Despite continuous improvements in medical technologies in recent decades, IPF remains an incurable disease associated with a high mortality rate [39]. However, various breakthroughs in surveillance strategies and treatment decision-making for IPF have been achieved. For instance, clinical, radiographic, and histopathological variables as well as some common comorbidities in IPF may help predict mortality [40]. Gene signatures have become an effective tool for identifying IPF patients with a poor prognosis. In terms of emerging therapeutic strategies, lysophosphatidic acids [41], connective tissue growth factor [42], recombinant human pentraxin 2 [43], and other novel pharmaceuticals have shown encouraging results in clinical trials. The clinical course of IPF is characterized by high variability and unpredictability. Therefore, the investigation of new genetic targets could provide novel insights into potential therapies and treatment strategies for IPF.
In this study, we first extracted 97 plasma cells-related genes from the GSE132771 dataset. Then, gene coexpression networks of the GSE150910 dataset were generated, and 44 hub genes most relevant to the extent of plasma cells infiltration were determined. Moreover, 1937 and 378 DEGs were obtained from the GSE150910 and GSE70866 datasets, respectively. Subsequently, the plasma cells-related genes were combined with hub genes and three groups of DEGs to acquire 30 candidate genes, and univariate Cox regression analysis was used to select eight key genes as model features for inclusion in subsequent analyses. We fitted 101 prognostic models via the 10-fold cross-validation framework in the training cohort and screened for PCRGS including seven genes. Further validation demonstrated that the combination of CoxBoost and Enet (α = 0.7) was considered the optimal model, remarkably simplifying signatures and implying underlying patterns. PCRGS was determined as an independent deleterious indicator of OS in IPF, which demonstrated high precision and stable performance in the validation cohort according to the ROC and C-index analysis. Calibration curves indicated that the observed values were consistent with true values. Meaningfully, PCRGS can effectively distinguish IPF patients from healthy subjects according to their expression level. These findings attest that PCRGS has great potential for clinical application.
Notably, PCRGS worked independently of traditional risk factors (e.g. age, gender, and GAP score) in assessing the clinical outcomes of IPF patients and likewise had considerably better prognostic efficacy in accordance with the C-index evaluation. Additionally, 22 published signatures with diverse biological functions were retrieved, yet few of them have been applied in clinical practice. In addition, they displayed superior performance in the training cohort but performed weakly in the validation cohort when compared with PCRGS, suggesting their poor robustness and generalizability. It was conceivable that the hybrid signature selection approach ensured the stability and applicability of PCRGS.
Interestingly, ST6GAL1 may be a more crucial gene in PCRGS. ST6GAL1 is a critical sialyltransferase enzyme located within the ER-Golgi complex that dominates the addition of α2,6-linked sialic acids to the terminal galactose of N-glycoproteins that are destined for secretion or the cell surface [44]. Previous studies have reported that ST6GAL1 is an immune modifier [45] that regulates monocyte-macrophage development [46] and immunoglobulin production in B cells [47], and shifts IgG to the anti-inflammatory pattern by sialylation [48]. It is noteworthy that specific anti-inflammatory salivary IgG Fcs bind to the DC-SIGN receptor on DC-SIGN+ macrophages, activating the Raf-1/ATF3 pathway [49]. ATF3 protein over-expression activates the Wnt/β-catenin signaling pathway to polarize macrophages toward an M2 phenotype [50]. In the IPF lung, M2-like macrophages could produce pro-fibrotic mediators, such as transforming growth factor-β (TGF-β) and platelet-derived growth factor, which induce sustained activation of fibroblasts and promote myofibroblast proliferation [51]. Another potential mechanism by which ST6GAL1 promotes IPF development may be activating the AKT pathway through mediating sialylation modification of β1-integrin [52]. AKT regulates macrophage-derived TGF-β expression by increasing mitophagy, leading to apoptosis resistance, myofibroblast differentiation and fibrosis development in mice [53]. Furthermore, ST6GAL1 was previously reported to promote TGF-β-mediated EMT in the liver [54], which is also a typical pathological process of lung fibrosis. In our study, ST6GAL1 was upregulated in the lung tissues of IPF patients and BLM-injured mice. To sum up, ST6GAL1 may participate in the pathological process of IPF through the aforementioned mechanisms and may act as a potential therapeutic target for IPF.
To predict potentially effective therapeutic agents for IPF, we applied the HERB database to filter out three chemical components related to the ST6GAL1 target, namely quercetin, 17-β-estradiol and mannose-b. Subsequently, a molecular docking procedure was performed to simulate the patterns of interactions between the above agents and the ST6GAL1 protein, which showed that ST6GAL1 and quercetin might have a more stable binding conformation. Quercetin, as a flavonoid alcohol compound with various biological activities, has shown therapeutic potential for PF treatment in several studies, associated with dasatinib [55, 56] or alone [57, 58]. For example, quercetin could enhance the expression of FasL receptor and caveolin-1 and inhibit AKT activation, thus reversing the resistance to death ligand-induced apoptosis and mitigating the progression of established PF [57]. Meanwhile, the effectiveness of quercetin in diminishing the proinflammatory expression of the senescence-associated secretory phenotype was also observed in a model of BLM-induced senescence in fibroblasts in vitro [58]. As ST6GAL1 is highly expressed in plasma cells, quercetin may also have a unique influence in regulating plasma cells effects. Our findings may provide a novel theoretical basis for the pharmacological activity of quercetin in IPF treatment.
Notably, a Korean epidemiological study reported a relationship between IPF and a diversity of cancers [59]. First, IPF has a relationship with the risk of overall cancer incidence, including lung cancer, lymphoma, skin cancer, uterine cervical cancer, multiple myeloma, thyroid cancer, leukemia, pancreatic cancer, liver cancer and prostate cancer. Second, some pathogenic similarities exist between IPF and cancer [60]. Third, IPF shares common risk factors with cancer, such as cigarette smoking, environmental exposure, microbial agents, gastro-oesophageal reflux and viral infection [61]. Thus, we further explored the role of ST6GAL1 across cancers.
In our study, up-regulation of ST6GAL1 was observed in various malignancies. Moreover, we found that a high expression level of ST6GAL1 was associated with poor outcomes in LAML, LGG, ACC, and GBML. A large amount of research has determined that ST6GAL1 impacts cancer hallmarks. For instance, ST6GAL1 promotes the proliferation, migration and invasion of non-small cell lung cancer cells [62]; mitigates the growth and invasive behavior of breast tumor cells in 3D in vitro cultures [63]; mediates the up-regulation of Fas α2,6 sialylation contributing to colon carcinoma cells apoptosis resistance [64]; and increases the capacity of the brain tumor-initiating cells in vitro growth, self-renewal, as well as tumorigenicity by regulating PDGFRB α2,6 sialylation [65]. It is apparent that ST6GAL1-mediated α2,6 sialylation shows a significant effect on cancer pathobiology, which is not only a driver of malignant progression but also a poor prognostic signature for some cancers [66–69]. The exact mechanisms by which ST6GAL1 facilitates oncogenic behaviors are still under investigation. Nonetheless, some critical tumorigenic pathways involving PI3K/AKT [70], Wnt/β-catenin [71], Notch/Hes1/MMPs [62] and targets for ST6GAL1-mediated sialylation, such as TGF-β [54], epidermal growth factor receptor [72], vascular endothelial growth factor [73], tumor necrosis factor [74] and platelet and endothelial cell adhesion molecule [75] have been identified. Furthermore, pan-cancer analysis based on immune infiltration levels revealed a strong connection between ST6GAL1 and immune cells, especially B cells, in various tumors, which was similarly observed in IPF. Therefore, targeted therapy for ST6GAL1 might alleviate both tumors and IPF, and improve the prognosis of patients with combined cancer and IPF.
The present study does have some limitations that need to be mentioned. First, all included datasets were derived from Western retrospective studies, and in the future, PCRGS should be validated in a multicenter longitudinal study, particularly in the Chinese population cohort. Second, the size of the public cohorts we used was relatively smaller than that of compared with the larger cohorts of other diseases, and the inclusion criteria may lead to some undesirable deviations. Third, although we validated the protein expression and cellular localization of ST6GAL1 in BLM-treated mice, due to the mismatched sample types in computer-aided design, more experiments in vitro or in vivo via human samples are warranted to confirm our results. Finally, ST6GAL1 has not been reported in IPF. The mechanism of its roles in IPF needs to be further explored in basic experiments.