IPF is a condition that worsens with time, is fatal, has few therapy choices, inadequate therapeutic outcomes, and an unfavorable prognosis. Notwithstanding the booming development of multiple diagnoses and treatment strategies, the survival rate of patients with IPF has not been improved. IPF is characterized by a clinical course that is highly variable and prone to unpredictability; hence, improved risk assessment methodologies and tailored targeted therapy strategies are necessary (32).
In this research, we found two molecular clusters that are considerably distinct from one another premised on the expression of genes associated with FA metabolism. Enrichment analyses revealed different immune and metabolism actions between the two clusters. Further immune analyses identified several immune cells in BALF potentially associated with prognosis in IPF. Additionally, we developed a predictive risk model that was on the basis of FAMRGs and was able to correctly predict the prognosis of patients who were diagnosed with IPF. Our findings could help in creating therapies that specifically target IPF.
The use of consensus clustering as a method for categorizing samples into various groups premised on the gene expression matrix was shown to be reliable. We initially identified two molecular clusters using consensus clustering based on the FAMRGs expression in patients diagnosed with IPF and discovered that these two molecular clusters had remarkably different OS rates. These findings provided further evidence that the FA metabolic subtypes of a patient's IPF influenced their prognoses and that the consequent construction of prediction models premised on FAMRGs was reliable.
Next, enrichment studies between the two clusters were carried out so that the inherent biological processes could be investigated. GSEA is a classic approach for incorporating gene expression data. Using this method, the expression pattern of gene sets in various groups may be elucidated directly. (33). First, GSEA findings illustrated an enrichment of immune-related pathways, including natural killer cell-mediated cytotoxicity, the B cell receptor signaling pathway, the Toll-like receptor signaling pathway, leukocyte transendothelial migration, the chemokine signaling pathway, and the Nod-like receptor signaling pathway, in cluster 1, which indicated poor OS. Based on these findings, it seems that an immune system that is dysregulated may be a potential mediator of the function that FA metabolism plays in the onset and progression of IPF. Mechanistically, the association between FA metabolism and immune dysregulation may be through epigenetic regulation, such as DNA methylation influenced by genetic variation (34). Notably, the results showed that clusters 1 and 2 had specific metabolism signatures. The pathways enriched in cluster 1 were predominantly linked to the metabolism of amino acids, including arginine, proline, and histidine metabolism, and other lipid metabolism pathways, including the metabolism of glycerophospholipid and arachidonic acid. Short-chain FA metabolism pathways, including propanoate metabolism and butanoate metabolism, and other metabolism pathways, including aminoacyl tRNA biosynthesis, riboflavin metabolism, limonene, and pinene degradation were mainly enriched in cluster 2. Considering that the classification was based on FA metabolism relevant genes, the result suggested that there was crosstalk between FA metabolism and other nutrient metabolisms that affected the process of IPF and were worth further exploration.
As mentioned previously, immunology performs an integral function in the onset and progression of IPF, which is also associated with FA metabolism. CIBERSORT is a biological information analysis tool that evaluates the expression levels of immune cells premised on RNA-seq and obtains different immune cell ratios from samples. This algorithm is extensively utilized to examine the infiltration of immune cells in various human immune-related diseases, such as tumors, osteoarthritis, and lupus nephritis (35–37). Li et al. established a hypoxia immune-related prediction model for determining the prognosis among IPF patients using CIBERSORT (38); therefore, we implemented CIBERSORT to ascertain the infiltration levels of immune cells between the two clusters. We identified several BALF immune cells that had the potential to be associated with IPF prognosis, mainly involving activated mast cells, naive B cells, resting mast cells, M0 and M2 macrophages, monocytes, activated NK cells, resting dendritic cells, and resting NK cells. Activated NK cells, activated mast cells, and monocytes showed increased infiltration in cluster 1 patients with worse survival. Similarly, in the constructed risk model, these three immune cell populations also had an elevated infiltration level in the high-risk group with poor prognosis and were positively correlated with Riskscore. Earlier research findings have shown that individuals with pulmonary fibrosis exhibit a greater infiltration level of NK cells in BALF in contrast with those who have sarcoidosis (39). Scott et al. indicated that elevated circulating monocyte count could be a cellular biomarker for poor outcomes in IPF (40). According to the findings of Kawanami et al., individuals with fibrotic lung disease had significantly higher numbers of mast cells in their lungs. These mast cells are often localized around the thickened regions of the alveolar septa and are located near aberrant epithelial cells (41). Taken together, these results revealed that activated mast cells, monocytes, and activated NK cells in the BALF of patients with IPF may promote disease progression.
Synthesizing the findings of this study, it is reasonable to infer that FA metabolism-related dysregulation, resulted in the disorder of immune and metabolism status in BALF, hence contributing to unsatisfactory IPF prognosis. As aforementioned, the reprogramming of the FA metabolism was identified as a distinctive characteristic of a grim prognosis in individuals with IPF. To additionally verify the influence of FA metabolic disorders in IPF and examine the prognostic significance of FAMRGs in patients diagnosed with IPF, we created a prognostic risk model using FAMRGs, and then we confirmed it using a separate validation cohort. It has been shown that the five genes that were employed for the establishment of the risk model in this research are remarkably related to the onset and progression of IPF. GGT5 encodes gamma-glutamyl transferase 5, which cleaves glutathione peptides to maintain the glutathione balance in the human body (42). Prior studies have reported that mice lacking gamma-glutamyl transpeptidase had less severe cases of bleomycin-induced pulmonary fibrosis (43). However, the role of GGT5 in FA metabolism and the progression of IPF is unknown. The branched-chain acyl-CoA oxidase that is encoded by the ACOX2 gene is a peroxisomal enzyme that is thought to play a role in the metabolism of bile acid intermediates as well as branched-chain FAs. A deficiency in ACOX2 has been shown to be linked to an increased risk of developing liver fibrosis (44), but its role in pulmonary fibrosis is unknown. CYP4F3, also referred to as leukotriene-B4 omega-hydroxylase, consists of enzymes CYP4F3A and CYP4F3B that are responsible for the metabolism of leukotriene B4 and most probable 5-hydroxyeicosatetraenoic acid through an omega oxidation reaction. This results in the inhibition and deterioration of well-recognized inflammatory markers (45, 46). Thus, CYP4F3 is considered to be associated with inflammation-related diseases, such as inflammatory bowel disease (IBD) (47), but the role of CYP4F3 in IPF is unknown. HACD4 encodes 3-hydroxyacyl-CoA dehydratase 4, which is involved in FA elongation and the biosynthetic process of very-long-chain FA; therefore, its role in IPF is worth exploring. Although ODC1 is known to encode the rate-limiting enzyme of the polyamine biosynthetic pathways, which acts as a catalyst of ornithine to putrescine, its significance in IPF is only partially recognized. Survival analysis illustrated that the developed risk model showed outstanding prediction capacity for the survival of patients diagnosed with IPF in both cohorts. The patients' prognoses could be determined independently by each of the five genes, and independence and subgroup analyses demonstrated that the FAMRGs-based risk model can independently anticipate the prognosis of IPF, regardless of age and sex. Ultimately, a nomogram that incorporates both the risk score and the clinical characteristics was developed, calibrated, and tested; the results revealed that it had a powerful predictive power regarding survival. Taken together, these findings provided more evidence that FAMRGs have a prognostic predictive role in IPF.
Owing to the poor prognosis, variability, and unpredictability of the clinical course of IPF, there is a need for a method of efficient risk classification as well as a treatment plan that emphasizes personalized targeting. Our research demonstrates highlights in comparison to earlier studies. First, our research was on the basis of FA metabolism, a topic that has been receiving more and more interest in IPF-related research. Using consensus clustering, we discovered two molecular clusters with distinct prognoses and immunological statuses. Second, we examined the biological processes based on the clustering findings, and we partially elucidated the fundamental mechanisms. Thirdly, we elucidated the function that the metabolism of FA plays in influencing the infiltration level of immune cells in BALF as well as the prognosis. Moreover, the role of each of the five selected genes is unclear in IPF, and further exploration is necessary. Our study offers a good theoretical foundation for future IPF research.
Nevertheless, our study also has several limitations. First, we could not establish the involvement of FAMRG in the progression of IPF because there was insufficient data regarding the pulmonary function of the individuals who had IPF. Second, our findings were derived from a bioinformatics study, which was not followed up with further experimentations to confirm its validity. Thirdly, the data that were utilized in this research were obtained by downloading them from a free-access database since our clinical practice had an insufficient number of patients diagnosed with IPF. Consequently, to additionally evaluate the performance of the predictive model and the mechanisms of the five FAMRGs in the pathogenic process behind IPF, a prospective cohort study as well as molecular biology tests ought to be devised and conducted.