Benign tracheal stenosis is a common clinical disease in which excessive proliferation of granulation tissue leads to tracheal and bronchial stenosis or complete obstruction due to repeated self-repair after long-term stimulation and injury of tracheal mucosa. The life quality of the patients is seriously affected, which could eventually progress into respiratory failure and death (Sarper et al., 2005). In recent years, with the improvement in surgical technologies and critical care medicine, mechanical ventilation is more often used in clinics, and the incidence of benign tracheal stenosis has increased. Benign tracheal stenosis has many causes, and its pathogenesis mainly results from trauma, injury and inflammation of tracheal mucosa during operation or intubation (Wujtewicz et al., 2009; Bernon et al., 2013). Under these conditions, the fibroblasts begin to proliferate and migrate after induced by inflammatory mediators and growth factors, which increases the production of extracellular matrix and results in the formation of granulation, tissue scar, tissue hyperplasia, and tracheal stenosis (Shitrit et al., 2010; Liang et al., 2015; Kumar et al., 2017; Oberg et al., 2018). However, the molecular mechanism underlying tracheal stenosis has not yet been clarified (Farzanegan et al., 2017). Benign tracheal stenosis is often found after hospitalization, and the delay in diagnosis can affect the treatment efficacy (Shadmehr et al., 2017). Although fiberoptic bronchoscopy can be used for cryotherapy, stent implantation, and other treatments for tracheal stenosis, the recurrence rate and follow-up rate are still high, and there is no better treatment method. The patients with benign airway stenosis had a long survival time and many complications. Therefore, exploring the pathogenic mechanism of tracheal stenosis is of great significance for its early diagnosis and targeted treatment.

The rapid development of gene sequencing technology plays an essential role in the early screening and diagnosis of diseases, as well as the individualized treatment. It greatly improves the speed, efficiency, and sensitivity of detecting circulating free tumor DNA. In addition, high-throughput sequencing technology can accurately detect gene mutations to formulate individualized treatment plan for patients. Also, abnormal genes can be identified through whole genome detection, so that the early intervention treatment can be performed in time (Patch et al., 2015). Through high-throughput sequencing and bioinformatics analysis, it has been reported that about 75% of the human genome can be transcribed into RNA, of which up to 74% are non-protein coding RNA (Kapranov et al., 2007), including long-non-coding RNA (lncRNA), microRNA (miRNA) and circular RNA (circRNA); these RNA species may represent a new disease diagnosis marker and therapeutic target (Beermann et al., 2016; Matsui and Corey, 2017). The regulatory functions of ceRNA provide a new perspective for clarifying the gene expression regulatory network constructed by transcriptome, and add more dimensions for analyzing the molecular mechanism of important biological processes.

In this study, the differentially expressed mRNA, lncRNA and circRNA genes between patients with benign tracheal stenosis and unaffected individuals were explored. And a ceRNA regulatory network was constructed, the main functions and signaling pathways of the key genes were analyzed, and the potential therapeutic drugs were explored. Our results provide a new direction for further investigation on the pathogenesis of benign tracheal stenosis.

In recent years, the respiratory and critical care medicine, respiratory treatment technology, and lung transplantation and other operations have been rapidly developed and improved, resulting in an increased incidence of benign tracheal stenosis (Chen et al., 2021). Although bronchoscopic intervention is a standard treatment for benign tracheal stenosis, it can cause secondary damage to the airway wall structure in some patients, and the excessive granulation tissue and scar contracture formed during the repair process will cause airway obstruction and force patients to undergo intervention therapy again, resulting in a vicious circle of “stenosis-treatment-repair-restenosis” (Jeong et al., 2020). The management of airway stenosis has become a challenge in the field of respiratory interventional therapy. Therefore, it is of great significance to explore the pathogenesis of benign tracheal stenosis and use effective methods to control its development.

In this study, by comparing tracheal stenosis samples with normal control samples, 229 differentially expressed mRNA, 40 differentially expressed lncRNA and 583 differentially expressed circRNA were identified, suggesting that these genes may be involved in the pathogenesis of benign tracheal stenosis. So far, there are few studies on lncRNA and circRNA in benign airway stenosis, and their functions need to be explored. Functional enrichment analysis showed that the differential mRNAs were significantly enriched in biological processes such as neutrophil degranulation, neutrophil activation involved in immune response, neutrophil mediated immunity, and neutrophil activation. It is speculated that the development of benign airway stenosis is related to the activation of neutrophils. It has been found that neutrophils can gather to the wound site within 24 h after injury (Welshhans and Hom, 2017), release inflammatory mediators and cytokines such as TNF- α, IL-1 and IL-6, enhance inflammatory response, and stimulate vascular endothelial growth factor and IL-8 production, which all serve to facilitate injury repair (Reinke and Sorg, 2012). However, neutrophils also produce a large amount of oxygen free radicals, which can cause oxidative stress in wound tissue, delay the repair process and change the healing result (Koh and DiPietro, 2011). It has been reported that after the operation on New Zealand rabbits, the tissue wound was accompanied by a large number of inflammatory cell infiltration such as neutrophils and macrophages (Zhang et al., 2021).

Protein-protein interaction network is a network with proteins as nodes, and the edges indicates that the proteins participate in the same metabolic pathway, biological process, structural complex, functional correlation, or have physical contact. Proteins interact with each other to regulate various life processes such as biological signal transduction, gene expression regulation, energy and substance metabolism, and cell cycle (Wang et al., 2014). In this study, the PPI network of 229 differential mRNAs was constructed, and the interaction relationships of 160 proteins was obtained. The upregulated genes such as matrix metalloproteinase (MMP9), calprotectin S100A9/12/8, serum haptoglobin (HP) and camp were located in the center of the network, suggesting that these genes were involved in disease development. 46 key genes were obtained by further screening. The key genes were significantly related to cellular immune response, HIF-1 signaling pathway, erythropoietin triggering signaling pathway in endothelial cells, neuroinflammatory pathway, wound healing signaling pathway and interleukin regulation. It has been found that IL-8 can not only promote the gap junction communication between fibroblasts, accelerate the maturation of granulation tissue (Moyer et al., 2002), but also promote the proliferation of fibroblasts and change the morphology of fibroblasts (Lim et al., 2009). As a chemokine, IL-8 can participate in the inflammatory response, induce the aggregation of polymorphonuclear leukocytes (Lira and Furtado, 2012; Puyo and Dahms, 2012), and participate in the activation of leukocytes in the wound healing process (Mukaida et al., 1998). Through co-expression network analysis and construction of ce-RNA network, a ce-RNA network of “lnc-RNA- miRNA-mRNA” composed of 8 lnc-RNAs, 31 mRNAs and 116 miRNAs, and a ce-RNA network of “circRNA- miRNA-mRNA” composed of 30 circRNAs, 36 mRNAs and 275 miRNAs were obtained. Some drugs have been found to inhibit cell cycle progression and promote the apoptosis of primary human airway granulation fibroblasts through Mir143Hg (LNCRNA)/Mir-1275 (Mirna)/ILK axis, thus effectively inhibiting the proliferation of airway granulation tissue and ultimately alleviating airway stenosis, which is a potential therapy for benign airway stenosis (Zhang et al., 2021). This study can provide more theoretical support for similar regulatory pathways and therapeutic targets.

Hemoglobin A2(HbA2), calcium binding protein S100A, matrix metalloproteinase (MMP) and growth differentiation factor-related serum protein (WFIKKN1) are the key genes identified in this study, and they are important nodes in each relationship network, suggesting that these genes play an important role in the pathogenesis of benign airway stenosis. Patients with benign airway stenosis are often accompanied by edema and congestion of airway mucosa in pharynx, or clinical symptoms such as congestion, bleeding, erosion or scar formation of trachea and bronchial mucosa; moreover, hemoglobin A2 level is often increased in these patients (Takaishi et al., 2021). Therefore, the abnormal level of hemoglobin A2 is one of the important features of benign airway stenosis. Calcium binding protein S100A participates in innate immunity and plays a role in recruiting leukocytes and neutrophils in inflammatory tissues (Tong et al., 2014). The first stage of airway injury repair is inflammatory response: neutrophils, macrophages and other inflammatory cells infiltrate and release chemokines. It has been proposed that calcium binding protein S100A may be involved in the repair of airway injury. In addition, calcium binding protein S100A interacts with extracellular matrix proteins to stimulate fibroblast proliferation and regulate leukocyte metastasis (Zhong et al., 2016). During the repair process of airway injury, granulation tissue is fibrosed, new epithelium is formed, matrix is remodeled, and the repair of injury is completed (Stunova and Vistejnova, 2018). Low expression of lncRNA LINC00665 inhibits the invasion and proliferation of bladder cancer by regulating S100A13. It has been found that S100A9 is involved in idiopathic non-fibrotic process and can be used as an index for disease evaluation and monitoring (Korthagen et al., 2010). The expression of S100A9 was significantly upregulated in our study, but whether it functioned as previously reported needs further investigation. It has been shown that matrix metalloproteinase MMP9 can effectively degrade the main components of extracellular matrix and play an important role in the process of cancer cell abscission, invasion and metastasis. The activation of MMP-9 is responsible for the degradation of tight junction protein and the opening of tight junction, resulting in increased endothelial cell permeability (Liu et al., 2018). Sevoflurane treatment can relieve neuron damage by inhibiting the expression of MMP-9 in microglia after spinal cord ischemia-reperfusion injury in rats (Li et al., 2014). In our study, the expression of MMP9 was up-regulated, and its protein level needs to be measured to explore its function. Growth factor-related serum protein (WFIKKN1) is a myostatin and growth differentiation factor inhibitor related to muscle growth (Szláma et al., 2013). The WFIKKN expression is increased in human serum during resistance training, as well as during the growth of rat longitudinal muscle (Aoki et al., 2009). In our study, WFIKKN1 expression decreased in benign airway stenosis, which is opposite to what we expected; thus, its protein level needs to be measured and its functions need to be explored in future studies.

In animal cells, HIF (hypoxia-inducible factor) and its regulator pVHL (VonHippel-Lindau tumor suppressor protein) were involved in the perception and response of oxygen. In the presence of hypoxia, the HIF-1α subunit could not be recognized by pVHL. The dimer was formed and transferred to the nucleus, interacted with the cofactor CBP/p300 and Pol II complex, and bound with HRE (hypoxia response element) to activate the transcription of target genes. In addition, airway stenosis was closely related to mechanical compression. Under the compression of the air bag, the tracheal mucosa was ischemia and hypoxia, and the synthesis of hypoxia-inducible factor (HIF-1α) promoted the overexpression of TGF-β and VEGF. Fibroblasts activated and proliferated excessively, which resulted in airway scarring (Cai et al., 2013).

At present, the treatment of benign airway stenosis mainly includes surgery, intervention therapy, and drug therapy. In this study, the potential therapeutic drugs targeting the 46 key genes were analyzed, the results showed that the key gene PPAR γ had the most targeting drugs, indicating that this gene may play an important role in the development of benign airway stenosis. PPARs belonged to the nuclear hormone receptor superfamily and it consisted of three members, PPAR-α, PPAR-β/δ, and PPAR-γ (Han et al., 2017). The function of PPARs was mainly through the interaction of ligands with co-activators or co-repressors (Brunmeir and Xu, 2018). Carvalho et al. (2021). found that PPAR γ suppressed the gene expression of inflammatory mediators by trans-repressing other transcription factors, thereby controlling the exacerbated inflammation that occured in lung injury/acute respiratory distress. In recent years, with the gradual deepening of the research on PPAR-γ, it had been found that PPAR-γ played an important role in the pathogenesis of chronic respiratory diseases, such as asthma, chronic obstructive pulmonary disease (COPD), lung cancer and pulmonary fibrosis. The application of PPAR-γ agonists was expected to become a new method for the treatment of chronic diseases of the respiratory system (Zhang et al., 2019; Pan et al., 2021; Xu et al., 2021). It has been proposed that PPAR γ may be responsible for initiating the endogenous mechanism of wound repair and controlling fibrotic response by activating PPAR γ through its natural ligand (Wu et al., 2009; Wei et al., 2010). Therefore, PPAR γ may be a potential therapeutic target for treating benign airway stenosis.

In this study, the expression of benign airway stenosis and control transcripts was rapidly obtained by whole transcriptome sequencing. Compared with conventional transcriptome sequencing, whole transcriptome sequencing can simultaneously obtain three types of RNA information: mRNA, lncRNA and circRNA. So far, no whole transcriptome sequencing articles related to this disease have been consulted. The innovation of this study was sequenced using its own samples, and to explore potential therapeutic drugs for key gene-related diseases. This study constructed a ceRNA network through bioinformatics methods, analyzed the main functions of key genes, and predicted potential therapeutic drugs, which provided a certain reference for the pathogenesis of benign tracheal stenosis, but this study lacked in vitro and in vivo experimental verification. We would further explore its mechanism of action and verify the function of key genes in follow-up clinical experiments.

In conclusion, our study analyzed the differential mRNA, lncRNA and circRNA between patients with benign airway stenosis and normal controls. Two ceRNA networks were constructed, which provide a new perspective for further exploring the pathogenesis of benign airway stenosis. In addition, the discovery of potential drugs for treating benign airway stenosis, opening up a new direction for the drug development of this disease.