LUAD is a solid tumor that mainly originates from the distal airway with high tumor heterogeneity[36, 37]. In the current era of precision medicine, there is an urgent need to establish a more accurate method to evaluate prognosis and guide the treatment of LUAD patients. Previous data has shown that modification of m7G may be involved in translational regulation of cancer development[38]. However, the cause of dysfunctional m7G regulator remains unclear. miRNAs typically consist of 21–23 nucleotides, are evolutionarily conserved endogenous noncoding RNAs and are vital modulators of gene expression[27]. A recent study has demonstrated that the m7G modification of miRNA inhibits proliferation migration in lung cancer cell lines[28]. Thus, we constructed a prognostic model of LUAD, which combined m7G-related miRNAs signature. A total of 321 m7G-related DEMs were identified in LUAD patients based on the TCGA data. Subsequently, we constructed a prediction model of 28 m7G-related miRNAs signature, contributing to better LUAD survival prediction. Besides, to better understand the immune state and provide a new perspective for clinical treatment, the immune cell infiltration and immune function of the high-risk and low-risk subgroups were further evaluated. Our data, for the first time, revealed the relationship between the m7G-related miRNAs signature and TME, and was helpful in predicting prognosis in patients with LUAD.
Numerous studies have demonstrated the role of miRNAs as biomarkers for carcinogenesis, tumor suppression, diagnosis, and prognosis in LUAD[39]. In this study, a total of 28 miRNAs were included in the risk model, and both the ROC curve and survival curve suggested that the model had a good degree of discrimination. Among them, 8 m7G-related miRNAs were shown to be associated with NSCLC. A previous study showed that knockdown of NEAT1 promotes apoptosis by sponging miR-153-3p, thereby inhibiting cell proliferation, migration, and invasion in NSCLC[40]. miRNA-32-5p has been reported to inhibit epithelial-mesenchymal transition (EMT) and metastasis in LUAD by targeting SMAD family 3 (SMAD3)[41]. In addition, the expression of miR-890 was negatively regulated by small nucleolar host gene 3 (SNHG3). Notably, SNHG3 was found to promote the progression of LUAD by targeting miR-890[42]. It has also been reported that suppression of hsa_circ_0000729 could induce pyroptosis and tumorigenesis in NSCLC cells by targeting miR-1281/FOXO3[43]. Furthermore, let-7f-1-3p may act as a suppressor gene targeting integrin β1 and enhance doxorubicin's inhibition on lung cancer cell viability in vitro[44]. Chuang Li et al. showed that miR-665 was significantly up-regulated in NSCLC[45]. Consequently, exosomal miR-665 can regulate the expression of HEY-like protein (HEYL), a downstream transcription factor of Notch pathway and promote lung cancer cell invasion and migration[46]. Moreover, up-regulation of miR-579-3p was shown to fuel NSCLC cell proliferation[47]. The other m7G-related miRNAs such as hsa-miR-548t-5p[48], hsa-miR-3922-5p[49], hsa-miR-490-5p[50], hsa-miR-4666a-5p[51], hsa-miR-383-3p[52], hsa-miR-6795-5p[53], hsa-miR-4476[54], hsa-miR-6825-5p[55] and hsa-miR-4665-3p[56] were shown to play important roles in other tumors. Our findings demonstrated that these m7G-related miRNAs are crucial in the occurrence and development of LUAD. However, the mechanisms of these m7G-related miRNAs in LUAD are still unclear and need further exploration. Besides, their relationship with LUAD prognosis need urgent verification in large clinical samples.
Immunotherapy strategies targeting immune checkpoint proteins, such as programmed death-1 (PD-1) and programmed death ligand-1 (PD-L1), have transformed the treatment paradigm in LUAD over the past decade[57]. Although these results are encouraging, increased immune tolerance is frequently documented in patients with LUAD[58]. Mechanistically, this phenomenon may be associated with the fact that developing LUAD dynamically communicates with the surrounding TME, hijacking and evading host immune surveillance[59]. Therefore, there is a need for more research attention on the interaction between LUAD and immune cells within the TME. According to the GSEA analysis, the PD-1 checkpoint pathway was significantly different between the high-risk and low-risk subgroups. This data suggests that the TME may differ between the high-risk and the low-risk subgroups.
The estimated, immune, and stromal scores between the high-risk and low-risk subgroups were then investigated. Our data showed that all three scores were lower in high-risk patients than in the low-risk group. Due to the insufficient information on immunotherapy in the TCGA-LUAD cohort, the relationship between the estimated score, immune score, and stromal score as well as immunotherapy response could not be well analyzed. Furthermore, we performed ssGSEA analysis to evaluate the correlation between the subgroups and immune infiltration and function. The results showed that activated CD8+ T cells, effect or memory CD8+ T cells, central memory CD4+ T cells, T follicular helper cells, type 1 T helper cells, regulatory T cells, activated B cells, immature B cells, natural killer cells, myeloid-derived suppressor cells, activated dendritic cells, immature dendritic cells, eosinophils, mast cells, monocytes, and neutrophil infiltrated were lower in the high-risk group, while memory B cell infiltrated were higher in the high-risk group. Among them, neoantigen-driven T follicular helper cells and B cells synergistically promoted the responses of anti-tumor CD8+ T cells in LUAD[60]. In addition, dendritic cells, known as antigen-presenting cells, have been shown to stimulate the differentiation of T cells to eliminate tumor cells[61], while eosinophils in the metastatic TME promote lymphocyte-mediated antitumor immunity[62]. Activated neutrophils have been reported to interact with T cells in two distinct ways. Several studies reported that peripheral blood neutrophils inhibit antigen-non-specific T cell proliferation by releasing argininase-1 and producing ROS[63, 64], while other studies showed that neutrophils can provide antigens and auxiliary signals for T cell activation[65, 66]. Besides, Eruslanov et.al indicated that tumor-associated neutrophils can stimulate the proliferation of T cells and IF-γ release in the early stages of lung cancer[67]. A recent study revealed that tumor-associated mast cells (TAMCs) in NSCLC were a group of heterogeneous population with different subsets of CD103 expression, which need further analysis, especially to understand whether the TAMCs are phenotypically and functionally shaped by growing tumors[68]. Besides, myeloid-derived suppressor cells (MDSCs) are important components of the immune suppressive network and could inhibit host protective antitumor immunity[69]. On the other hand, tumor-infiltrating regulatory T cells have been shown to inhibit the response of endogenous cytotoxic T cells in LUAD[70]. These results demonstrate that immune cell infiltration in the TME of LUAD is an extremely complex and dynamic process.
Consistent with this result, the immune function revealed significant differences in the CCR, checkpoint, HLA, T-cell_co-inhibition, T-cell_co-stimulation, and type_II_IFN_response between the high and low-risk subgroups. Defective HLA-I antigen processing and presentation are involved in acquired resistance toward an immune checkpoint inhibitor in lung cancer[71]. We then analyzed the association of immune checkpoint genes between the high and low-risk subgroups. The results showed that the expression of BTLA, TIGIT, CD28, ICOS, CTLA4, TNFRSF14, CD27, HAVCR2, TNFSF9, CD244, CD48, CD83, CD276 and CD40LG was higher in low-risk group patients with LUAD. CTLA-4 is an immunoglobulin superfamily member receptor, mainly expressing on the surface of activated and regulatory T cells, and inhibiting the initiation, activation and migration of T cells[72]. Although the expression of CTLA4 in NSCLC tumor tissues and cell lines has been reported, its expression in normal bronchial epithelium has not been evaluated[73]. Previous reports showed that anti-CTLA4 antibodies may induce PD-L1 expression in NSCLC with wild-type EGFR and high expression of CTLA4, which enhances the efficacy of anti-PD-1 therapy[74]. According to the KEYNOTE-001 clinical trial, high PD-L1 expression was necessary for using pembrolizumab in NSCLC[75]. These findings indicate that patients in the low-risk group may benefit from combined anti-CTLA4 and anti-PD-1 immunotherapy. CheckMate-227, a large phase 3 trial in metastatic or recurrent NSCLC, showed that patients treated with a combination of nivolumab and ipilimumab had significantly longer progression free survival (PFS) compared to those treated with chemotherapy[76]. However, there is no data on specific biomarkers for the combination of anti-PD-1 and anti-CTLA4 therapy in patients with advanced NSCLC. In addition, macrophage-associated immune checkpoints, CD47 and LILRB1, were highly expressed in the low-risk group. CD47 is a ligand of the negative immune checkpoint regulator signal regulatory protein α (SIRPα), which could trigger macrophage-mediated elimination of relapsed NSCLC cells when targeted. Notably, simultaneous targeting of CD47 and VEGF via VEGFR1-SIRPα fusion protein could induce infiltration of macrophages and sensitize NSCLC to antiangiogenic agents and CD47 blockade[77]. Meanwhile, leukocyte Ig-like receptor B1 (LILRB1) is an immunoreceptor tyrosine-based inhibitory motif-containing receptor that binds to MHC class I molecules[78]. In vitro model studies have shown that blocking the LILRB1 signaling pathway can activate macrophage activity against solid tumors[79]. According to these data, we believe that low-risk patients with LUAD may be able to benefit from LILRB1-targeted therapy. Consistent with these results, Amira A. Barkal et al. demonstrated that concurrent intercepting of the MHC class I-LILRB1 signaling axis may fuel macrophages to eliminate tumor cells and indirectly promote the functions of other immune cells[79].
We further explored the drug response of patients with LUAD between the high-risk and low-risk subgroups. The data suggested that related signaling pathways such as TKI pathway, HDAC pathway, PI3K/AKT/mTOR pathway and Bcl-2 apoptosis may be involved in regulating m7G-related miRNAs. Clinical trials suggest that afatinib is active in NSCLC tumors harboring specific uncommon EGFR mutations, which include Leu861Gln, Gly719Xaa and Ser768Ile[80]. A randomized phase II study of pemetrexed/cisplatin with or without axitinib showed that although patients with axitinib combinations had non-significant differences in PFS, they exhibited a higher ORR compared to chemotherapy alone in non-squamous NSCLC[81]. A recent phase I trial in NSCLC showed that a combination of avelumab, axitinib, and palbociclib exhibited desribale activity and tolerability in NSCLC[82]. Ibrutinib, an irreversible inhibitor of bruton tyrosine kinase, may be a candidate for the treatment of EGFR-mutated NSCLC, even in erlotinib-resistant tumors[83]. Histone acetylation is one of the post-translational modifications that occurs on DNA-packaging proteins, which often results in increased accessibility of promoter regions and transcription of genes in chromosomal local regions[84]. Many cancers express high levels of HDAC and are more sensitive to HDAC inhibitors[85]. Previous studies have shown that HDAC inhibitors can be used to sensitize EGFR-TKIs in treating NSCLC[86, 87]. On the other hand, it has shown that HDAC inhibitors promote chemokines' expression and enhance T-cell infiltration and response to PD-1-blocking immunotherapy[88]. This data suggests that HDAC can sensitize targeted therapy or immunotherapy in NSCLC. However, there is a need for multi-center clinical trials to validate this finding.
Although our study highlights several exciting findings, data we extracted is obtained from a public database. Therefore, a real-world prospective cohort study may be needed to validate our risk scoring model. Besides, the nomenclature and annotation of the miRNAs in the included studies were not uniform. For instance, some miRNAs were named according to their origin from the 3' or 5' arm, while others refer to the miRNAs based on their relative abundance[89]. In addition, the interactions between these prognostic miRNAs and m7G and the molecular mechanisms in LUAD remain unclear. In-depth analysis of these miRNAs' biological functions may provide a new perspective to further understand the mechanism of carcinogenesis and therapeutic strategies in LUAD.