Abstract
Background
Neutrophil extracellular traps (NETs) might be promising targets for the evaluation and treatment of pulmonary hypertension. However, no study has systematically screened NETs as feature genes and therapeutic targets in pulmonary hypertension. The present study aims precisely to address this issue.
Methods
GSE15197, GSE48149 and GSE53408 were used as the training group, whereas GSE113439 was used as the validation group for comprehensive bioinformatics analyses, and the screened feature genes were further confirmed using immunoblotting with animal lung tissues.
Results
First, 12 differentially expressed NETs-related genes were significantly upregulated in pulmonary hypertension patients compared with healthy controls. Second, immune infiltration analysis revealed that immature B cells, immature dendritic cells, T helper cells and memory CD8+ T cells were increased in pulmonary hypertension patients. Third, three machine learning models identified five feature genes for pulmonary hypertension, including IQGAP2, HSP90AA1, HK2, MYO5A and NCL, which were further confirmed through animal experiments. In addition, clustering analysis revealed two distinct subtypes of pulmonary hypertension with different immune microenvironment profiles. Finally, hypothesis-generating computational screen of drug-gene interactions yielded potential inhibitors against the five feature genes.
Conclusions
Five NETs-related genes, namely IQGAP2, HSP90AA1, HK2, MYO5A and NCL, were identified as feature genes for pulmonary hypertension. The immune microenvironment played an important role in the pathogenesis of pulmonary hypertension, and patients in cluster C1 were considered the main disease subtype. In addition, potential inhibitors targeting the five feature genes were further analyzed by computational screen.
Keywords: Neutrophil extracellular traps (NETs), pulmonary hypertension, feature genes, immune infiltration, inhibitors
Highlight box.
Key findings
• Five neutrophil extracellular traps (NETs)-related genes, namely IQGAP2, HSP90AA1, HK2, MYO5A and NCL, were identified as feature genes and potential therapeutic targets for pulmonary hypertension, and their potential inhibitors were further analyzed.
What is known and what is new?
• NETs play a crucial role in the formation of pulmonary hypertension, and some of them might be promising targets for the evaluation and treatment of pulmonary hypertension.
• Five NETs-related genes were identified as feature genes for pulmonary hypertension through comprehensive bioinformatics analyses, and they were further confirmed using immunoblotting with animal lung tissues. The immune microenvironment played an important role in the pathogenesis of pulmonary hypertension, and patients in cluster C1 were considered the main disease subtype. In addition, potential inhibitors targeting the five feature genes were further analyzed.
What is the implication, and what should change now?
• The study provided a theoretical basis for the exploration of diagnostic markers and potential therapeutic targets for pulmonary hypertension.
Introduction
Pulmonary hypertension refers to a hemodynamic state in which the mean pulmonary artery pressure measured by a right cardiac catheter is ≥20 mmHg at sea level at rest, resulting in right ventricular dysfunction, a series of clinical symptoms, and even death (1). There is no effective traditional treatment for pulmonary hypertension. In recent years, more than 10 new targeted drugs have been developed to intervene against important targets in the pathogenesis of pulmonary hypertension (2). Although these targeted drugs reduce the pulmonary artery pressure of patients to a certain extent and delay the progression of right heart failure, the treatment compliance of patients with pulmonary hypertension has been greatly reduced due to their side effects and high treatment costs, bringing new challenges to the treatment of pulmonary hypertension (3-5). Therefore, further exploration of new intervention targets in the pathological mechanism of pulmonary hypertension and the development of safer, more effective and more economical new drugs are imperative.
Neutrophil extracellular traps (NETs) are web-like filamentous extracellular structures released by neutrophils in response to supernumerary or oversized pathogens and entrap pathogens in a network of DNA, histones, proteases, and other cytotoxic and highly inflammatory compounds (6). An increasing number of studies have demonstrated that NETs play critical roles not only in the eradication of pathogens but also in cancer initiation and progression (7,8). Until recently, a few studies have reported that the levels of NETs are elevated in patients with pulmonary hypertension (9,10), suggesting their potential as promising targets for the evaluation and treatment of pulmonary hypertension.
However, no study has systematically screened NETs as feature genes and therapeutic targets in pulmonary arterial hypertension. With the rapid development of bioinformatics technology, disease feature genes can be accurately identified via comprehensive analysis of gene expression profiles.
The present study aimed to identify NETs-related genes associated with pulmonary hypertension through integrated bioinformatics analysis and further confirmed them through animal experiments, providing novel insights for the exploration of diagnostic markers and potential therapeutic targets in pulmonary hypertension. We present this article in accordance with the TRIPOD and ARRIVE reporting checklists (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0557/rc).
Methods
Data acquisition
The GSE15197, GSE48149, GSE53408 and GSE113439 datasets were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/, accessed on 10 May 2025) and the joint analysis was conducted. As specified in the description provided by the original submitter, the data had already undergone background correction and normalization. An examination of the matrix confirmed the absence of any missing values. These datasets were all from human lung tissue, including the pulmonary hypertension group and the healthy control group. Among the 4 datasets, GSE15197, GSE48149 and GSE53408 were used as the training group, while GSE113439 was used as the validation group for subsequent analysis. A total of 61 pulmonary hypertension samples and 46 normal control samples were analyzed via R software (version 4.5.0) in this study. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
Gene expression and functional enrichment analyses in pulmonary hypertension
The differentially expressed genes (DEGs) between the pulmonary hypertension group and the control group were identified with the “limma” package. A P value <0.05 and a |log2 fold change (FC)| >0.379 were considered statistically significant. This threshold (|log2FC| >0.379) was chosen to ensure an adequate number of genes for downstream functional enrichment analyses, as these analyses require a gene set of sufficient size to produce biologically meaningful results. To validate the robustness of the DEG results, a sensitivity analysis was subsequently performed using a threshold of |log2FC| >0.585. In addition, a heatmap and volcano plot were generated to display the DEGs. The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment results of the DEGs were subsequently analyzed with the “clusterProfiler” package.
Filtering for key module genes by weighted gene co-expression network analysis (WGCNA) in pulmonary hypertension
WGCNA was used to construct the co-expression network in the training group with the “WGCNA” package. First, to ensure the accuracy of the analysis, the outliers were removed, and the samples were clustered. The optimal soft threshold (β) was then chosen so that the network approximated a scale-free distribution. Next, the cluster dendrogram was obtained by calculating adjacency and similarity. The modules were partitioned via a dynamic tree-cutting algorithm, and similar modules were merged by setting MEDissThres to 0.3. The correlation between each module and pulmonary hypertension was subsequently evaluated, and the module with the highest absolute value of the correlation coefficient was selected as the key module in pulmonary hypertension.
Identification of NETs-related genes in pulmonary hypertension
NETs-related genes were downloaded from the GeneCards database (https://www.genecards.org/, accessed on 10 May 2025). NETs-related genes with a relevance score >3 from the GeneCards database were retained for subsequent analysis, producing a total of 3,312 candidate genes. To assess the biological relevance of our initially screened NETs-related genes, we performed a comparison with an experimentally supported NETosis gene set. Specifically, we obtained a core NET gene set from published literature (11), which comprises 137 genes with direct experimental evidence for involvement in NET formation or regulation. The overlap between our GeneCards-derived list (3,312 genes) and this experimentally curated set (137 genes) comprised 107 genes (available at: https://cdn.amegroups.cn/static/public/jtd-2026-0557-1.xlsx). Finally, to identify NETs-related genes relevant to pulmonary hypertension, the intersecting genes from three sources were identified: the DEGs, genes in the key WGCNA module, and the GeneCards-derived list described above. The results were visualized using a Venn diagram and box plots. In addition, significant NETs-related genes associated with pulmonary hypertension were mapped to chromosomal positions via the “circlize” package and “CMplot” package.
Identification of feature genes in pulmonary hypertension
Three methods were used to screen for feature genes related to pulmonary hypertension in the training group. First, the feature genes associated with pulmonary hypertension were selected via a least absolute shrinkage and selection operator (LASSO) binary logistic regression model with the “glmnet” package. Second, the feature genes associated with pulmonary hypertension were identified via the support vector machine recursive feature elimination (SVM-RFE) method with the “e1071” package. Third, the feature genes associated with pulmonary hypertension were identified via a random forest model by creating 500 trees per data point with the “randomForest” package, and genes with DecreaseGini scores exceeding 1 were selected. Ultimately, the feature genes supported by at least two methods were used for subsequent analysis.
Construction and validation of the column line plot model
A column line plot model was established with the “rms” package and “rmda” package to evaluate the risk of pulmonary hypertension using the levels of feature genes. Each gene was used to predict a score, and the total score was the cumulative sum of these scores. The predictive accuracy of the model was assessed through calibration curves and decision curve analysis (DCA). A box plot was generated with the “reshape2” package and “ggpubr” package to illustrate the differences in the expression of feature genes between the pulmonary hypertension group and the control group. Finally, receiver operating characteristic (ROC) curves were computed with the “pROC” package, and the predictive powers of the feature genes were assessed via the area under the curve (AUC).
Unsupervised clustering in pulmonary hypertension
Five feature genes previously identified in pulmonary hypertension patients were further clustered via the “ConsensusClusterPlus” package in an unsupervised manner. The cumulative distribution function (CDF) curve, as a key index for determining the optimal number of pulmonary hypertension subtypes and evaluating the stability of our consensus clustering analysis, was used to cluster 61 pulmonary hypertension samples into distinct groups. In addition, the distinctions between subtypes were determined via principal component analysis (PCA).
Generation of the rat pulmonary hypertension model
The twenty male Sprague-Dawley (SD) rats weighing 150 to 200 g used in this study were purchased from Lanzhou Veterinary Research Institute, Chinese Academy of Agricultural Sciences. The rats were kept under standard conditions (18–22 ℃ and 40–60% humidity) in a specific pathogen-free animal laboratory, and they had free access to food and water. To mitigate experimental bias, the feeding, model establishment, identification, and random grouping of rats were conducted by a dedicated experimenter who was blinded to subsequent data collection and analysis. The experimenter assigned the animals using a random number table and applied color coding to cage cards, thereby precluding visual identification of group allocation by other researchers performing experimental procedures. Twenty rats were randomly divided into 2 groups (n=10 in each group): the control group and the monocrotaline (MCT) group. The rat pulmonary hypertension model was induced by intraperitoneal injection of MCT (60 mg/kg) on day 1. On the 28th day, all surviving rats were anesthetized with 10% chloral hydrate (3–5 mL/kg) via intraperitoneal injection. Next, the rats were placed in the supine position, and a polyvinyl catheter filled with heparin saline solution was inserted into the right ventricle via the isolated right external jugular vein. Finally, the right ventricle systolic pressure (RVSP) was measured via a Grass polygraph (Power Lab, Australia). A protocol was prepared before the study without registration. All animal experiments were approved by the Laboratory Animal Care Committee of Lanzhou University Second Hospital (No. D2023-054), and were conducted in compliance with the U.S. National Research Council’s guide for the care and use of laboratory animals.
Immunoblotting
In the study, two animals in the MCT group died, whereas all control animals survived. From the surviving rats in each cohort, lung tissues from three individuals with RVSP values closest to the group median were selected for Immunoblotting. First, rat lung tissues were homogenized in RIPA lysis buffer (HEART, Xi’an, China) containing protease inhibitors, phosphatase inhibitors and phenylmethylsulfonyl fluoride (PMSF) and centrifuged at 12,000 rpm at 4 ℃ for 20 min. Then, the supernatants were collected as sample proteins. Next, the proteins were separated on a 6–10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA, USA) via semidry transfer. The membranes were subsequently incubated overnight at 4 ℃ with primary antibodies against IQGAP2 (#BS72586; Bioworld Technology, Louis Park, MN, USA; 1:1,000), HSP90AA1 (#R24635; Zhengbio, Shanghai, China; 1:1,000), hexokinase II (HK2) (#WL02454; Wanleibio, Shenyang, China; 1:1,000), MYO5A (#DF4200; Affinity Biosciences, Cincinnati, OH, USA; 1:1,000), Nucleolin (NCL) (#WL04279; Wanleibio; 1:1,000) and β-actin (#WL01372; Wanleibio; 1:1,000). After being washed three times in Phosphate-Buffered Saline with Tween-20 (PBST) for 15 min, the membranes were incubated at room temperature for 2 h with horseradish peroxidase-conjugated goat anti-rabbit secondary antibodies (#ab7090; Abcam, Cambridge, UK; 1:5,000). Finally, the protein bands were visualized via SuperSignal West Pico Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL, USA) and quantified via Quantity One software (Bio-Rad).
Statistical analysis
Statistical analysis was performed with SPSS 13.0 software (SPSS Inc., Chicago, IL, USA). All the data were expressed as the means ± standard deviations and were analyzed via independent-sample t-tests. P<0.05 was considered to indicate a statistically significant difference.
Results
Identification of DEGs and functional enrichment analyses in pulmonary hypertension
The datasets GSE15197, GSE48149 and GSE53408 were used as the training group to obtain the DEGs. After batch correction, the differences among the three datasets were eliminated, resulting in a mixed random distribution. After batch correction, the three datasets were more suitable for further analysis as a whole training group. Then, DEG analysis was carried out on 48 pulmonary hypertension patients and 33 healthy controls in the training group. The results revealed that 69 genes were upregulated and that 130 genes were downregulated in pulmonary hypertension patients compared with controls, as presented in the volcano plot and heatmap (Figure 1A,1B). These DEGs were subsequently annotated via GO and KEGG pathway enrichment analyses. GO analysis revealed that these dysregulated genes were enriched in muscle contraction- and development-related biological processes (Figure 1C). KEGG pathway analysis revealed that these dysregulated genes were involved mainly in cardiovascular disease-related and cell proliferation-related pathways, such as the PI3K-Akt signaling pathway, calcium signaling pathway, and Notch signaling pathway (Figure 1D).
Figure 1.
DEGs and functional enrichment analyses in pulmonary hypertension. (A) The DEGs between pulmonary hypertension patients and healthy controls are visualized in a volcano plot. (B) DEGs between pulmonary hypertension patients and healthy controls are presented in a heatmap. (C) GO enrichment analysis of the DEGs. (D) KEGG pathway enrichment analysis of the DEGs. BP, biological process; CC, cellular component; DEGs, differentially expressed genes; FC, fold change; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.
Construction of key gene modules and co-expression networks in pulmonary hypertension
Key gene modules and co-expression networks were constructed to identify the critical gene modules related to pulmonary hypertension via WGCNA. During the process of WGCNA, soft threshold powers (β) from 1 to 20 were evaluated. A value of β=6 was selected as optimal. At this threshold, the scale-free topology fit index reached R2=0.89, satisfying the recommended criterion of R2>0.8. The corresponding mean connectivity at β=6 was 5.7, striking a balance between retaining biologically relevant co-expression relationships and minimizing spurious connections (Figure 2A). Five expression modules, blue, yellow, brown, turquoise and gray, were subsequently built via the dynamic cutting algorithm, of which the brown module had the strongest correlation with pulmonary hypertension (r=−0.56, P<0.001), as presented in Figure 2B-2D. In addition, the brown module contained a total of 73 genes, which were considered closely associated with pulmonary hypertension and were used for subsequent analysis.
Figure 2.
Key gene modules and co-expression networks in pulmonary hypertension. (A) The optimum soft threshold value. (B) Construction of expression modules via the dynamic cutting algorithm. (C) Clustering of the gene dendrogram via WGCNA. (D) Relationships between the five modules and pulmonary hypertension presented by a histogram. WGCNA, weighted gene co-expression network analysis.
Screening of differentially expressed NETs-related genes in pulmonary hypertension
To identify the most biologically relevant genes involved in the pathogenesis of pulmonary hypertension, differentially expressed NETs-related genes were screened by taking the intersection of genes in the brown module, DEGs and NETs-related genes. Finally, 12 differentially expressed NETs-related genes were identified (Figure 3A). In addition, 12 differentially expressed NETs-related genes, including IQGAP2, FNBP1L, MYO5A, ALCAM, HSP90AA1, NCL, ANKRD50, AKAP13, CYBB, EGFL6, HK2 and ITGA2, were significantly upregulated in pulmonary hypertension samples compared with healthy control samples, as shown in Figure 3B (P<0.001). The chromosomal positions of the 12 differentially expressed NETs-related genes are depicted in a circular diagram (Figure 3C), and the relative expression levels of the 12 NETs-related genes between pulmonary hypertension patients and healthy controls are shown in a Manhattan map (Figure 3D).
Figure 3.
Differentially expressed NETs-related genes in pulmonary hypertension. (A) Identification of differentially expressed NETs-related genes. (B) The levels of differentially expressed NETs-related genes between pulmonary hypertension patients and healthy controls. (C) Chromosomal positions of differentially expressed NETs-related genes. (D) Comparison of the relative expression levels of NETs-related genes between pulmonary hypertension patients and healthy controls via a Manhattan map. *, P<0.05; ***, P<0.001. DEGs, differentially expressed genes; NET, neutrophil extracellular trap; WGCNA, weighted gene co-expression network analysis.
Additionally, we re-performed the differential expression analysis using a stricter threshold of |log2FC| >0.585 while maintaining a P value <0.05. The resulting DEGs (available at: https://cdn.amegroups.cn/static/public/jtd-2026-0557-2.xlsx) were then intersected again with genes in the brown module and NETs-related genes. The results show that 8 of the original 12 NETs-related DEGs (IQGAP2, FNBP1L, MYO5A, ALCAM, HSP90AA1, ANKRD50, CYBB and EGFL6) remained significantly differentially expressed under the more stringent cutoff. These 8 genes constitute a robust core gene set, indicating that our central findings remain solid even under more rigorous criteria.
Immune cell infiltration in pulmonary hypertension
The different levels of immune cells in lung tissues from pulmonary hypertension patients and healthy controls were obtained via the single-sample gene set enrichment analysis (ssGSEA) method and further displayed with a heatmap and a violin plot. As presented in Figure 4A,4B, the content of immature B cells, immature dendritic cells, T helper cells and memory CD8+ T cells were increased in pulmonary hypertension patients compared to the controls (P<0.05), while the content of activated CD4+ T cells, natural killer cells and eosinophils were reduced in pulmonary hypertension patients compared to the controls (P<0.05). In addition, correlation analysis revealed that the differentially expressed NETs-related genes were positively correlated with T helper cells, dendritic cells and macrophages and negatively correlated with regulatory T cells, natural killer cells, eosinophils and CD8+ T cells (Figure 4C). These results indicated that the differentially expressed NETs-related genes might play important roles in immune cell infiltration during the pathogenesis of pulmonary hypertension.
Figure 4.
Immune cell infiltration in pulmonary hypertension. (A) Enriched immune cells in every sample shown in a heatmap. (B) The distinct fractions of immune cells between pulmonary hypertension patients and controls are displayed in a violin plot. (C) Correlation between differentially expressed NETs-related genes and immune cells. NET, neutrophil extracellular trap.
Identification of feature genes for pulmonary hypertension
The feature genes associated with pulmonary hypertension among the 12 differentially expressed NETs-related genes were identified via LASSO binary logistic regression, the SVM-RFE method and a random forest model. First, the expression patterns of 12 differentially expressed NETs-related genes were integrated into LASSO regression, and the optimal lambda value of five was determined, indicating that five feature genes are involved in pulmonary hypertension, as shown in Figure 5A. Second, the SVM-RFE method identified 9 feature genes associated with pulmonary hypertension with the highest accuracy (Figure 5B). Third, the random forest model revealed 10 feature genes associated with pulmonary hypertension with importance scores ≥1 (Figure 5C). Ultimately, the genes screened by at least two methods were identified as feature genes for pulmonary hypertension (Figure 5D), including IQGAP2, HSP90AA1, HK2, MYO5A and NCL.
Figure 5.
The feature genes associated with pulmonary hypertension. (A) LASSO binary logistic regression was used to determine five feature genes. (B) Screening of feature genes via the SVM-RFE method. (C) Acquisition of feature genes via the RF model. (D) Identification of feature genes for pulmonary hypertension via at least two methods. CV, coefficient of variation; LASSO, least absolute shrinkage and selection operator; RF, random forest; RFE, recursive feature elimination; SVM, support vector machine.
Diagnostic efficiency of the feature genes for pulmonary hypertension
The nomogram indicated that these five feature genes had high predictive efficiency for pulmonary hypertension. Among them, IQGAP2 had the highest diagnostic value, followed by HSP90AA1 and MYO5A (Figure 6A). Then a moderate disparity between the predicted and actual cluster risk for pulmonary hypertension was shown by the calibration curve (Figure 6B). The DCA results corroborated the high accuracy of nomogram (Figure 6C), pointing to its value as a resource for directing treatment strategies in pulmonary hypertension. Next, the levels of five feature genes, including IQGAP2, HSP90AA1, HK2, MYO5A and NCL, were analyzed in the validation set GSE113439. The results revealed that the levels of all five genes in pulmonary hypertension patients were significantly greater than those in the controls (Figure 6D). In addition, the ROC curve analysis indicated a minimum AUC value of 0.749 for the five feature genes in the validation set and an overall AUC value of 0.889 (Figure 6E,6F), suggesting that the five feature genes have good predictive ability for pulmonary hypertension.
Figure 6.
Diagnostic efficiency of the feature genes for pulmonary hypertension. (A) Nomogram for predicting the occurrence of pulmonary hypertension on the basis of feature genes. (B) Calibration curve. The y-axis and x-axis show the actual and predicted probabilities of pulmonary hypertension, respectively. The diagonal dotted line indicates perfect calibration. The solid line represents the bias‑corrected performance of the model; closer fit to the diagonal denotes better prediction. (C) Decision curve analysis. The red represents the net benefit of the nomogram in the prediction of pulmonary hypertension occurrence. The grey line represents the assumption that all people have pulmonary hypertension. The black line represents the assumption that all people do not have pulmonary hypertension. (D) Levels of the five feature genes in the validation group. (E) ROC curves of the five feature genes for the diagnosis of pulmonary hypertension. (F) Merged ROC curve of the five feature genes for the diagnosis of pulmonary hypertension. ***, P<0.001. AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristic.
Classification of NETs-related gene clusters in pulmonary hypertension
As shown in Figure 7A, pulmonary hypertension patients were stably classified into two distinct molecular subtypes via consensus clustering analysis. The PCA plot further visualized the optimal clustering with vital intracluster consensus and clear separation between clusters, forming distinct groups C1 and C2 (Figure 7B). Next, the five feature genes associated with pulmonary hypertension were increased in cluster C1 in the heatmap and box plots, as indicated in Figure 7C. The differences in the immune microenvironment between the two clusters were subsequently analyzed, and the results indicated that patients in cluster C1 were highly similar to the overall pulmonary hypertension population, making it the main disease subtype (Figure 7D).
Figure 7.
Classification of NETs-related gene clusters associated with pulmonary hypertension. (A) Classification of two distinct molecular subtypes of pulmonary hypertension by consensus clustering analysis. (B) PCA of the 2 clusters. (C) Differential expression levels between the two clusters in the box plots. (D) Differences in immune cell infiltration between the two clusters. *, P<0.05; **, P<0.01; ***, P<0.001. NET, neutrophil extracellular trap; PCA, principal component analysis.
Hypothesis-generating computational screen of drug-gene interactions
The drug-gene interaction data were downloaded from the Drug SIGnatures DataBase. Then, Cytoscape software was used to construct the drug-gene interaction network. As shown in Figure 8, the potential inhibitors against IQGAP2 were COBALT and Periodate-oxidized adenosine, whereas the potential inhibitors for HSP90AA1 included 1,10-phenanthroline, ketoconazole, sodium dichromate, and clotrimazole baclofen. Potential drugs that act on HK2 include 1,10-phenanthroline, ketoconazole, COBALT, clotrimazole, quinidine, cobalt sulfate, and beta-D-glucose 6-phosphate; potential drugs for MYO5A include sodium dichromate and periodate-oxidized adenosine. In addition, predicted drugs against NCL include 2-bromo-3-hydroxy-4-methoxybenzaldehyde, Baclofen and arsenic(III) chloride. These computationally predicted drug inhibitors provide candidate compounds for future investigations.
Figure 8.
Hypothesis-generating computational screen of drug-gene interactions. The red circles indicate the feature genes, and the blue boxes indicate the drugs.
Confirmation of the feature genes associated with pulmonary hypertension in rat lung tissues
After 28 days of MCT injection, the RVSP in MCT-treated rats was significantly increased compared with the control group (50.62±3.04 versus 20.80±1.95 mmHg, P<0.05), as shown in Figure 9A,9B. The results suggest that pulmonary hypertension was successfully induced by MCT in rats. Finally, we collected lung tissues from rats with pulmonary hypertension and healthy controls to assess the protein levels of the five feature genes by immunoblotting. The results revealed that the expression levels of all five feature genes were upregulated in the pulmonary hypertension group compared with those in the control group (all P<0.05, Figure 9C), which was consistent with our findings from the bioinformatics analysis.
Figure 9.
The protein levels of the feature genes associated with pulmonary hypertension in rat lung tissues. (A) RVSP waveform in each group. (B) Changes of RVSP in each group (n=8–10). (C) The protein levels of IQGAP2, HSP90AA1, HK2, MYO5A, and NCL in lung tissues from the pulmonary hypertension group and control group were examined by immunoblotting (n=3). *, P<0.05 versus the control group. MCT, monocrotaline; RVSP, right ventricular systolic pressure.
Discussion
This study screened 12 differentially expressed NETs-related genes, among which IQGAP2, HSP90AA1, HK2, MYO5A, and NCL were identified as feature genes for pulmonary hypertension. Immune cell infiltration played an important role in the pathogenesis of pulmonary hypertension, and patients in cluster C1 were considered the main disease subtype. Finally, computational prediction of drug-gene interactions were performed, and potential inhibitors of the five feature genes for pulmonary hypertension were yielded.
Five NETs-related genes were identified as feature genes for pulmonary hypertension through comprehensive bioinformatics analyses, and they were further confirmed using immunoblotting with animal tissues. However, there are some limitations in this study. First, the results were based on transcriptomic data rather than protein data. Second, although the five feature genes were upregulated in lung tissues of rats with pulmonary hypertension, the specific experimental evidence that the five genes could be used as promising candidate markers and potential therapeutic targets for pulmonary hypertension was lacking. Third, the |log2FC| >0.379 threshold used for DEG screening is relatively lenient, which may increase the inclusion of false positives with limited biological relevance. Fourth, the NETs-related feature genes were identified via a text-mining approach and thus do not include canonical genes directly involved in NET formation or regulation. These candidate genes may instead represent disease-contextual modulators or secondary consequences of the NETosis process. Fifth, due to the limited sample size—particularly in the external validation cohort—our prognostic model and feature genes should be considered a robust hypothesis-generating resource; their definitive clinical utility requires prospective validation in larger, multi‑center cohorts. Finally, the two molecular subtypes (C1 and C2) identified in this study have not been linked to key clinical parameters; therefore, their prognostic or therapeutic utility remains unclear. Future studies in well‑annotated, prospective cohorts are needed to determine whether these subtypes correlate with disease progression, differential treatment response, or survival, which would establish their potential clinical applicability.
Twelve differentially expressed NETs-related genes were screened, and they were upregulated in the lung samples of pulmonary hypertension patients. In addition to the feature genes, the remaining upregulated NETs-related genes attracted our attention. FNBP1L encodes different protein isoforms that are involved in the regulation of cell shape, polarity, motility and signal transduction, which are essential for the elimination or suppression of intracellular pathogens (12,13). As a member of the immunoglobulin superfamily, ALCAM regulates cell adhesion, differentiation, proliferation, and migration, participating in the process of immune defense and the inflammatory response (14). The AKAP13 protein has been shown to coordinate signals originating from the cell membrane and has many common targets, including RhoA and PKA, which play important roles in tumor and nontumor proliferative diseases by affecting many effector molecules (15,16). The CYBB protein is an important membrane-bound oxidase in phagocytes and is critical for superoxide production. Multiple studies have shown that CYBB plays an indispensable role in many diseases, such as tumors, excessive inflammation and immune disorders, through a variety of molecular mechanisms (17,18). EGFL6 is a member of the EGF superfamily of proteins; it is upregulated in tumorigenesis, epithelial-to-mesenchymal transition, angiogenesis and osteogenesis (19,20). ITGA2 is a subunit of the collagen receptor integrin alpha-2 and plays a critical role in a variety of cancers and cardiovascular diseases (21,22). Therefore, we speculate that these upregulated NETs-related genes might play crucial roles in the occurrence and development of pulmonary hypertension, which needs further confirmation in future experiments.
The differences in immune cell infiltration between pulmonary hypertension patients and controls indicate that the pathogenesis of pulmonary hypertension involves complex immune dysregulation and that the differentially expressed NETs-related genes might participate in this process. Multiple key immune cell disorders, such as those involving neutrophils, monocytes, macrophages, and CD4+ and CD8+ T cells, occur in patients with pulmonary hypertension and promote pulmonary vascular injury as well as abnormal repair through aberrant immune responses (23,24). These findings are consistent with our current research, which revealed the upregulation of immature B cells, immature dendritic cells and memory CD8+ T cells and the downregulation of activated CD4+ T cells, natural killer cells and eosinophils in pulmonary hypertension patients. The increase in immature B cells and immature dendritic cells indicates an increase in adaptive immunity and is involved in the pathogenesis of pulmonary hypertension. The increase in CD8+ T cells and decrease in CD4+ T cells suggest adaptive immune injury, which contributes to major pathophysiological changes in vascular cells and leads to vascular remodeling, increased pulmonary vascular resistance and eventually pulmonary hypertension. The decrease in natural killer cells and eosinophils indicates that innate immunity is impaired, which may be associated with adaptive immunity in the pathogenesis of pulmonary hypertension (25). In addition, the relationships between differentially expressed NETs-related genes and different immune cell populations indicate that NETs-related genes might participate in the pathogenesis of pulmonary hypertension by affecting the immune microenvironment.
Five differentially expressed NETs-related genes, IQGAP2, HSP90AA1, HK2, MYO5A and NCL, were identified as feature genes for pulmonary hypertension in the present study. IQGAP2 has been identified as a large cytoplasmic scaffold protein and acts as a tumor suppressor in malignancies (26). At present, there is no research on the role of IQGAP2 in the pathogenesis of pulmonary hypertension. However, studies have shown that IQGAP2 is correlated with the MAPK/ERK, Wnt and PI3K-AKT signaling pathways, which are crucial signaling pathways in the progression of pulmonary hypertension (27). The HSP90AA1 gene encodes heat shock protein 90α, which belongs to the ATP-dependent chaperone family and is involved in the cell stress response, signal transduction and mitochondrial protein transport. Deng et al. reported that the expression of HSP90AA1 is significantly increased in rat models of pulmonary hypertension and that the inhibition of HSP90AA1 reduces the proliferation and migration of pulmonary artery smooth muscle cells (PASMCs) (28). HK2 is the first enzyme in the glycolytic pathway. Wang et al. and Zhang et al. demonstrated that the HK2 protein is upregulated in hypoxic PASMCs from humans and mice, while the inhibition of HK2 alleviates pulmonary artery remodeling by reducing PASMC proliferation (29,30). MYO5A is a dimeric motor protein that transports molecular cargos along actin filaments. Kasavi reported that MYO5A is a feature gene for pulmonary hypertension (31), which is consistent with our results. NCL is a core protein present in the nucleolus of eukaryotic cells and is involved in basic biological processes such as chromatin remodeling and RNA metabolism. NCL protein levels are increased in hypoxia-induced pulmonary hypertension mice, and the inhibition of NCL suppresses the proliferation and migration of PASMCs (32).
Another important finding of the present study was the classification of two distinct molecular subtypes with different immune microenvironment features. Cluster C1 was the main disease subtype due to its high similarity to the overall pulmonary hypertension population in terms of the levels of the five feature genes. Furthermore, our computational drug-gene interaction screen identified several candidate compounds (e.g., ketoconazole, clotrimazole, quinidine) inhibiting our feature genes. This analysis serves an exploratory, hypothesis-generating role. Future studies using pulmonary hypertension-relevant experimental models would be valuable to validate the biological efficacy and specificity of these candidates.
The study lays an important theoretical foundation for elucidating the roles of the five NETs-related genes in pulmonary hypertension, and their specific mechanisms need to be explored in future experiments in vivo and in vitro.
Conclusions
Five NETs-related genes, namely, IQGAP2, HSP90AA1, HK2, MYO5A and NCL, were identified as feature genes for pulmonary hypertension. The immune microenvironment played an important role in the pathogenesis of pulmonary hypertension, and patients in cluster C1 were considered the main disease subtype. In addition, potential inhibitors targeting the five feature genes were further analyzed by computational screen.
Supplementary
The article’s supplementary files as
Acknowledgments
The authors would like to thank contributions from the GEO database, GeneCards database and Drug SIGnatures database.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All animal experiments were approved by the Laboratory Animal Care Committee of Lanzhou University Second Hospital (No. D2023-054), and were conducted in compliance with the U.S. National Research Council’s guide for the care and use of laboratory animals.
Footnotes
Reporting Checklist: The authors have completed the TRIPOD and ARRIVE reporting checklists. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0557/rc
Funding: This work was supported by the National Natural Science Foundation of China (No. 82360014, to F.L.), Cuiying Scientific and Technological Innovation Program of Lanzhou University Second Hospital (No. CY2023-MS-B05, to F.L.), and the Key Incubation Project Funds of the Second Hospital & Clinical Medical School, Lanzhou University (No. 2025-23-zdfy-004, to F.L.).
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0557/coif). All authors report that this work was supported by the National Natural Science Foundation of China (No. 82360014, to F.L.), Cuiying Scientific and Technological Innovation Program of Lanzhou University Second Hospital (No. CY2023-MS-B05, to F.L.), and the Key Incubation Project Funds of the Second Hospital & Clinical Medical School, Lanzhou University (No. 2025-23-zdfy-004, to F.L.). The authors have no other conflicts of interest to declare.
Data Sharing Statement
Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0557/dss
References
- 1.Ricchi B, Jagana V, Singh E, et al. Advances in diagnosis and patient profiling in pulmonary arterial hypertension for precision medicine. Ther Adv Respir Dis 2025;19:17534666251367312. 10.1177/17534666251367312 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Adu-Amankwaah J, Shi Y, Song H, et al. Signaling pathways and targeted therapy for pulmonary hypertension. Signal Transduct Target Ther 2025;10:207. 10.1038/s41392-025-02287-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Manzi G, Badagliacca R, D'Alto M, et al. 2025. expert consensus on the use of parenteral prostanoids in incident and prevalent PAH patients: The Italian perspective. Vascul Pharmacol 2026. [Epub ahead of print]. doi: . 10.1016/j.vph.2026.107590 [DOI] [PubMed] [Google Scholar]
- 4.D'Alto M, Scelsi L, Giuliani L, et al. Real-world observational study on pulmonary arterial hypertension: Italian cohort treated with macitentan and/or selexipag as a part of a combination treatment (INSPECTIO). Vascul Pharmacol 2026;162:107585. 10.1016/j.vph.2026.107585 [DOI] [PubMed] [Google Scholar]
- 5.Swietlik EM, Fay M, Morrell NW. Exploring Diagnostic and Therapeutic Odyssey in Pulmonary Arterial Hypertension: Insights from In-Depth Semi-Structured Interviews. Respiration 2025;104:26-39. 10.1159/000540556 [DOI] [PubMed] [Google Scholar]
- 6.Khanmohammadi M, Mirzaalikhan Y, Danish H, et al. Biomechanic regulation of neutrophil extracellular traps in the cardiovascular system. Trends Immunol 2025;46:690-703. 10.1016/j.it.2025.08.001 [DOI] [PubMed] [Google Scholar]
- 7.Shen J, Lin H, Mo K, et al. Bidirectional roles of neutrophil extracellular traps in oral microbiota carcinogenesis: A systematic review. Transl Oncol 2025;56:102361. 10.1016/j.tranon.2025.102361 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhang J, Miao C, Zhang H. Targeting neutrophil extracellular traps in cancer progression and metastasis. Theranostics 2025;15:5846-69. 10.7150/thno.111096 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Liu Y, Guo Y, Wu X, et al. Neutrophil Extracellular Traps: Potential Therapeutic Targets of Traditional Chinese Medicine and Natural Products for Cardiovascular Diseases. Pharmaceuticals (Basel) 2026;19:183. 10.3390/ph19010183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sheng Y, Min F, Wang Y, et al. Crocin inhibits neutrophil migration and activation to treat hypoxic pulmonary hypertension through targeting HCK. Phytomedicine 2025;148:157334. 10.1016/j.phymed.2025.157334 [DOI] [PubMed] [Google Scholar]
- 11.Wu J, Zhang F, Zheng X, et al. Identification of renal ischemia reperfusion injury subtypes and predictive strategies for delayed graft function and graft survival based on neutrophil extracellular trap-related genes. Front Immunol 2022;13:1047367. 10.3389/fimmu.2022.1047367 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Janevska M, Witkowski W, Vermeire J, et al. Impairment of endocytosis-related factors FNBP1L, ARHGAP24, and ATP6V1B1 increases HIV-1 entry into dendritic cells. J Virol 2025;99:e0206624. 10.1128/jvi.02066-24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Li H, Zhang J, Sun Y. Insights into transcriptomic changes in blood of a mouse model of LPS-induced peritonitis. Toxicol Appl Pharmacol 2026;506:117623. 10.1016/j.taap.2025.117623 [DOI] [PubMed] [Google Scholar]
- 14.Chu D, Schwartz N, Ampudia J, et al. Soluble urine activated leukocyte cell adhesion molecule is a strong predictor of lupus nephritis. Rheumatology (Oxford) 2025;64:2676-87. 10.1093/rheumatology/keae559 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fang Y, Wang J, Liu C, et al. Targeting P2Y14R alleviates platelet-induced NET formation and venous thrombosis through PKA/AKAP13/RhoA axis. Eur Heart J 2026;47:1982-97. 10.1093/eurheartj/ehaf836 [DOI] [PubMed] [Google Scholar]
- 16.Zhang C, Li H, Gao Y, et al. Targeting the FBXW7-AKAP13-YAP axis suppresses gliomagenesis by dual inhibition of tumor invasion and macrophage recruitment. Biochem Pharmacol 2026;246:117677. 10.1016/j.bcp.2026.117677 [DOI] [PubMed] [Google Scholar]
- 17.Wolff JH, Skov TW, Haslund D, et al. Targeted gene editing and near-universal cDNA insertion of CYBA and CYBB as a treatment for chronic granulomatous disease. Nat Commun 2025;16:7475. 10.1038/s41467-025-62738-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Li J, Di Y, Kang X, et al. Integrated Transcriptomic Analysis Leveraging Single-Cell and Bulk RNA Sequencing Data to Uncover Pyroptosis-Related Prognostic Signatures in HCC. J Hepatocell Carcinoma 2025;12:2971-88. 10.2147/JHC.S557035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Li Q, Muheteer N, Ayoufu G, et al. Expression and prognostic value of EGF-like domain multiple 6 in uterine corpus endometrial carcinoma and its correlation with immune cell infiltration. Sci Rep 2025;15:22620. 10.1038/s41598-025-07379-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Bai S, Gupta N, Liu V, et al. EGFL6 is a novel HER3 ligand, inducing HER3/integrin heterodimers to induce pERK centrosomal deposition and therapeutic resistance. Neoplasia 2025;70:101246. 10.1016/j.neo.2025.101246 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Guo X, Hou P, Zhang S, et al. Enhancer-Associated LncRNA-ITGA2 Promotes Vascular Remodeling Through ITGA2. Circ Res 2025;136:1610-28. 10.1161/CIRCRESAHA.124.325443 [DOI] [PubMed] [Google Scholar]
- 22.Qi X, Zhou J, Wang P, et al. KLF7-regulated ITGA2 as a therapeutic target for inhibiting oral cancer stem cells. Cell Death Dis 2025;16:354. 10.1038/s41419-025-07689-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhou X, Tian W, Gu S, et al. Microbiome-Immune Interaction in Pulmonary Arterial Hypertension: What Have We Missed? Research (Wash D C) 2025;8:0669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhang X, Cheng L, Ma X, et al. Research progress on the mechanisms of cGAS-STING signaling in immune cell infiltration associated with vascular remodeling. J Transl Med 2025;24:112. 10.1186/s12967-025-07536-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Tobal R, Potjewijd J, van Empel VPM, et al. Vascular Remodeling in Pulmonary Arterial Hypertension: The Potential Involvement of Innate and Adaptive Immunity. Front Med (Lausanne) 2021;8:806899. 10.3389/fmed.2021.806899 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bai Z, Liang J, Nie Y, et al. Unveiling the therapeutic potential of senescence-related IQGAP2 in pancreatic Cancer through post-GWAS genomic and scRNA-seq analyses. Discov Oncol 2025;16:1301. 10.1007/s12672-025-03078-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Song F, Dai Q, Grimm MO, et al. The Antithetic Roles of IQGAP2 and IQGAP3 in Cancers. Cancers (Basel) 2023;15:1115. 10.3390/cancers15041115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Deng ZH, Chen YX, Xue-Gao, et al. Mesenchymal stem cell-derived exosomes ameliorate hypoxic pulmonary hypertension by inhibiting the Hsp90aa1/ERK/pERK pathway. Biochem Pharmacol 2024;226:116382. 10.1016/j.bcp.2024.116382 [DOI] [PubMed] [Google Scholar]
- 29.Wang X, Guan X, Zhu X, et al. CircNAP1L4 regulates pulmonary artery smooth muscle cell proliferation via the NAP1L4-mediated super-enhancer-driven glycolysis gene hexokinase II (HK II) in pulmonary hypertension. FASEB J 2024;38:e23868. 10.1096/fj.202400585RRR [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zhang C, Sun Y, Guo Y, et al. JMJD1C promotes smooth muscle cell proliferation by activating glycolysis in pulmonary arterial hypertension. Cell Death Discov 2023;9:98. 10.1038/s41420-023-01390-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kasavi C. Idiopathic Pulmonary Arterial Hypertension: Network-Based Integration of Multi-Omics Data Reveals New Molecular Signatures and Candidate Drugs. OMICS 2023;27:315-26. 10.1089/omi.2023.0066 [DOI] [PubMed] [Google Scholar]
- 32.Kinoshita D, Shishido T, Takahashi T, et al. Growth Factor Midkine Aggravates Pulmonary Arterial Hypertension via Surface Nucleolin. Sci Rep 2020;10:10345. 10.1038/s41598-020-67217-w [DOI] [PMC free article] [PubMed] [Google Scholar]