Blog

IH exacerbates lung remodelling in mice

We exposed C57BL/6J mice to either IH or normoxia for 6 weeks to investigate the effects of IH on lung tissue. After the exposure period, lung tissues were collected for histopathological staining and analysis. Compared with control mice, those exposed to IH exhibited significant increases in pulmonary inflammation, mucus staining, and collagen deposition (Fig. 1A-D). Phalloidin staining revealed enhanced cytoskeletal organization and increased stress fiber formation, which may reflect actin remodeling associated with both structural cells and infiltrating inflammatory cells in the lungs of IH-exposed mice (Fig. 1A, E). Additionally, histochemical staining revealed increased expression of Cd11b, Mpo, and Cd68 in the lungs of IH-exposed mice, indicating that IH exposure promoted infiltration of immune cells. The upregulation of Cd31 expression in lung tissue further confirmed that IH exposure stimulated remodeling of pulmonary blood vessels (Fig. 1F-G). In summary, IH exposure induced pronounced inflammatory infiltration and structural alterations, contributing to extensive pulmonary tissue and vascular remodeling in mice.

Construction of a single-cell transcriptome map of IH-exposed mouse lungs

Next, we retrieved scRNA-seq data from the GEO database, including lung tissue samples from both IH model mice (n = 3) and healthy control mice (n = 3), to gain a comprehensive understanding of the cellular landscape in the lung tissue of IH model mice (Supplementary Table 1). After rigorous quality control and filtering, a total of 48,272 cells were retained from the scRNA-seq data for downstream analysis (Supplementary Fig. 1 and Supplementary Table 1). In the scRNA-seq dataset, the median number of unique molecular identifiers (UMIs) per cell was 1978, and the median number of genes detected per cell was 996. After PCA-mediated reduction of these high-quality data, the first 15 PCs were included in the follow-up analysis (Supplementary Fig. 1). We applied the Harmony algorithm to each dataset again to minimize batch effects between different scRNA-seq datasets. Subsequently, we adopted a clustering method based on t-SNE to initially identify 12 clusters according to typical markers of different cell types. These cell types include type 1 alveolar epithelial cells (AT1), type 2 alveolar epithelial cells (AT2), macrophages/monocytes, megakaryocytes, B cells, T cells, fibroblasts, endothelial cells, eosinophils, natural killer (NK) cells, neutrophils, and erythroid cells (Fig. 2A-B, Supplementary Table 2). These cells were present in each sample (Fig. 2E and Supplementary Fig. 1). The specific number of cells identified in each sample is provided in Supplementary Table 2. A total of 195 cells in Clusters 19 and 21 carried markers of two cell types whose expression differed significantly at the same time. We conducted a secondary analysis and divided these 195 cells into three clusters to exclude the possibility of insufficient dimensionality reduction (Fig. 2C). A total of 154 cells in Clusters 0 and 2 expressed endothelial cell and B-cell markers simultaneously. Forty-one cells in Cluster 1 expressed both B-cell and T-cell markers (Fig. 2D-F). We decided not to include these double-positive cells in subsequent analyses to ensure the accuracy and reliability of our results. The calculation of the cell proportion for each cell type indicated significant changes (Fig. 2G). Among all the cells collected, endothelial cells accounted for the largest proportion and exhibited the most obvious increase, and macrophages presented the greatest change among immune cells. Consistent with previous reports, IH exposure increased the number of type 1 alveolar epithelial cells [21]. Notably, IH also promoted the aggregation of fibroblasts in the lung (Fig. 2H) [22, 23].

Characterization of fibroblasts in IH-exposed mouse lungs

Fibroblast heterogeneity was analyzed in scRNA-seq data from six samples (Fig. 3A). Reclustering revealed four fibroblast subtypes (F1-F4) with distinct gene expression profiles (Fig. 3B-C). Acta2⁺ fibroblasts (F1) expressed high levels of smooth muscle contraction genes such as Actc1 and Myh11 (Fig. 3E), consistent with myoid fibroblasts described previously [24]. GO analysis showed enrichment in actomyosin and muscle system processes for F1 (Fig. 3D). Tcf21⁺ fibroblasts (F2) highly expressed vascular development genes (Fig. 3E), with GO terms related to blood vessel morphogenesis (Fig. 3D). Dcn⁺ fibroblasts (F4) expressed extracellular matrix (ECM) genes (Col1a1, Col3a1) and were enriched in ECM organization pathways (Fig. 3D-E). A proliferative fibroblast cluster (F3) was identified by high ribosomal gene expression and low apoptosis gene expression (Fig. 3E).

Compared to controls, all four fibroblast subtypes expanded after IH exposure, with the most significant increases seen in F1 and F2 cells (Fig. 3F-G). Through transcription factor target analysis, we characterized the predominant regulatory networks in each fibroblast subtype. In F2 and F4 cells, Sp1—a transcription factor involved in transcriptional activation—was predicted to regulate numerous target genes. F1 cells were predominantly regulated by serum response factor (Srf), a well-known transcription factor implicated in the contractile phenotype of vascular smooth muscle cells (Fig. 3H) [25]. In F3 cells, Erg was identified as a key regulator, consistent with its known role in fibroblast activation [26]. Gene set scoring revealed F1 fibroblasts had the highest contractility score, F4 fibroblasts the highest ECM score, and F3 the lowest apoptosis score (Fig. 3I). Glycolytic activity scores did not differ significantly among the subtypes. During disease progression, significant proportional changes in fibroblast subtypes were observed. Trajectory analysis showed that F1 and F2 cells were evenly distributed along the developmental timeline, while F4 cells were mainly present during the intermediate stages, and F3 cells were predominantly enriched at the late stages. These trends may explain the greater incidence of hypertension and pulmonary fibrosis in OSAHS patients than in the general population(Fig. 3J-K, Supplementary Fig. 2) [27, 28].

Characterization of monocytes and macrophages in IH-exposed mouse lungs

The t-SNE analysis grouped monocytes and macrophages into 14 clusters based on 2,000 highly variable genes (Fig. 4A, Supplementary Fig. 3). Clusters 0, 5, and 8 were annotated as macrophage subtype 1 (Mφ1), marked by Wfdc21, Ear1, and Plet1. Clusters 2, 10, 11, and 12 formed monocyte subtype 1 (M1), expressing Treml4 and Cx3cr1, corresponding to nonclassical monocytes. Clusters 3, 4, 6, 7, and 13 were macrophage subtype 2 (Mφ2), marked by Cd209a, H2-Dmb1, Klrd1, and Ccnd1. Clusters 1 and 9 were monocyte subtype 2 (M2), expressing Ly6c2 and Ccr2, corresponding to classical monocytes (Fig. 4B-C).

Compared to controls, the proportion of M1 cells decreased while Mφ1 cells increased significantly. Pseudotime analysis indicated that monocytes progressed from M1 to M2, then to Mφ2 and ultimately Mφ1, supporting enhanced monocyte-to-macrophage transition in the lungs under IH exposure (Fig. 4F, Supplementary Fig. 4). Functional scoring revealed that Mφ1 exhibited the highest fatty acid oxidation and lowest inflammation levels among the four subtypes, while M2 showed the strongest chemotaxis, likely due to recruitment by inflammatory signals. Both macrophage subtypes had lower glycolysis levels than monocytes, and Mφ1 displayed disordered apoptosis-related gene expression, suggesting terminal differentiation (Fig. 4E-F). Transcription factor analysis indicated that Sp1 predominantly regulated Mφ1-related genes, with higher expression in the IH group (Fig. 4G-I). GO enrichment showed that M1 was associated with innate immune responses, M2 with immune regulation, Mφ2 with lymphocyte activation, and Mφ1 with lipid metabolism and endocytosis (Supplementary Fig. 4H-K).

Additionally, we analyzed cytokine expression in Mφ1 and Mφ2 from the lungs of IH-exposed mice. The results showed that Mφ1 is characterized by a high expression level of Il18, whereas Mφ2 displayed higher levels of Ccl5, Ccr2, and Cd68, indicative of a classical pro-inflammatory phenotype (Fig. 4J). GSEA of Mφ1 signature genes linked them to processes such as monovalent inorganic cation homeostasis, cell adhesion, and carbohydrate metabolism (Fig. 4K), indicating roles in ion balance and metabolic regulation. KEGG analysis showed significant enrichment of PPAR-related pathways in Mφ1 (Supplementary Fig. 4L), highlighting PPAR’s role in metabolism and immune function. Using lactate-treated macrophage data (GSE115354), we found 43 overlapping upregulated genes enriched in PPAR pathways (Fig. 4L-M), confirming PPAR’s involvement in IH-induced Mφ1 changes. Expression of key PPAR pathway genes (Acaa1a, Acaa1b, Acox1, Cpt1a, Lpl, Nfe2l2, Pparg) increased over pseudotime (Fig. 4N), indicating enhanced fatty acid oxidation and lipid metabolism in macrophages after IH exposure.

Characterization of T cells in IH-exposed mouse lungs

The t-SNE analysis of scRNA-seq data from six samples revealed the heterogeneity of T cells and identified nine distinct subtypes through clustering (Fig. 5A and B). By calculating the similarity between cell clusters (Fig. 5C), we further classified these clusters into five major subtypes (Fig. 5D). Subtype T1 (Cluster 6) corresponds to exhausted CD8⁺ T cells, characterized by markers of chronic antigen stimulation and functional exhaustion. Subtype T2 (Clusters 0, 1, 2, 3) represents memory CD8⁺ T cells with high Cd8a and Ccr7 expression, involved in antigen recognition and immune surveillance. Subtype T3 (Clusters 7, 8) is marked by elevated Gata3, Ccr2, and Cxcr6, consistent with CD4⁺ Th2 cells that regulate allergic inflammation. Subtype T4 (Cluster 4) shows effector memory features, with Ccl5, Cd8a, and Cxcr3 expression indicating strong migratory and antiviral capacity. Subtype T5 (Clusters 5, 9) comprises regulatory T cells (Tregs), characterized by Foxp3, Ctla4, and Lcos expression, crucial for immune tolerance and suppressing excessive immune responses.

After IH exposure, the proportion of memory T cells (T2) decreased significantly, while CD4⁺ Th2 cells (T3) increased (Fig. 5F-G). Notably, SP1, a transcriptional regulator linked to T3 signature genes, was significantly upregulated in the IH group, suggesting its role in T3 cell regulation (Fig. 5H). KEGG and GO analyses of T3 highly expressed genes showed enrichment in pathways related to CD4-positive, alpha-beta T-cell differentiation and activation, including Th17 and Th1/Th2 differentiation (Fig. 5I-J). Trajectory analysis showed that T1 and T2 cells were evenly distributed along the developmental timeline, while T4 cells were mainly enriched during the intermediate-to-late stages, and T3 cells were predominantly present at the late stage (Fig. 5L). T3 cells exhibited the highest inflammation score, indicating that their activation and cytokine secretion may exacerbate pulmonary inflammation in late-stage IH (Fig. 5K). Fatty acid oxidation was similar across subtypes, while T1 cells displayed markedly lower glycolysis, consistent with an exhausted phenotype. These findings demonstrate that IH reshapes T-cell composition and function, promoting Th2-skewed inflammation and T-cell exhaustion.

Characterization of endothelial cells in IH-exposed mouse lungs

We analyzed endothelial cells from six samples using scRNA-seq data. t-SNE analysis identified nine clusters (Fig. 6A), which were grouped into five subtypes based on cluster similarity (Fig. 6B): endo1 (Clusters 1 and 5), endo2 (Clusters 1, 3, and 4), endo3 (Clusters 6, 7, and 9), endo4 (Cluster 0), and endo5 (Cluster 8). The distribution of these subtypes across samples is shown in Fig. 6C and D. Marker gene analysis revealed distinct profiles: Cyb5r3 and Rps29 for endo1; Cldn5 and Foxf1 for endo2; Emp2 and Cdh5 for endo3; Vwf and Cytl1 for endo4, consistent with venous endothelial cells [29]; and Ccl21a and Mmrn1 for endo5, matching lymphatic endothelial cell signatures (Fig. 6E) [29]. Among these, endo2 was the most abundant, while endo5 was the least prevalent. Notably, IH exposure increased the proportions of endo1 to endo3 cells (Fig. 6F). Pseudotime trajectory analysis revealed a dynamic transition of endothelial subtypes over time: endo2 and endo5 predominated in the early stages, endo3 was evenly distributed throughout the trajectory, and endo1 and endo4 became dominant at the terminal stage (Fig. 6G, Supplementary Fig. 5). KEGG enrichment analysis demonstrated subtype-specific functional programs: endo1 was enriched in “ribosome” and “oxidative phosphorylation”; endo2 in “antigen processing and presentation” and “fluid shear stress and atherosclerosis”; endo3 in “IL-4/IL-13 signaling” and “apoptosis”; endo4 in “TGF-beta signaling” and “stem cell pluripotency”; and endo5 in “ferroptosis” and “autophagy” (Fig. 6H). These findings suggest that during IH exposure, endothelial cells undergo functional reprogramming from immune and stress responses toward a more reparative and metabolically active phenotype.

Analysis of interactions between monocytes/macrophages and other altered cell types

We analyzed receptor‒ligand interactions between monocytes/macrophages and other altered cell types to further elucidate the cellular interactions within lung tissue exposed to IH. The results revealed that the interactions between the endo2 and other cell subtypes were particularly strong. Among these interactions, endo2 cells exhibited similarly strong interactions with both Mφ1 and M2 , and these interactions were markedly stronger than those observed with other cell types (Fig. 7A and B). Additionally, interactions between Tgfb1–(Tgfbr1 + Tgfbr2), Ccl6–Ccr2, Sema3a–(Nrp1 + Plxna2), and Sema3c–(Nrp1 + Plxna2) are highly probable within the network of monocytes/macrophages and other altered cell types, with Ccl6–Ccr2 showing the strongest potential for interaction. Notably, the receptor for Ccl6, Ccr2, was expressed in multiple cell subtypes, with prominent expression in endothelial cells and macrophage/monocyte-related subtypes (Fig. 7E). We performed immunohistochemistry to investigate this phenomenon further, and the results confirmed that Ccr2 expression was indeed upregulated in the lung tissue of the IH-exposed group (Fig. 7F), suggesting that Ccr2 may play a significant role in the pathophysiology of IH.

Effect of the SP1 intervention on the lung tissue of IH-Exposed mice

In this study, we first confirmed that SP1 expression was significantly upregulated in the lungs of mice exposed to IH, as evidenced by both immunofluorescence staining and western blot analysis (Fig. 8A-D). Based on this observation, we established an IH mouse model combined with the SP1 inhibitor plicamycin to further investigate its therapeutic potential. Western blot analysis and immunofluorescence analysis (Fig. 8E-F and H-I) showed that plicamycin treatment partially reversed the IH-induced upregulation of SP1 in lung tissue. Similarly, measurement of lung hydroxyproline levels suggested a modest reduction following plicamycin treatment compared with IH alone (Fig. 8G). Subsequently, histopathological evaluation revealed that plicamycin tended to attenuate IH-induced pulmonary inflammation, mucus secretion, and collagen deposition (Fig. 8H and J–L), indicating that SP1 plays a central role in mediating lung injury under IH exposure. Additionally, phalloidin staining revealed increased F-actin remodeling and stress fiber formation in the lungs of IH-exposed mice, which was partially alleviated by plicamycin treatment (Fig. 8H and M). Furthermore, PCR analysis revealed that plicamycin effectively reduced the expression of genes involved in inflammation and fibrosis, including Fizz1, Ym1, Il6, α-Sma, Cd68, Ccl2, Cd80, Ccl7, Col1a1, and iNOS (Supplementary Fig. 6 A), further suggesting that SP1 plays a key role in pulmonary inflammation and fibrosis induced by IH exposure. In conclusion, the SP1 inhibitor plicamycin significantly suppresses the inflammatory response, cytoskeletal remodeling, and fibrosis in the lung tissue of IH-exposed mice.