Autophagy plays a pivotal role in maintaining cellular homeostasis and survival (Kundu and Thompson, 2008; Kota et al., 2017). Three types of autophagy have been identified: macroautophagy, microautophagy, and chaperone-mediated autophagy. Among them, macroautophagy is the most prominent form of autophagy and has been extensively studied (Patel et al., 2013). Macroautophagy refers to an evolutionarily conserved pathway that operates via lysosome-dependent degradation. The process involves a complicated molecular mechanism that includes more than 30 autophagy-related proteins encoded by autophagy-related genes (Atgs) to ensure that the whole autophagy process proceeds in an orderly manner (Kanthasamy et al., 2006; Komatsu and Ichimura, 2010).

Mounting research has revealed the role of autophagy in pulmonary diseases, such as chronic obstructive pulmonary disease (COPD), pulmonary fibrosis, pulmonary hypertension, cystic fibrosis, and airway remodeling processes (Ryter et al., 2012; Kota et al., 2017). After identifying the role of autophagy in these diseases, several drugs targeting the autophagy process have been developed as therapeutic strategies in the treatment of related diseases and applied in clinical trials (Margaritopoulos et al., 2013). Importantly, whether autophagy represents a target for pulmonary fibrosis therapy is of great interest.

Idiopathic pulmonary fibrosis (IPF) is a kind of progressive and ultimately fatal lung disease that affects more than five million patients worldwide (Chakraborty et al., 2014). Many biomechanical and biochemical factors contribute to the development of IPF, such as aberrant activation of myofibroblasts, injury of the pulmonary epithelium or endothelium, abnormal tissue repair, dysregulation of wound-healing pathways, and accumulation of the extracellular matrix (ECM) (Mora et al., 2017; Richeldi et al., 2017). The median survival of patients diagnosed with IPF is approximately 3 years (Ley et al., 2011). Unfortunately, effective therapies are very limited (Richeldi et al., 2017).

The role of autophagy in IPF is controversial, and decreased or increased expression has been reported in lung tissue obtained from patients with IPF (Patel et al., 2012; Kota et al., 2017). Such inconsistent results were also noted in animal studies with a bleomycin-induced pulmonary fibrosis model in mice (Gui et al., 2015; Zhao et al., 2020a).

Intriguingly, our previous research also found compromised autophagy in mice with an aberrant activation of mTOR in alveolar epithelial cells, which had the potential to develop into more severe pulmonary fibrosis. However, enhancement of autophagy via blocking mTOR with rapamycin protected mice from pulmonary fibrosis (Gui et al., 2015). To explore the relationship between autophagy and pulmonary fibrosis, we explored the dynamic changes in autophagy in different phases of a bleomycin-induced mouse fibrosis model. Additionally, we also identified the differentially expressed genes and signaling pathways in different stages of bleomycin-induced mouse fibrosis by transcriptomic sequencing.

Intratracheal instillation of bleomycin is commonly used for studying the pathogenesis of pulmonary fibrosis because of the strong fibrosis process response. This model is based on the initiation of the inflammatory process followed by the conversion to fibrosis. At the beginning of the first 7 days post-bleomycin, a strong inflammation was induced in response to the lung injury. The fibrosis formation can be observed at day 14 and reached its plateau period between day 21 and day 28 in single-dose intratracheal administration of bleomycin, followed by the resolution of the fibrosis within 6 weeks (Moore and Hogaboam, 2008; Degryse et al., 2010). Hence, the mice treated with bleomycin were divided into five groups to better explore the role of autophagy in the different stages as well as differentially expressed genes and their related functions in the development of pulmonary fibrosis.

To observe the dynamic changes in autophagy markers in the development of a mouse model of bleomycin-induced pulmonary fibrosis, lung samples were collected at the baseline and after 3, 7, 14, 21, and 28 days post-bleomycin. Results from immunoblotting have showed that autophagy marker LC3B and SQSTM1/p62 expressed no difference in all the mice among the groups. In addition, other autophagy-related genes also showed little change in the whole lung tissue of fibrosis. Given that the formation of autophagosomes is the gold standard of autophagy, we then observed autophagosomes by TEM. Previous research studies have observed the autophagosomes in epithelial cells. In a study, the numbers of autophagosomes between IPF patients and the healthy lung had no significance, and autophagy was not induced in IPF patients (Patel et al., 2012). Cabrera S. et al. observed autophagosomes in pulmonary epithelial cells, which they thought that the autophagy gene was highly abundant in the pulmonary epithelial cells (Cabrera et al., 2015). Our observation did not show a significant difference in the number of autophagosomes in the epithelium by TEM. Based on our study, we cautiously concluded that autophagy is probably not critically involved in the development of pulmonary fibrosis.

Moreover, the RNA-Seq results indicated that autophagy-related genes expressed in different stages of fibrosis did not exhibit a significant continuous increase or decrease in expression. Validation by immunoblotting was consistent with the RNA-Seq results. Many genes are involved in the process of macroautophagy and exert a critical role. For instance, ATG4b was reported to be another important protein regulating the formation of autophagosomes by modifying LC3 (Satoo et al., 2009). ATG5 is responsible for autophagic death (Klionsky et al., 2003; Pyo et al., 2005). ATG16 L1, as a binding protein with ATG5, also plays a pivotal role in the process of autophagy (Mizushima et al., 2003). However, we did not find evidence supporting significant increases or decreases in autophagy during the development of fibrosis.

In the literature, autophagy has been reported to be at the low basal level in a mouse model of pulmonary fibrosis induced by bleomycin (Chitra et al., 2015). In human samples from IPF patients, autophagy was not induced in some studies (Patel et al., 2012; Araya et al., 2013). Based on controversial reports from the literature and the current research reported here, we postulate that autophagy may not exert a critical role in the process of fibrosis.

Autophagy is likely not consistently changed, overall, among the total lung tissue samples. Thus, we sought to determine whether variations in autophagy occur in specific cells, such as airway epithelial cells and fibroblasts/myofibroblasts. Recently, single-cell sequencing studies have provided some answers. In two research studies on human IPF, decreased expression of SQSTM1/p62 was shown. Meanwhile, in all three research studies, expression of LC3B and other autophagy-related genes showed no difference among the cell subtype and the disease. As for the results from a research study using bleomycin-induced mouse models, the expression of autophagy seems increased in a short period at the inflammatory stage based on the characteristics of mouse models induced by bleomycin which are dependent on the development of inflammation, while showing little change in the fibrosis plateau period that usually occurs between day 21 and day 28 (Moeller et al., 2008) (Figure 7).

To further understand the mechanisms of pulmonary fibrosis, a dynamic transcriptome profile provided interesting findings. Numerous genes were either upregulated or downregulated in the model of bleomycin-induced pulmonary fibrosis. DEGs in different stages of fibrosis were analyzed. GO terms explained the functions of these DEGs based on the molecular function, cellular component, and biological process categories. We found that in the biological process category, positive regulation of neutrophil degranulation was significantly upregulated at days 3 and 7 post-bleomycin, and the DEGs Ptafr, Cd177, Itgam, and Itgb2 were mainly involved in this process. In bleomycin-induced murine models, the inflammatory process occurred before fibrosis formation. The RNA-Seq results indicate that the conversion from inflammation to fibrosis was largely modulated by neutrophil activation. Intriguingly, among those DEGs involved in neutrophil degranulation, several have already been identified in the development of fibrosis. For example, Ptafr has been found to be significantly dysregulated in the liver (Cai et al., 2006) and cardiac fibrosis (Zhao et al., 2020b). Liang et al. demonstrated that inhibiting Ptafr could block the cell proliferation and secretion of collagen (Liang et al., 2018). Of note, Sala et al. also identified differentially expressed Ptafr in the nasal epithelium of IPF patients (Sala et al., 2018). All these findings of dysregulated Ptafr in animal fibrosis models or in patients could be used to establish the potential role of Ptafr in the development of fibrosis. In addition to elevated neutrophil degranulation, the immune response and chemokine levels were also markedly activated during the inflammatory process induced by bleomycin. A large number of these DEGs consisted of the CXC chemokine family. The CC chemokine family and CC chemokines that bind to CC receptor (CCR) family members do not play a critical role in the recruitment of leukocytes to local inflammation but are important in integrating cellular signal transduction (Starr et al., 2012). For cellular components, DEGs involved in the regulation of the extracellular space were significantly expressed. Some DEGs, such as Tnf, Col3a1, Mmp12, Serpina10, Cd109, Col8a1, Il6, Wnt7b, Fbn1, and Plau, were significantly upregulated in the extracellular space component. Intriguingly, DEGs such as Tff2, Fabp3, Adipoq, Azgp1, Scube2, Cfd, Ces1e, Pgc, Gkn3, Igfbp3, Rnase4, Sema4 g, Nrg4, and S100a7a were downregulated. These findings indicated that the extracellular space changed markedly after intratracheal instillation of bleomycin, and this result provided a molecular basis for bleomycin-induced pulmonary fibrosis in vitro. For annotations of molecular functions, DEGs were mainly enriched in the regulation of cytokine activity and binding, such as in CXC chemokines; CC chemokines and CCR family members were markedly upregulated, and calcium ion binding, troponin C binding, and haptoglobin binding were significantly downregulated. In terms of the role of these differentially expressed chemokines involved in the upregulation of related biological processes and molecular functions, we can infer that bleomycin had a great effect on the expression of chemokines and may enhance fibrosis through these chemokines.

Given that pulmonary fibrosis is a progressive fatal disease and that this pathological process can be mimicked in mice by bleomycin to some extent, dynamic gene pattern expression analyses are also needed to further study the disease and explore possible therapeutic targets.

STEM can reflect the gene expression trend at different times (Ernst and Bar-Joseph, 2006). Through the STEM analysis, two profiles caught our attention: Profile 39, which was a consistent upregulation profile, and Profile 8, which was a continuous downregulation profile. GO annotations revealed that the critical functions of the genes in these profiles were involved in three aspects. For Profile 39, in addition to collagen synthesis, extracellular space–related functions and inflammation-related responses were enriched, which have been well studied as a result of intratracheal instillation of bleomycin, and other annotations were also continually significantly upregulated, such as peptidyl–lysine oxidation. Previous research has reported that lysyl oxidase exerts a central role in collagen maturation (HM, 2000). Aumiller et al. demonstrated that Loxl2 is strongly expressed in human lung fibroblasts and airway epithelial cells after stimulation with TGF-β. In addition, they also found increased expression of Loxl2, Loxl3, and Loxl4 in whole lung tissue from pulmonary fibrosis mouse models using gene expression assays at day 14 post-bleomycin. Consistent with these findings, the expression of Loxl2, Loxl3, and Loxl4 was also increased in the tissue of IPF patients (Aumiller et al., 2017). Similarly, Tjin et al. found that Loxl2 was upregulated in IPF patients (Tjin et al., 2017). Consistent with these results, Loxl2, Loxl3, and Loxl4 were continually upregulated in our research (Supplementary Excel 2). Additionally, Nuggc was also found to be continually increased during the regulation of peptidyl−lysine oxidation. However, to the best of our knowledge, Nuggc has not been previously reported in the regulation of the development of fibrosis and deserves further research.

For the continually decreased Profile 8, annotations were mainly enriched in cardiac functions, such as cardiac muscle organization and contraction. In IPF patients, cardiovascular comorbidities, such as hypertension, coronary heart disease, and heart failure, represent important factors that affect the prognosis and mortality of IPF (Agrawal et al., 2016). Research is limited on the relationship between cardiac muscle function and pulmonary fibrosis. In the current research, sarcomere organization, muscle contraction, and muscle fiber development were continually downregulated post-bleomycin. Related genes, such as Tnnt2, Actc1, and Myh6, were implicated. However, dysregulation of these genes may be the leading factor for the continually decreased cardiac functions and could act as a stimulus for the severity of the disease. For example, mutations of Tnnt2 lead to the loss of sarcomeres and cause the disarrangement of myocytes, which can be a very important life-threatening hallmark for patients (Sehnert et al., 2002). Actc1 is reduced in patients with congenital heart disease, while its downregulation could be responsible for cardiomyocyte apoptosis (Jiang et al., 2010). Downregulation of Myh6 occurred concomitantly with a reduced ejection fraction of systemic right ventricular tissue in patients with hypoplastic left heart syndrome (Theis et al., 2015). In the current research, continually downregulated genes enriched in cardiac functions may be responsible for the poor prognosis and heart complications resulting from the pulmonary fibrosis, which also provides a molecular mechanism insight into the cardiac complications that result from the pulmonary fibrosis. However, whether these continually downregulated genes are directly involved in the regulation of pulmonary fibrosis remains unclear, and further research exploring the mechanism underlying their dysregulation in the context of pulmonary fibrosis is still essential.

In addition, we also explored the KEGG pathways enriched in Profile 39 and Profile 8. For Profile 39, the PI3K-Akt signaling, ECM–receptor interaction, and focal adhesion signaling pathways were continually upregulated. PI3K-Akt plays a critical role in a wide range of cellular processes, such as cell metabolism, proliferation, growth, and protein synthesis (Kuma et al., 2004). Previous research by Lu et al. found that aberrant activation of PI3K-Akt promotes fibroblast proliferation and collagen synthesis (Lu et al., 2010). Similarly, Zhang et al. also demonstrated that PI3K-Akt is involved in the development of pulmonary fibrosis using a bleomycin-induced rat pulmonary fibrosis model, while inhibition of PI3K-Akt could reverse pulmonary fibrosis and epithelial mesenchymal transition (Zhang et al., 2016). In addition, PI3K-Akt is also a bridge that communicates signals to various pathways, and it also plays an important role in pulmonary fibrosis, such as mTOR, VEGF, and TGF-β (Yan et al., 2014). Given the roles of PI3K-Akt in pulmonary fibrosis, research has shown that the PI3K-Akt signaling pathway could be a therapeutic target in IPF patients. Accordingly, a study reported that omipalisib (GSK2126458), an inhibitor of PI3K-mTOR, was under Phase I clinical trial in the treatment of IPF (Lukey et al., 2019). Compared with these studies, continually upregulated PI3K-Akt was found in our research based on a STEM analysis, and it was enriched along with genes that were previously reported in the context of fibrosis, such as collagen1, collagen4, Il7r, Itga11, and Itgb6. Our research is consistent with previous findings, suggesting a critical role of these genes in the upregulation of the PI3K-Akt signaling pathway throughout the development of pulmonary fibrosis. Of note, several genes, such as Reln and Ntrk2, were also enriched in PI3K-Akt. Reln was reported to be related to liver fibrosis (Mansy et al., 2014), and Ntrk2 encodes a tyrosine kinase receptor, TrkB, which is abundantly expressed in lung tissue and responsible for hypoxia, thus promoting the proliferation of pulmonary arterial smooth muscle cells (Kwapiszewska et al., 2012). These results and our findings could provide new insights into the genes involved in pulmonary fibrosis. However, further research is still essential.

For the KEGG analysis of Profile 8, continually downregulated pathways were the cAMP signaling pathway, the cGMP-PKG signaling pathway, the calcium signaling pathway, and cardiac muscle contraction. Enriched genes, such as Actc1, Adrb2, Adcy8, and Chrm1, may correspond to some critical functions. For example, Adrb2 controls the expression of the β2 adrenergic receptor, which belongs to the G-protein-coupled family and plays a critical role in regulating the pulmonary, cardiac, and vascular systems by binding to catecholamines and epinephrine (Litonjua et al., 2010). Adcy8 is known to regulate glucose homeostasis (Raoux et al., 2015). Chrm1 belongs to the muscarinic–cholinergic receptor (CHRM) family and encodes the Chrm1 receptor, which is extensively expressed in the lung and regulates the secretion of electrolytes and water (Jimenez-Morales et al., 2014). Their continuous downregulation may also provide therapeutic insights into new targets for the treatment of pulmonary fibrosis.

However, this research presented certain limitations. We monitored autophagy markers in the whole lung tissue, and it seems that autophagy did not exert a critical role in a continually changing manner, whereas it might play an important role at a certain time in a certain cell type which deserves further study. Given that epithelial cells and fibroblasts are the two main cell types involving pulmonary fibrosis, either monitoring the expression of autophagy-related gene expression or observing the specific structure such as autophagosomes in certain cell types besides these could be meaningful. Furthermore, in the current study we have assessed autophagy in the whole development of fibrosis. It is better to use the autophagic inducer or inhibitor to assess the dynamic change of autophagic flux and further, to evaluate the role of autophagy in the alleviation or aggregation of fibrosis. Also, a specific autophagy-deficient or aberrantly activated animal model for studying pulmonary fibrosis could be more indicative for the related research. Additionally, although bleomycin is widely used for establishing pulmonary fibrosis models, these models are not consistent with those for IPF patients. IPF is a progressive and irreversible disease, but the process is difficult to reproduce completely in a mouse model induced by intratracheal instillation of bleomycin. In the pathology of mouse fibrosis, as the disease progresses, lung pathology tends to recover 28 days after bleomycin treatment. Although many DEGs and pathways were identified by RNA-Seq, it is still difficult to determine the most important genes in the development of IPF since the disease was thought to be a result of multiple factor interactions. The results would be more robust if those genes or pathways were further validated in the patients’ tissue.

In the current research, we monitored autophagy at different stages of pulmonary fibrosis using bleomycin-induced mouse models. We detect autophagy markers through multiple aspects. Additionally, we analyzed autophagy-related genes from public datasets. All results consistently indicate that autophagy did not exhibit a continual expression in the development of fibrosis. We then explored the associated DEGs, including their enriched functions and signaling pathways, based on RNA-Seq. Continually upregulated annotations were related to collagen synthesis, inflammation, and extracellular space. The enriched signaling pathways were mainly related to the PI3K-Akt signaling, ECM–receptor interaction, and focal adhesion signaling pathways. Continually downregulated profiles were mainly related to sarcomere organization, muscle contraction, and muscle fiber development. Enriched KEGG signaling pathways were related to the cAMP signaling, cGMP-PKG signaling, calcium signaling, and cardiac muscle contraction pathways. In summary, autophagy may not be critically involved in the development of pulmonary fibrosis. These continually up- or downregulated profiles, including the associated genes and their functions and pathways, may provide mechanistic insights for future therapies for pulmonary fibrosis.