COPD led to 3.2 million deaths globally in 2015(20). In 2017, the incidence of COPD was about 3.9% and caused a mortality of 41.9 per 100,000 individuals around the world(21). In the United States, there are 12% of adults (30 million) have COPD and it is the fourth leading cause of death (126,000 deaths per year)(22). Moreover, COPD prevalence was higher in Chinese men (11.9%) than in women (5.4%), and 13.7% of Chinese aged 40 years suffer from COPD(23). Neutrophils, macrophages, lymphocyte subsets, and dendritic cells, both types of cells are immune-infiltrating cells and play a critical role in the airway inflammation and lung destruction of COPD(24). Ferroptosis is a new method of cell death, which positively participates in the inflammation and immune responses(25, 26). In the study, we aim to investigate the regulatory role of ferroptosis-related genes between the immune infiltration cells and DEGs in COPD, resulting in screening the potential hub FRDEGs for the treatment of COPD.
There is no research about incorporating the lung tissues and peripheral blood datasets together to COPD through bioinformatics analysis. Therefore, We downloaded a total of 10 datasets consisting of 5 COPD peripheral blood datasets and 5 COPD lung tissue datasets. We obtained differentially expressed genes from the merged dataset after removing the batch effect. According to these DEGs, we identified the potential drugs for COPD through the CMAP database. The results of the action mode of the top 50 compounds implied the co-action mechanism of oxprenolol, naftopidil, and labetalol via adrenergic receptor antagonist, whereas SR-57227A and mosapride interacted via serotonin receptor agonist, The interaction mechanism for AMG-232 via MDM inhibitor restraining ferroptosis related gene-MDM2. There is a need to further understand whether these drug-targeting pathways and treatments are effective in treating COPD, especially MDM inhibitors.
Next, The DEGs combined with ferroptosis-related genes, and validated by WGCNA analyses, 24 FRDEGs were identified in patients with COPD compared with normal controls from the training dataset. Subsequently, functional enrichment analysis of GO terms indicated that FRDEGs were mainly enriched in the oxidation-reduction process, cell death, and immune system. KEGG pathway analysis showed that the FRDEGs were mainly enriched in the p53 signaling pathway. Thus, we speculated that cell death may play an important role in COPD. In addition, 14 out of 24 genes were statistically significant in the merged dataset. We constructed a PPI network of FRDEGs and identified 14 genes with high scores, and 8 genes (MDM2, PIK3CA, CDKN1A, AKR1C3, CREB5, KIF20A, MTDH, and AKR1C2) were statistically significant in the training dataset. LASSO-COX regression and ROC analysis of these DEGs were also performed and finally, we validated 2 hub FRDEGs: MDM2 and CDKN1A. Correspondingly, the COPD dataset consisting of blood and lung tissues was used to confirm the expression of MDM2 and CDKN1A. There was also a significant difference in the level of these two genes in COPD patients compared with that in normal controls. This indicated that the two candidates may mainly participate in COPD by ferroptosis.
To further explore the function of these two critical genes in COPD, the training COPD dataset was divided into a high-expression and a lowexpression group according to the median value of MDM2 and CDKN1A, we then applied GO and KEGG analysis in the DEGs. The enrichment results also suggested that their functions were mostly involved in ferroptosis and immune response. In addition, we further investigated the correlation between MDM2 and CDKN1A expression and infiltrated immune cells in COPD samples. Showing that several kinds of significant immune cells have different performances in the two groups. Macrophages M0, neutrophils and neutrophils, dendritic cells activated, and mast cells activated were positively correlated with MDM2 and CDKN1A expression, respectively. Whereas, T cells follicular helper, T cells regulatory and T cells CD8, T cells CD4 memory resting, macrophages M2, mast cells resting were negatively correlated with MDM2 and CDKN1A expression, respectively. Neutrophils had a positive correlation in both MDM2 and CDKN1A, which indicated that the expression of both MDM2 and CDKN1A was critically involved in neutrophils.
CDKN1A (p21) belongs to a key cell cycle protein and is an important member of cyclin-dependent kinase inhibitors(27, 28). CDKN1A is best known for senescence-associated biomarkers and is transcriptionally controlled by p53 and p53-independent pathways(28, 29). Several recent studies verified that CDKN1A is a facilitator of ferroptosis and has a positive effect on COPD(30, 31). However, the precise role of CDKN1A in the progression and development of COPD is still unrevealed. According to our study, we speculate that CDKN1A promotes the development of COPD by ferroptosis. MDM2 is an E3 ubiquitin ligase that regulates target proteins through ubiquitination(32). The full-length MDM2 (MDM2-FL) is the founding protein of targeting wild-type p53 for degradation by the proteasome, but the functions of the various splice variants of MDM2 are needed to explore(33). Among these results, we found both MDM2 and CDKN1A had a correlation with p53, which was a pivotal target for aging(34, 35). Besides, cellular senescence is now considered a crucial driving mechanism for chronic lung diseases, particularly COPD(36, 37). We speculated that these two high-expressed hub genes in the COPD group may get involved in cell senescence by regulating ferroptosis and thus affect the infiltrating immune cells residing in lung tissues or recruited from peripheral blood. Meanwhile, we found that MDM2 might be an important target gene of CDKN1A. MDM2 and CDKN1A are markedly upregulated in CS and LPS-induced COPD mice.
In addition, as the two hub genes may carry out the dominant role, we further continued to analyze the interaction between the two genes and respiratory tract diseases. The inference scores of CDKN1A in all 17 respiratory tract diseases were higher than those in MDM2 and provided references for CDKN1A as a viable therapeutic target for COPD.
There were obvious limitations in our study. First, the ferroptosis-related regulators were obtained by bioinformatic methods based on GEO and FerrDb databases, which might lead to a high false-positive rate. Second, the in vivo function of these two genes should be verified in more experimental studies.