Fig. 3c, Supplementary Fig. 2c). Notably, the proportions of CD11c⁺ macrophages, Ly6C^low macrophages and total macrophages in mice at 2.2 m decreased to a much lower level compared to mice in the control group and at 1.6 m (Supplementary Fig. 2c). In addition, our results also demonstrated a strong and significant correlation between BLI and the expression of phosphorylated cyclic-AMP response element-binding protein (pCREB) and phosphorylated retinoblastoma protein (pRb) in B cells, DCs, granulocytes and Ly6C^high macrophages, as well as the expressions of p-p38 in granulocytes and pS6 in B cells. The expression levels of these markers in peripheral blood immune cells were significantly dysregulated by exposure to explosion (Fig. 3d and Supplementary Fig. 3a). Furthermore, we employed SPADE analysis to identify the detailed T cell clusters involved in blast-induced immune response. Among 50 T cell clusters, 13 clusters were strongly and significantly correlated to BLI (Fig. 3e and Supplementary Fig. 3b). The proportions of these 13 T cell clusters with distinct phenotypes were significantly dysregulated by exposure to explosion (Fig. 3f and Supplementary Fig. 3c). These data provide a comprehensive understanding of the immune landscape in peripheral blood in the acute phase after blast and demonstrate the early involvement of monocytes, and Ly6C^high macrophages in BLI.

The peripheral white blood (PWB) was collected from control mice or mice experiencing explosion for mass cytometry-based single-cell analysis. a viSNE maps colored by 13 cell populations of the PWB cells after clustering. b A heatmap showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the PWB cells and the LI score. c Bar plots showing the frequencies of indicated populations in the PWB cells of control mice (n = 10) and blast-treated mice 1.6 (n = 10) or 2.2 (n = 10) m from the explosion site. d A heatmap showing the normalized expression of significantly changed intracellular proteins in the indicated PWB cell populations. e SPADE tree describing 50 minor peripheral blood T cell clusters of one representative mouse colored by the median expression of the indicated markers. f Bar plots showing the frequencies of the indicated T cell SPADE cluster in peripheral blood T cells of control mice and blast-treated mice 1.6 or 2.2 m from the explosion site. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

Identification of immune response in the spleen in the acute phase after blast

Given that the spleen and BM are key immune organs and primary sources of immune cells, we investigated the effects of the explosion on immune cells within these organs using mass cytometry. As revealed by mass cytometry-based single-cell analysis, the immune environment in the spleens was also changed in the acute phase after blast. viSNE analysis identified 14 immune cell populations (Fig. 4a and Supplementary Fig. 4a, b). The percentage of seven cell populations, including PMN-MDSCs and M-MDSCs, CD4 T cells, CD8 T cells, B cells, Ly6C^high and Ly6C^low macrophages, displayed strong and significant correlation to lung injury (Fig. 4b). The proportions of PMN-MDSCs and M-MDSCs were significantly increased in mice placed 1.6 or 2.2 m from explosion site, while the percentages of B cells, Ly6C^high and Ly6C^low macrophages were decreased (Fig. 4c, d and Supplementary Fig. 4c). In terms of T cells, the percentages of TCRβ-T cells, and DN T cells were increased only in mice positioned 1.6 m from explosion site (Fig. 4c and Supplementary Fig. 4c). The other cell populations in the spleens were not significantly affected by blast (Supplementary Fig. 4c). Furthermore, our analysis revealed that several intracellular signaling pathways in the immune cells of the spleens were regulated after lung injury (Fig. 4e). The phosphorylation of p38 in B cells, CD4 T cells, and CD8 T cells, IκBα expression in CD4 T cells and CD8 T cells, and pS6 in CD8 T cells were significantly dysregulated by blast (Supplementary Fig. 5a). Based on SPADE analysis (Fig. 4f), ten T cell clusters were significantly and strongly correlated to lung injury (Supplementary Fig. 5b). These T cell clusters, characterized by distinct expression of surface and intracellular proteins were significantly dysregulated by blast. Of note, T cell cluster 22, which was identified as Treg cells expressing Foxp3, was increased in the spleens by blast (Fig. 4g and Supplementary Fig. 5c). Overall, these data indicate that the immune cell responses in the spleen are significantly perturbed during the acute phase after explosion and demonstrate the potential relevance to BLI.

The SP cells collected from control mice or mice experienced an explosion were used for mass cytometry-based analysis. a viSNE maps colored by 13 cell populations of the SP CD45⁺ cells after clustering. b A heatmap showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the SP CD45⁺ cells and the LI score. c, d Bar plots showing the frequencies of indicated populations in the SP CD45⁺ cells of control mice (n = 10) and blast-treated mice at 1.6 (n = 10) or 2.2 (n = 10) m from the explosion site. e A heatmap showing the normalized expression of significantly changed intracellular proteins in the indicated SP CD45⁺ cell populations. f SPADE tree describing 50 minor SP T cell clusters of one representative mouse colored by the median expression of the indicated markers. g Bar plots showing the frequencies of the indicated T cell SPADE cluster in SP T cells of control mice and blast-treated mice 1.6 or 2.2 m from the blast site. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

Immune cells in the BM are perturbed in the acute phase after blast

BM is the source of all immune cells. To evaluate the immunological changes in the BM, we prepared single cells of the BM and conducted mass cytometry. Through viSNE analysis, we identified 14 immune cell populations in the BM (Fig. 5a and Supplementary Fig. 6a, b). Different from the significant changes in immune cells in other tissues, the immune cells in the BM were not widely affected by blast in the acute phase. Whereas, the proportions of M-MDSCs, CD4 T cells, CD8 T cells and total T cells were strongly and significantly correlated to the severity of lung injury (Fig. 5b–e). Specifically, the proportions of CD4 T cells, CD8 T cells and total T cells were significantly increased, while the percentage of CD19⁺ DCs was decreased in mice located 1.6 m from explosion site (Fig. 5c–e). All the other populations were slightly or not affected by blast (Fig. 5f and Supplementary Fig. 6c).

The BM cells collected from control mice or mice experienced an explosion were used for mass cytometry-based analysis. a viSNE maps colored by 13 cell populations of the BM CD45⁺ cells after clustering. b–e Dot plots (n = 30) showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the BM CD45⁺ cells and the LI score (left). Bar plots showing the frequencies of indicated populations in the BM CD45⁺ cells of control mice (n = 10) and blast-treated mice 1.6 (n = 10) or 2.2 (n = 10) m from the explosion site (right). f Bar plots showing the frequencies of indicated populations in the BM CD45⁺ cells of control mice and blast-treated mice 1.6 or 2.2 m from the explosion site. g Bar plots showing the normalized expression of significantly changed intracellular proteins in the indicated BM CD45⁺ cell populations. h Radar plots showing the average frequencies of the indicated populations in the lung, SP, BM or PWB cells of control mice and blast-treated mice 1.6 or 2.2 m from the blast site. i PCA and corresponding vector plot of contributions for immune cell frequencies in the lung and LI score from control mice and blast-treated mice 1.6 or 2.2 m from the blast site. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

The phosphorylation of STAT1 in CD4 T cells, pCREB in CD8 T cells and DN T cells, and IκBα and pERK1/2 in PMN-MDSCs were found to be strongly and significantly correlated to lung injury, and these proteins were dysregulated by blast (Fig. 5g). These results suggest that the immune cells in the BM have less relevance to the immune response in the acute phase after blast and BM is not the main source of immune cells migrating to the lung.

Next, we further analyzed the systemic immune response to blast. At the early stage after blast, we observed a significant increase in granulocytes, especially neutrophils, in both peripheral blood and lung tissues. Additionally, the composition of granulocyte precursor cells (PMN-MDSCs) in the spleens was also markedly elevated, which was not observed for the precursor cells (PMN-MDSCs and M-MDSCs) in the BM (Fig. 5h). As a result, the proportion of B cells was decreased in peripheral blood, lungs, and spleens. DCs were decreased in the lungs and peripheral blood as well (Fig. 5h). Furthermore, our results demonstrated that T cells, macrophages, and monocytes in different tissues were also dysregulated by blast (Supplementary Fig. 6d). Remarkably, our principal component analysis showed that neutrophils in the lungs were highly correlated with BLI, implicating the vital role of neutrophils in the progression of BLI at the early stage (Fig. 5i).

Immune landscape in lung tissues during blast injury repair

Inflammation and cytokine storm are major causes of acute lung injury, whereupon, in the recovery phase, immune cells are critically involved in lung repair and injury resolution. To elucidate the immune landscape during BLI repair, we employed mass cytometry and assessed the dynamics of immune cells. Mice placed at 1.6 m from explosion site were euthanized and tissues of the lungs, spleens, BM, and peripheral blood were collected for H&E staining or mass cytometry analysis (Fig. 1a). H&E staining revealed visible injuries in the lungs of mice that experienced an explosion (Fig. 6a). Compared to control mice (day 0), the alveolar septa were widened and the alveoli cavities were fused with reduced individual volume in the lung tissues from mice exposed to a gas explosion. Moreover, degrees of variable inflammatory cells infiltration were observed in the lungs of model mice. Infiltration of inflammatory cells peaked on day 1 post-exposure to a gas explosion. This observation was further supported by pathological scores of microscopic changes due to the highest scores for lungs on day 1 and a decrease on days 3 and 7 (Supplementary Fig. 7a). Moreover, the immune cells infiltration gradually decreased over time, commencing three days after the explosion. (Supplementary Fig. 7b). These results suggest that our present mice model of BLI encompassing the acute and recovery phase of BLI and lung repair starts from at least day 3 following initial injury.

Mice placed at 1.6 m from the explosion point were subjected to a gas explosion. After the indicated days, the lung tissues were collected for hematoxylin-eosin (H&E) staining and mass cytometry-based analysis. Mice without explosion exposure at day 0 were used as controls. a Representative images of H&E staining of the lungs at indicated time post-explosion exposure (Scale bars: 200 µm). b Stacked bar plot showing the average frequencies of the indicated cell populations in the lungs of controls (day 0) and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. c A heatmap showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the lung and the lung injury (LI) score. d, e Bar plots showing the frequencies of indicated populations in lung CD45⁺ cells of controls (n = 10) and mouse models killed on Day(s) 1 (n = 9), 3 (n = 10), and 7 (n = 10) post-explosion exposure. f SPADE tree describing 50 minor lung T cell clusters of one representative mouse colored by the median expression of the indicated markers. g, h Bar plots showing the frequencies of the indicated T cell SPADE cluster in lung T cells of controls and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Day 0 (C) or D0 (C), Day 0 (control).

Full size image

After analyzing the immune cells in the lungs, we found that in the 7 days following the explosion, the proportion of immune cells displayed a regular change (Fig. 6b and Supplementary Fig. 7c–e). Correlation analysis demonstrated strong and significant correlations between lung injury during repair and nine immune cell populations (Fig. 6c). Neutrophils, inflammatory monocytes and alveolar macrophages accumulated in the lungs within 1 day following blast (Fig. 6d, e). Interestingly, the proportions of most immune cells returned to a state similar to the control group at 3 days after blast (Fig. 6d, e and Supplementary Fig. 7f). However, inflammatory monocytes remained elevated significantly till to 7th day after explosion (Fig. 6e). Additionally, our analysis also distinguished 12 lung T cell SPADE clusters that were strongly and significantly correlated with lung injury during repair (Fig. 6f and Supplementary Fig. 7g, h). Among these clusters, the proportions of most T cell clusters, including clusters 26, 31, 47, 35, 42, 22, 12, 21, and 30, changed on day 1 but returned to their pre-blast state 3 days after explosion (Fig. 6g, h). Clusters 37 and 32 remained decreased while cluster 43 remained elevated up to the 7th day after the explosion (Fig. 6h). All these data indicate that neutrophils, inflammatory monocytes and alveolar macrophages participate in the early stage of lung repair after BLI, and inflammatory monocytes continue to be involved in the lung repair process.

Immune landscape in peripheral blood during blast injury repair

Immune landscape in peripheral blood during blast injury repair was also analyzed using viSNE analysis to identify immune cell populations (Supplementary Fig. 8a–c). Immune cells in peripheral blood were also changed during blast injury repair (Fig. 7a). However, only granulocytes in peripheral blood exhibited a strong and significant correlation with lung injury during repair (Fig.7b). The proportion of granulocytes was significantly elevated at day 1 after blast, and returned to pre-blast levels (control, day 0) within 3 days (Fig. 7b). Although most dysregulated immune cells returned to their pre-blast levels within 3 days, the percentages of Ly6C^high and CD11c⁺ macrophages didn’t restore until day 7 after blast (Fig. 7c, d, Supplementary Fig. 8d). Furthermore, multiple intracellular proteins in distinct immune cell types, including pRb in B cells, CD4 T cells, CD8 T cells, monocytes, CD19⁺ DCs, and DCs, and pCREB in Ly6C^high macrophages and monocytes, were strongly and significantly correlated to lung injury during repair. The levels of these proteins were decreased at day 1 and returned to the pre-blast state 3 days after the blast (Fig. 7e). Additionally, we identified 12 peripheral blood T cell SPADE clusters strongly and significantly correlated to lung injury during repair (Fig. 7f and Supplementary Fig. 8e). The proportion of all these T cell clusters, including two Treg clusters 23 and 20, was changed on day 1 and returned to pre-blast state on day 3 after blast (Fig. 7g, h and Supplementary Fig. 8f). These data reveal that immune cells in peripheral blood, particularly granulocytes, are involved in lung injury resolution mainly at the early stage of lung repair. The recovery of Ly6C^high macrophages in peripheral blood appears later than that of other immune cells.

The peripheral blood collected from mice experienced an explosion for the indicated days was used for mass cytometry-based single-cell analysis. a Stacked bar plot showing the average frequencies of the indicated cell populations in the peripheral blood of controls and mouse models killed on Day(s) 1, 3, and 7 post-explosion exposure. b Dot plots (n = 37) showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the PWB cells and the lung injury (LI) score (left). Bar plots showing the frequencies of indicated populations in the PWB of controls (day 0, n = 9) and mouse models killed on day(s) 1 (n = 8), 3 (n = 10), and 7 (n = 10) post-explosion exposure (right). c and d Bar plots showing the frequencies of indicated populations in the PWB of controls and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. e Bar plots showing the normalized expression of significantly changed intracellular proteins in the indicated PWB cell populations. f SPADE tree describing 50 minor peripheral blood T cell clusters of one representative mouse colored by the median expression of the indicated markers. g and h Bar plots showing the frequencies of the indicated T cell SPADE cluster in peripheral blood T cells of controls and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

Immune landscape in the spleens during blast injury repair

In the spleens, there were slight changes in immune cell populations during the repair of blast injuries (Fig. 8a and Supplementary Fig. 9a–c). The proportion of PMN-MDSCs, total MDSCs and total T cells were strongly and significantly correlated with lung injury during repair (Fig. 8b, c). MDSCs, especially PMN-MDSCs, were significantly and greatly elevated on day 1 after blast, and returned to pre-blast state (control, day 0) 3 days after blast (Fig. 8b). The proportion of T cells, especially CD8 T cells, decreased on day 1 after the blast and returned to their pre-blast state 7 days after blast (Fig. 8c, d). The other immune cell populations were not clearly altered during blast injury repair (Supplementary Fig. 9d). However, intracellular pathways in multiple immune cells were dysregulated. The expression of IκBα in CD4 T cells, CD8 T cells, B cells, DCs, and Ly6C^low macrophages, as well as p-p38 in CD4 T cells, CD8 T cells and B cells, were either up- or down-regulated on day 1 and returned to pre-blast state 3 days after blast (Fig. 8e). Additionally, eight T cell SPADE clusters were also dysregulated during blast injury repair (Fig. 8f, g and Supplementary Fig. 9e). These results suggest that the spleen is also involved in the early stage of the repair process following blast injuries, but are less significant in the later stage.

The SP cells collected from mice experienced an explosion for the indicated days were used for mass cytometry-based analysis. a Stacked bar plot showing the average frequencies of the indicated cell populations in the SP CD45⁺ cells of controls and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. b, c Dot plots (n = 37) showing the Pearson correlation coefficients for relationships between the proportions of the indicated cell populations in the SP CD45⁺ cells and the LI score (left). Bar plots showing the frequencies of indicated populations in the SP CD45⁺ cells of controls (day 0, n = 10) and mouse models killed on day(s) 1 (n = 9), 3 (n = 10), and 7 (n = 10) post-explosion exposure (right). d Bar plots showing the frequencies of indicated populations in the SP CD45⁺ cells of controls and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. e Bar plots showing the normalized expression of significantly changed intracellular proteins in the indicated SP CD45⁺ cell populations. f SPADE tree describing 50 minor SP T cell clusters of one representative mouse colored by the median expression of the indicated markers. g Bar plots showing the frequencies of the indicated T cell SPADE cluster in SP T cells of controls and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

Immune landscape in the BM during blast injury repair

After the identification of the immune atlas in the BM cells using viSNE, regular changes different from those in other tissues were observed (Fig. 9a and Supplementary Fig. 10a–c). The proportions of PMN-MDSCs and Ly6C^high macrophages were not changed on day 1, but significantly increased on days 3 and 7 after the blast (Fig. 9b, c). CD4 T cells were elevated on day 1 and returned to their pre-blast state (control, day 0) 3 days after blast (Fig. 9c). As a result of increased PMN-MDSCs, the percentages of CD8 T cells, DN T cells, TCRβ-T cells, B cells, and CD19⁺ DCs were decreased on days 3 and 7 after blast (Fig. 9b, c and Supplementary Fig. 10d). The rest of the cell populations were not changed after blast (Supplementary Fig. 10e). The phosphorylation of CREB in CD4 T cells, p-p38 in CD8 T cells, M-MDSC and B cells, pRb in CD4 T cells and CD8 T cells, PMN-MDSCs, M-MDSCs and B cells, and IκBα in PMN-MDSCs were significantly dysregulated during blast injury repair (Fig. 9d and Supplementary Fig. 10f). These data clearly demonstrate that the immune cells, especially PMN-MDSCs and Ly6C^high macrophages, in the BM are mainly involved in the late stage of the repair process following blast injuries.

The BM cells collected from mice experienced an explosion for the indicated days were used for mass cytometry-based analysis. a Stacked bar plot showing the average frequencies of the indicated cell populations in the BM CD45⁺ cells of controls (day 0) and mouse models killed on day(s) 1, 3, and 7 post-explosion exposure. b, c Bar plots showing the frequencies of indicated populations in the BM CD45⁺ cells of controls (n = 10) and mouse models killed on day(s) 1 (n = 9), 3 (n = 10), and 7 (n = 10) post-explosion exposure. d Bar plots showing the normalized expression of significantly changed intracellular proteins in the indicated BM CD45⁺ cell populations. e Schematic diagram showing the pattern of immune response to blast within 7 days. Created with BioRender.com/m14n514. Error bar, mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

Globally, explosions are caused by military conflicts, terrorist acts, industrial accidents, and civilian gas explosions⁶. Due to its substantial air-tissue surface area, the lung is particularly vulnerable to blast injuries during explosions, and BLI has a profound impact on survivors’ quality of life and survival. A blast causes acute lung injuries that activate the immune system through the damage-associated molecular patterns (DAMPs) system and cause a severe local inflammatory response³⁶. During this process, the immune system plays a crucial role in restoring pulmonary homeostasis and repairing damaged lungs. In this project, we developed a mouse model of gas explosion and used high-throughput single-cell proteomics to investigate how systemic immunity responds to BLI throughout the pathological process (Fig. 9e). Our findings strengthen our understanding of the important role of systemic immunity both during the acute and recovery phases of BLI, as well as providing new potential strategies for improving BLI repair.

In our study, we discovered that blast triggered a similar process of granulopoiesis in a short period as infection. As a result of blast, neutrophils migrated to the lungs and accumulated there, and peripheral blood granulocytes increased significantly. In addition, PMN-MDSCs, the progenitor of granulocytes, are increased in the spleen, but not in the BM, suggesting extramedullary involvement in granulopoiesis at the early stage post-blast. A serious tissue injury can release DAMPs from necrotic cells, which serve as proinflammatory non-microbial stimuli²¹. DAMPs bind to pattern recognition receptors (PRRs) on neutrophils and cause inflammation. Thus, early inflammation in the lungs after blast is primarily induced through the DAMP-PRR system. Interestingly, we observed that neutrophils in the lung and granulocytes in the peripheral blood returned to pre-blast levels within 3 days after blast, while BM PMN-MDSCs remained significantly elevated, suggesting granulopoiesis in the BM during recovery.

In addition to neutrophils and granulocytes, other types of immune cells are also dysregulated after blast. In spite of no obvious changes in monocytes in the lungs following blast exposure, their proportion in peripheral blood increased significantly, indicating that monocytes played a role in the early immune response to BLI. Previous studies have reported that classical monocytes, alongside neutrophils, participate in the cellular immune response to primary BLI in a rat model¹⁰. Monocytes in peripheral blood alter only during the acute phase after blast and return to normal one day later. The progenitors of monocytes, M-MDSCs, are also increased by blast at the early stage in the spleen, suggesting the involvement of myelopoiesis in the acute phase. Damaged tissues and dead cells are primarily eliminated by phagocytes, especially monocytes and macrophages³⁷. Monocytes migrate to damaged tissues or inflammatory sites and differentiate into macrophages removing dead and dying cells³⁸. In our study, the proportion of neutrophils in the lungs and blood, as well as PMN-MDSCs in the spleen, returned to normal levels 3 days after blast, while the proportion of monocytes, especially inflammatory monocytes, remained elevated in the lungs, implicating their importance in the repair process of BLI.

Increasing granulopoiesis or neutrophil migration from circulating and marginated pools is responsible for changes in the proportion of other immune cells, such as lymphocytes. These results are consistent with a previous study which has shown a decrease in T cells caused by increased neutrophils in the blood after blast³⁹. Lymphocyte-mediated adaptive immune response typically begins 96 h after antigen stimulation. We found significant dysregulation of T cells in the BM, including CD4 T cells and CD8 T cells, but not PMN-MDSCs, 6 h after blast exposure, suggesting their early role in the immune response to blast. In the present study, we observed dysregulation of intracellular pathways in T cells, as well as T cell components, frequently following blast exposure, despite lymphocytes commencing the immune response much later than innate immune cells. A new insight into the understanding of systemic immune response to blast injuries comes from the involvement of adaptive immune cells in the early immune response.

In the present study, we systemically demonstrated the effects of blast on intracellular pathways in innate and adaptive immune cells. We found that phosphorylation of pRb, which controls G1-S phase transitions and determines cell cycle entry⁴⁰, was reduced in most immune cell types after blast, implicating a decrease in their proliferation. After returning to normal pRb levels, their proliferation may return to normal within 3 days. The NF-κB signaling pathway regulates innate and adaptive immune functions by targeting multiple genes involved in cell survival, proliferation, inflammatory response, and cell adhesion. IκB inhibits NF-κB-targeted gene transcription by preventing the translocation of NF-κB proteins to the nucleus⁴¹. Blast dysregulated the levels of IκB in CD4 T cells, CD8 T cells, B cells, DCs, Ly6C^high macrophages, and PMN-MDSCs, and thus influenced their functions. As a transcription factor, CREB interacts with the NF-κB pathway and regulates immune response, such as promoting the survival of macrophages, B cells, and T cells⁴². Multiple immune cells were dysregulated in CREB activation by blast, resulting in changes in their functions. S6 is phosphorylated by S6K1, a key regulator of the mTOR pathway, and regulates protein synthesis, cell differentiation, and immune function⁴³. Thus, blast also affects the functions of B cells and CD8 T cells by regulating S6 protein phosphorylation. ERK1/2 and p38 are critical serine-threonine protein kinases in the MAPK signaling cascade, one of the most critical pathways for cell proliferation, differentiation, and survival. The binding of phosphorylated ERK1/2 and p38 to AP-1 regulates gene transcription and contributes to inflammatory diseases⁴⁴. Blast significantly dysregulated the activation of ERK1/2 and p38 in B cells, CD4 T cells, CD8 T cells, M-MDSC, and granulocytes, suggesting that blast modulates their function. Although multiple signaling molecules in immune cells were significantly altered in vivo post-blast exposure, the exact nature of these effects and specific mechanisms underlying immune regulation remain to be thoroughly clarified in the future.

A detailed atlas of the immune response to BLI is presented in this study by fine-grained analysis of the systemic immune system in a mouse model. The present study identifies the intracellular pathways relevant to BLI in immune cells and reveals immune recovery patterns during BLI repair. These data will be a valuable resource for future research to explore more efficient strategies tailored to improve BLI treatment through systemic immunity regulation. It is important to note that our results not only validate previous discoveries but also provide significant empirical evidence and insights that will lead to a better understanding of blast-induced systemic regulation.

Animals

Eight- to nine-week-old male C57BL/6 J mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). After one week of adaptive feeding under the experimental conditions, mice were subjected to a gas explosion for model establishment of explosion-induced injury as described below. All animal experiments were approved by the Medical Animal Care and Welfare Committee of Xinxiang Medical University (XYLL-20220531) and carried out in compliance with the approved institutional and national guidelines as well as guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes and NIH. We have complied with all relevant ethical regulations for animal use.

Mouse model of explosion-induced injury

The mouse explosion-induced injury model was established with a self-made adjustable shock tube simulation device consisting of five major parts: a remote lighter, vacuum pump, signal detector of pressure and temperature, tube body and data analysis system, according to a previously published device with direction adjusted to be horizontal⁴⁵. The tube body was composed of three segments of tubes with an equal length of 1.2 m and a shorter segment of 0.8 m. These segments can be connected and sealed to one. The shorter segment served as the explosion chamber while the others were used to hold mice (Fig. 1a). In the present study, mice were randomly divided into four major groups: control group, and gas explosion groups at distances of 1.6 or 2.2 m. The explosion group at 1.6 m was further divided into four subgroups. After administering anesthesia, mice were fixed on a wire mesh in the animal holder in a standing position facing the explosion chamber. The explosion chamber and animal holder were connected and sealed but separated with a transparent polyethylene plastic sheet. The explosion chamber was filled with a mixture consisting of 10% methane and 90% normal air in a total volume of 30 liters and stood for 5 min after inflation. Finally, all valves were closed and the mice were then subjected to a gas explosion achieved by remote ignition action. The signal detector of pressure and temperature monitored the pressure and temperature during the explosive course (Maximum pressure: 0.42 ± 0.02 MPa; Maximum temperature: 788.95 ± 54.34 °C). Control mice and animals that experienced a gas explosion at 2.2 m were euthanized at 6 h post-explosion. Blood, lung, BM, and spleen were dissected for single-cell suspension preparation. Mice from the explosion group at 1.6 m were euthanized for tissue collection at different time points: 6 h, 1 day, 3 days, and 7 days post-explosion.

Blood cells preparation

Blood was collected from the orbital sinus into 1.5 ml tubes pre-treated with EDTA disodium salt dihydrate (Sangon Biotech, Shanghai, China) and then subjected to centrifugation. The cell pellet was resuspended with cold phosphate-buffered saline (PBS) (Sangon Biotech) and fixed with Fix I buffer (Standard BioTools, CA, USA) at room temperature (RT) for 10 min. After washing with cold PBS, the red blood cells were lysed with red blood cell lysis buffer to remove residual erythrocytes, centrifuged again, resuspended in FACS buffer supplemented with 10% dimethyl sulfoxide (DMSO) and stored in liquid nitrogen.

Single cell preparation for the lung

The lung tissue was dissected and washed twice with cold PBS to remove residual red blood cells. Approximately 30 mg of the right lobe was excised, minced and placed in 12-well plates with tissue digestion buffer pre-heated to 37 °C, containing 1 mg ml^-1 collagenase Type I (Sigma-Aldrich, MO, USA) and 300 U ml^-1 DNase I (Beyotime, Shanghai, China) in RPMI 1640 medium (Gibco, MA, USA). Minced lung slices in plates were incubated at 37 °C for 1.5-2 h with gentle shakes at an interval of 5 min. Digestion was terminated by transferring plates onto ice after complete dissociation. Cells were gently dispersed, filtered through a 200-mesh strainer and centrifuged at 300 g and 4°C. Cells were resuspended with 1 ml of PBS and fixed with Fix I buffer (Standard BioTools) for 10 min at RT followed by washing with cold PBS. The obtained cells were lysed with red blood cell lysis buffer to remove residual erythrocytes, centrifuged again, resuspended in FACS buffer supplemented with 10% DMSO and stored in liquid nitrogen.

BM cells preparation

The femur and tibiae were dissected and carefully cleaned by removing the unwanted muscles and washing them with cold PBS. The femur and tibiae were cut open on both sides and BM cells were immediately flushed into one well of pre-chilled 6-well plates using a 1 ml syringe filled with pre-chilled RPMI 1640 medium (Gibco) until all BM was recovered. Then, cell suspensions were filtered into 15 ml tubes with a 200-mesh strainer and fixed with Fix I buffer (Standard BioTools) for 10 min at RT. After washing with cold PBS, BM cells were lysed with red blood cell lysis buffer to remove residual erythrocytes, centrifuged again, resuspended in FACS buffer supplemented with 10% DMSO and stored in liquid nitrogen.

Spleen cells preparation

The spleen was dissected and evenly divided into three sections. The middle part was placed into one well of a 6-well plate containing 3 ml RPMI 1640 medium (Gibco). Spleen cells were separated by gently crushing the spleen. Cells were then filtered into 15 ml tubes with a 200-mesh strainer and fixed with Fix I buffer (Standard BioTools) for 10 min at RT. After washing with cold PBS, the obtained cells were lysed with red blood cell lysis buffer to remove residual erythrocytes, centrifuged again, resuspended in FACS buffer supplemented with 10% DMSO and stored in liquid nitrogen.

Histological analysis

Fresh murine lungs were fixed in 10% (vol/vol) neutral buffered formalin for 48 h at RT. After fixation, the lung tissues were dehydrated in serial ethanol solutions of gradient concentrations and xylene and embedded in paraffin. Finally, paraffin sections of 4 μm in thickness were made using a Leica RM2016 slicer (Leica, Wetzlar, Germany). Then, H&E staining was performed as previously described⁴⁶. The lung injury was examined under a light microscope and the pathological changes were quantified according to a semi-quantitative scoring system described by Mrozek and colleagues⁴⁷. Pathological changes including alveolar and interstitial inflammation, alveolar and interstitial hemorrhage, edema, atelectasis, and hyaline membrane formation were evaluated. Each variable was scored on a scale of 0–4 according to the percentage of pathological area relative to whole slide imaging as follows: 0 (no injury observed), 1 (injured area ≤25%), 2 (injured area 25– 50%), 3 (injured area 50–75%), and 4 (injured area >75%). The lung injury score was calculated as the sum of all individual scores.

immunohistochemical staining of CD45 in mouse lung

For immunostaining, paraffinized 4 μm thick sections were dewaxed and rehydrated. After blocking internal peroxidase activity with 3% H₂O₂/methanol, sections were subjected to microwave-assisted high-pressure antigen retrieval in l mM EDTA disodium salt dihydrate buffer for 2.5 min followed by cooling down at RT. Sections were then blocked with 2% bovine serum albumin and incubated at RT with rabbit anti-CD45 antibody (ABclonal, Wuhan, China) and secondary HRP-conjugated goat anti-rabbit antibody (ABclonal, Wuhan, China). Sections were then developed with VECTASTAIN Elite ABC Peroxidase Staining Kit (Vector Laboratories, CA, USA) and DAB Peroxidase Substrate Kit (Vector Laboratories) and finally counterstained with Meyer’s hematoxylin.

Samples preparation for mass cytometry

Single cells obtained from peripheral blood, BM, spleen, or lung tissues were washed with Cell Staining Buffer (CSB, Standard BioTools) and barcoded separately using the Cell-ID^TM 20-Plex Pd Barcoding Kit (Standard BioTools) following the manufacturer’s instructions. After washing with CSB, 20 barcoded samples derived from the same tissues were pooled together, and the rest of the ten samples were pooled together and stained with cisplatin (Standard BioTools). All samples obtained from the same type of organ were mixed together for antibody staining. Samples were blocked with anti-mouse CD16/32 antibody (FcR III/II, BioLegend, CA, USA) to reduce nonspecific antibody binding. These samples were stained with a cocktail of metal isotope-conjugated antibodies against 16 surface antigens (Supplementary Table S1). After washing, these cells were stained with anti-mouse Foxp3-158Gd and anti-mouse RORgt-159Tb using Nuclear Antigen Staining Buffer Set (Standard BioTools). After washing, samples were incubated with a cocktail of metal isotope-conjugated antibodies against 7 intracellular proteins (supplementary Table S1). After antibody staining, samples were resuspended in Fix & Perm Buffer (Standard BioTools) containing Cell-ID^TM Intercalator-Ir (Standard BioTools) overnight at 4 °C. Cells were washed and resuspended in ultrapure water supplemented with 15% EQ Four Element Calibration Beads (Standard BioTools). All stained cells were washed with CSB and cell acquisition solution (Standard BioTools), and then analyzed using a Helios mass cytometer (Standard BioTools).

Mass cytometry data processing and analysis

Mass cytometry data were randomized, normalized, and debarcoded using the CyTOF 6.7 software (Standard BioTools). Debarcoded data were uploaded to and analyzed using Cytobank⁴⁸ with the available algorithms viSNE⁴⁹ and SPADE⁵⁰. All cell debris and doublets were excluded before analysis.

Statistics and reproducibility

One-way ANOVA followed by multiple comparisons was used to compare multiple groups. A parametric test (Student’s t test) was used to determine the statistical significance between two groups if the data is normally distributed. Error bars represent the mean ± standard deviation (SD). P < 0.05 was represented for as statistically significant. Correlation analyses were performed using Person’s correlation coefficient. Heatmaps showing correlations, stacked bar plots, and radar plots were generated on ChiPlot (www.chiplot.online).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.