Glioblastoma (GBM) is the most common malignant central nervous system (CNS) tumor in adults, and this tumor is very aggressive (Batash et al., 2017; Louis et al., 2021). Even after surgery, radiotherapy, and chemotherapy, the median overall survival (OS) of patients with GBM is approximately 15 months (Johnson et al., 2018; Witthayanuwat et al., 2018). Tumors may continue to grow (progress) and accompany by recurrence after treatment (Weller et al., 2017; Bates et al., 2018; Tan et al., 2018). Although some advances have been realized in the treatment of GBM, such as the application of temozolomide and bevacizumab, patients with GBM still encounter poor treatment effect and drug resistance. (Tan et al., 2017; Jiapaer et al., 2018; Kickingereder et al., 2020; Reardon et al., 2020; Nayak et al., 2021; Tesileanu et al., 2022; Zeng et al., 2022). Tumor therapeutic field (TTF) therapy has also been approved by the FDA for recurrent (2011) and newly diagnosed (2015) GBM, but the lack of generalizability of data from previous studies has prevented TTF from being widely used. (Regev et al., 2021). Therefore, no effective treatment is currently available for GBM, which remains one of the most difficult and complex cancers to treat. Therefore, understanding the complex mechanism of GBM and exploring new therapeutic strategies are urgently required.

Orexin is involved in the interaction between cancer and neurodegenerative diseases such as narcolepsy. (Liu et al., 2015; Boss and Roch, 2017; Wan et al., 2017; Tan et al., 2020). Orexin A and B (OXA and OXB), also known as hypocretin 1 and 2, respectively, are peptides expressed by hypothalamic neurons and are first identified in 1988. (Sakurai et al., 1998). Orexin binds to two G-protein-coupled receptors, namely, orexin receptors 1 and 2 (OR1 and OR2) (Sakurai et al., 1999), which are widely expressed in the CNS (de Lecea et al., 1998), and is consistent with widespread expression of orexin neurons. (Tsujino and Sakurai, 2009). Orexin signaling is multifaceted and complex, with similar mechanisms for OR1 and OR2. A recent in vivo study reported that OXA may increase mitophagy and disrupt mitochondrial structure. (Zhu et al., 2021). In addition, OXA may affect the interaction between brain and gastrointestinal tract by acting on intestinal microorganisms, thereby affecting brain function (Foster et al., 2016). In recent years, a series of studies has focused on modulating orexin-related signaling, which may play a surprising therapeutic role in the treatment of certain types of cancer (Graybill and Weissig, 2017; Mogavero et al., 2021). OXA stimulates neovascularization, a key step in chronic inflammation and tumor growth (Kim et al., 2015). OR1 signaling has a pro-apoptotic role signaling in neuroblastoma cell lines (Rouet-Benzineb et al., 2004). These results were later confirmed in other tumor cells both in vivo and in vitro (Voisin et al., 2011; Wen et al., 2016). In this process, OXA directly activates caspase-3 to promote tumor cell apoptosis in vivo, or this process is possibly mediated by two immunoreceptor tyrosine-based inhibitory motifs in OR1 and OR2, which participate in phosphotyrosine phosphatase SHP2 and induce mitochondrial apoptosis (Mogavero et al., 2021). Studies on the potential modulation of the orexin system in cancer are still pioneering. Although promising results have been obtained, further human data are needed. Therefore, a multi-omics analysis was conducted based on transcriptome, proteome, and metabolome.

In recent years, molecular biology technology has been developed rapidly. Meta-analysis of single-omics datasets is very valuable for biological and medical research (Ni et al., 2014; Hanna et al., 2017; Goveia et al., 2020; Rijkschroeff et al., 2020; Hoogstrate et al., 2022). With the continuous development of deep sequencing and other high-throughput methods, the gradual reduction of cost and the increasing maturity of technology, a large number of omics data can be obtained, thus allowing the measurement of multi-omics data. Multi-omics analysis can reduce the effect of biological and experimental bias in the data, and different omics can reveal different cellular aspects, such as the effects manifested at the genomic and proteomic levels (Subramanian et al., 2020; Zhou et al., 2021). Multi-omics analysis is a comprehensive assessment of multiple sets of characteristics. Specifically, transcriptomics is the study of expressed RNAs; it usually focuses on protein-coding RNA (mRNA) and includes non-coding RNA that coordinate and regulate gene expression, which provides attention to the underlying mechanisms involved in genes and biological processes (de Jong and Bosco, 2021). Proteomics is the study of expressed proteins and is used to describe protein abundance, properties, post-translational modifications, and protein interactions (Aslam et al., 2017). Metabolomics focuses on the analysis of small molecules (i.e., metabolites), including carbohydrates, fatty acids, amino acids, and other compounds (Patti et al., 2012). Transcriptomics, proteomics, and metabolomics are complementary and synergetic for the effective understanding of molecular interactions and disease mechanisms.

In the present study, orexin-induced changes in GBM cells were analyzed at the transcriptomic, proteomic, and metabolomics levels by using a multi-omics approach. We measured multiple omics data, such as transcriptomics, proteomics, and metabolomics, of GBM cells cultured with and without Orexin A, and identified differentially expressed molecules that were significantly altered in each omics layer in orexin-added GBM cells relative to those without OXA, and linked these layers to differential regulation. The correlation of differential molecules was analyzed. Simultaneously, the KEGG pathways enriched by the combination of pairwise omics and the three omics were analyzed, and the differentially expressed molecules and differential regulators in KEGG pathways were extracted. Finally, data from TCGA and CGGA public databases were combined to screen out the molecules with prognostic effect. The underlying mechanism of OXA action in GBM cells was explored.

GBM treatment remains extremely challenging, and its prognosis remains unsatisfactory despite surgery, radiotherapy, chemotherapy, and some innovative therapies. Additionally, limited improvement has been realized in the treatment of GBM. OXA inhibits the growth of rat C6 glioma cells through a caspase-dependent mechanism (Biegańska et al., 2012). Programmed cell death has been induced by sustained orexin receptor stimulation in other malignancies (Smart et al., 1999; Ammoun et al., 2006). Similarly, both OXR subtypes can activate cell death as observed in recombinant CHO-S cells (Nicole et al., 2015), acinar cell adenoma (Kaczmarek et al., 2017), colon cancers, and neuroblastoma (Rouet-Benzineb et al., 2004). However, the specific mechanism of orexin’s action has not been fully understood. Detailed exploration of the mechanism of OXA acting on glioblastoma cells could facilitate the understanding of the disease and promote the discovery of therapeutic targets. The transcriptomic, proteomic, and metabolomics data of GBM in the orexin-treated and control groups have been measured, and significant differentially expressed molecules have been identified at each omics level caused by Orexin A, which could well identify whether GBM cells were treated with OXA or not. Further analysis of the correlations of DEGs, DEPs, and DEMs revealed significant correlations among them. Subsequently, the biomolecules and functional pathways, in which OXA might play a role in GBM, have been analyzed.

In our combined transcriptomic-proteome analysis, proteoglycans (PGs) in cancer, chemical carcinogenesis—ROS, and pathways such as mitophagy-animal and p53 signaling pathway were enriched. PGs play important roles in cancer initiation and progression by regulating cellular metabolism, influencing immune surveillance, acting as sensors of mechanical properties, and participating in resistance to various forms of therapy (Wade et al., 2013; Mouw et al., 2014; Baghy et al., 2016). Aberrant ROS plays contradictory roles at different stages of cancer formation, especially in the regulation of cell growth and death, and further understanding of the complex mechanisms of ROS in tumorigenesis is crucial for conquering cancer (Valavanidis et al., 2013; Wang Y. et al., 2021). Huang T. et al. (2021) found that the inhibition of mitophagy partially reversed cannabidiol-induced glioma cell death, suggesting the positive role of mitophagy against tumors. However, FOXO3A-induced mitophagy protects glioma from temozolomide induced cytotoxicity, suggesting that mitophagy could be considered as a double-edge sword for glioma (He et al., 2021; Hu et al., 2021). OXA reduces mitochondrial biogenesis, enhances mitophagy, and damages mitochondrial structure in AD patients (Zhu et al., 2021). OXA may play the same role in tumor and non-tumor diseases. Therefore, it may promote anti-tumor effect by enhancing mitophagy. The P53 tumor suppressor is a key genetic event associated with disease development and progression, and this factor is mutated or absent in 35% of GBM (Ham et al., 2019). P53 signaling pathway plays an important role in the development of glioma. Therefore, OXA might target p53-related pathways to exert its inhibitory effect on cancer. In addition, the majority of proteins showed higher abundances than the relevant RNA in quadrant 8, suggesting post-transcriptional, translational regulation, or accumulation of proteins. Combined transcriptomic-proteome analysis revealed the relationship between transcriptomics and metabolomes, and the FoxO signaling and PI3K-Akt signaling pathway were enriched. Previous mechanistic studies have found that the downregulation of FoxO/Smad signaling promotes cancer cell proliferation in glioblastoma (Wang et al., 2018). The activation of PI3K/Akt/mTOR pathway enhances the proliferation, migration, and invasion of glioma cells, resulting in the occurrence of drug resistance, thereby inhibiting the therapeutic effect of TMZ (Wu et al., 2016; Xu et al., 2018; Zhao et al., 2021). As a result, targeting these pathways by developing corresponding activators or inhibitors may inhibit tumor development and improve patient treatment outcomes. Combined proteomic and metabolome analysis revealed the relationship between the proteome and metabolome, showing only the enrichment of glyoxylate and dicarboxylate metabolic pathways. Glyoxylate and dicarboxylate metabolism plays an important role in prostate cancer (Chen et al., 2021a), but its role in GBM remains to be further investigated. The significant correlation between the differentially expressed molecules in the transcriptomic-proteome-metabolome confirmed the complex regulatory mechanism of OXA acting on GBM. Consequently, further exploration of this regulatory network will help us understand disease processes and identify therapeutic targets.

In the combined transcriptome-proteome-metabolome analysis, 11 common pathways were identified. In addition to PGs described above in cancer, chemical carcinoma-ROS and mitophagy-animal, P53 signaling pathway, also identified pyrimidine metabolism (PyM). A continuous supply of dNTPs is essential for cancer cell survival. Therefore, the permanent activation of the PyM gene is essential for tumor growth (Siddiqui and Ceppi, 2020). OXA may be associated with inhibiting GBM by downregulating the pyrimidine metabolism pathway. In addition, in the FoxO signaling and PI3K-AKT signaling pathway, more DEGs and DEPs are enriched. OXA increases AKT/PDK-1 phosphorylation through phosphatidylinositol 3-kinase and FOXO-1-dependent pathways (Göncz et al., 2008). Ju et al. (2014) revealed that OXA may affect apoptosis in rat hepatocytes by regulating Foxo1 and mTORC1 through the OX1R/PI3K/AKT signaling pathway. Moreover, ferroptosis has attracted our interest, and drugs that target different molecules involved in ferroptosis and stimulating the ferroptosis process are potential adjuvant anticancer treatment options (Liang et al., 2019; Wang et al., 2019; Xu et al., 2019; Chen et al., 2020). In GBM, ferroptosis stimulation can inhibit tumor growth, improve patient survival, and enhance the efficacy of chemoradiotherapy (Yee et al., 2020; Yang et al., 2021). Therefore, OXA targets molecules in the ferroptosis process and activates the ferroptosis pathway to exert anti-tumor effects. These findings need to be further confirmed.

For a good understanding of the mechanism of OXA action, the genes in 11 pathways were extracted, and 138 genes were selected. Survival analysis of TCGA-GBM data was carried out, and 10 genes (CCL2, UPP1, F2R, ITGA5, SFN, IRS1, CXCL8, MAP1LC3A, MET, and ISG15) out of 138 genes had significant survival differences between high- and low-expression groups. The results were further verified in CGGA data, and the results show that six genes (ITGA5, MET, F2R, CCL2, SFN, and UPP1) had significant survival differences between the high- and low-expression groups. The mechanism of action of these genes has been explored. Chen et al. (2021b) reported that ITGA5 accumulation can activate FAK signaling pathway to promote cell growth. Therefore, NEAT1/Mir-128-3p/ITGA5 axis is involved in the occurrence and progression of gliomas. Shi et al. (2021) suggested that ITGA5 expression is upregulated in glioma cells resistant to TMZ, and the overexpression of ITGA5 could increase the resistance of cells to TMZ by promoting the formation of vascular mimicry. OXA downregulated ITGA transcription, thus serving as a pathway for OXA inhibition of glioma. Invasion-related protein ITGA5 in previous studies may be an effective anti-invasion target, which is correlated with advanced tumor grade, recurrence, and overall survival of GBM. [PMID:25853691 and PMID:34873473] (Mallawaaratchy et al., 2015; Wei et al., 2021) Moreover, Wang Q. W et al. (2021) suggested that the MET-STAT4-PD-L1 axis may act with tumor-associated macrophages to enhance immune escape in gliomas and cause poor prognosis in patients with GBM. And Huang et al. (Huang R. et al., 2021) verified that PTPRZ1-MET (ZM) fusion was a key genetic change that drives the progression of low-grade glioma and helped ZM-carrying glioma patients benefit from MET inhibitors. In addition, the inhibition of receptor tyrosine kinases, including MET and/or its ligand hepatocyte growth factor (HGF), is a promising therapeutic strategy against tumor (Suzuki et al., 2010; Ge et al., 2013; Song et al., 2020). Auvergne et al. (Auvergne et al., 2016) strongly demonstrated the importance of the F2R gene encoding PAR1 for the self-renewal and tumorigenicity of glioma A2B5-defined tumor-initiating progenitor cells. CCL2 and its related receptor (CCR2) promote the migration of brain tumors and monocytes across the vascular endothelium (Vakilian et al., 2017; Cho et al., 2019). Besides, Aretz et al. (Aretz et al., 2022) demonstrated that the cross-talk between CCL2 and β-catenin could affect the dryness and immune escape mechanism of GBM by regulating the activity of immune cells and glioblastoma stem cells. In animal models, mNOX-E36 blocked angiogenesis and macrophage recruitment, and tumor volume and blood volume were reduced. SFN, which is converted from glucosinolates in broccoli/broccoli buds, prevents chemically induced cancers and inhibits tumor growth in rats (Soundararajan and Kim, 2018; Kaiser et al., 2021; Yan et al., 2021). In tumors, SFN may act by regulating multiple survival signaling pathways by inhibiting carcinogen metabolism, inducing oxidative stress, regulating metabolism, inhibiting cell cycle, and inducing apoptosis (Juengel et al., 2017; Lei et al., 2019; Li et al., 2020; Li et al., 2021). Zhou et al. (Zhou et al., 2020) also revealed the subcellular mechanism by which SFN-CYS (SFN analogue) inhibits human GBM invasion by regulating proteome expression. Methylation of UPP1 was confirmed as a prognostic factor for GBM multiforme in two bioinformatics analyses (Weng and Salazar, 2021; Yu et al., 2021). UPP1 is a potential biomarker of thyroid cancer, and the possible mechanism of regulating epithelial-mesenchymal transition (EMT) plays the role of oncogene (Guan et al., 2019). MethSurv database (Modhukur et al., 2018), a web tool that is used to perform multivariable survival analysis by using DNA methylation data, has been used to perform survival analysis for a CpG located in or around the proximity of a UPP1 in GBM. Results showed multiple methylation modification sites associated with GBM prognosis. Detailed results are shown in Supplementary Figure S5. However, the mechanism of action of these six molecules with OXA is still unclear and requires further investigation.

Our study comprehensively investigated the changes in the transcriptomics, proteomics, and metabolomics levels of GBM cells after OXA treatment, explored the consistency between omics, and screened some possible key pathways, including the FoxO signaling pathway, PI3K-AKT signaling pathway, and ferroptosis. OXA may affect the biological functions of GBM cells by regulating these pathways, and the multi-omics integration tool enhances our ability to focus on specific pathways with potential biological significance in GBM after OXA drug treatment, which will be verified in later studies. In addition, among the 138 genes in 11 pathways screened by multi-omics, six genes (ITGA5, MET, F2R, CCL2, SFN, and UPP1) had independent prognostic roles in both TCGA and CGGA cohorts. Previous in vitro and in vivo studies have found that ITGA5, MET, F2R, CCL2, and SFN are involved in the regulation of the occurrence, development, invasion, metastasis, and drug resistance of GBM. These processes may be related to GBM inhibition by OXA. Many of these processes still lack validation of biomolecular mechanisms, and the data generated by a large number of omics studies are still underutilized. In addition, the effect of OXA on glioma was determined by performing in vitro and in vivo experiments. CCK8 and colony formation assays indicated that OXA could inhibit the cell proliferation of glioma cells. Transwell assay results show that OXA impaired the migration/invasion of GBM in vitro. Our constructed subcutaneous xenograft mouse models also indicated that OXA could effectively inhibit glioma cell growth.

We acknowledge some limitations of this study. Despite our focus on biologically plausible mechanisms, validation is still lacking, thus requiring further investigation. Additionally, the broad applicability of our conclusions needs to be confirmed. The rigor and credibility of the study results can be improved by integrating multiple data types and mainly reporting the pathways that were jointly found by combining different data types, which were found to be related to the prognostic genes. In conclusion, multi-omics methods were used to effectively understand the interaction mechanism and combined effects of drugs and disease processes.