Ovarian cancer is a frequent malignancy in women all around the world, with a high rate of occurrence and mortality. Most individuals can only be diagnosed at an advanced stage due to a lack of adequate screening tools. Unfortunately, at this point, the tumor has frequently migrated to the peritoneal cavity and other organs, resulting in a dismal prognosis(39). After tumor reduction surgery, the typical treatment for advanced OV is combined chemotherapy with platinum(40). Platinum resistance in OV is a process involving multiple molecular mechanisms and pathways, and multidrug resistance, DNA damage repair, and cell cycle regulation are examples(41). However, so far, drug resistance in OV has not been fully explained and cannot be explained by a single process. As a result, it is crucial to identify novel biomarkers for prognosis as well as therapy targets specific to invasive OV.
The demand for energy increases during tumor cell multiplication, resulting in a reprogramming process in tumor cell metabolism. As a result, even under aerobic conditions, tumor cells primarily make ATP via glycolysis, a process known as the Warburg effect(42). Oxidative stress is a phenotype in oxygen ions, in which reactive oxygen species (ROS) decompose the molecular structure through reaction, or reaction with proteins, which can bring oxidative damage. ROS has been shown to influence angiogenesis, proliferation, invasion, metastasis, transformation, survival, and all aspects of tumor development(43). Glycolysis and oxidative stress have been discovered to influence tumor growth and prognosis. Jianlei Bi et al.suggest that glycolysis has a relationship with OV prognosis (44). Zheng et al. confirm that the oxidative stress state is an independent predictor of OV and can predict platinum resistance (45). There is little research combining glycolysis and oxidative stress to evaluate the prognosis and treatment of OV. The current study explored the potential role of a classifier based on glycolysis and oxidative stress expression profiles in predicting the prognosis and therapeutic targets of OV. It suggested that OV patients with a high or low glycolysis-oxidative stress score (GOSs) had different metabolic pathways. The GSEA showed that the genes between Group High/Low in the TCGA-OV dataset were significantly enriched in the assembly of collagen fibrils and other multimeric structures, met promotes cell motility, inflammatory response, and oxidative phosphorylation. The GSVA enrichment shows 20 lanes in TCGA-OV were significantly higher data the risk scores in the GOSs High, while the other 14 pathways were significantly higher in the GOSs Low.The different biological behaviors of GOSs High/Low may bring different prognostic of OV.
Next, we explored the genetic variations and expression of gene mutation status of GOSRGs in TCGA-OV dataset samples. Mutations in TP53 were the largest. MYC had the most CNV expansion. TP53 is a driver gene of OV revealed in studies, TP53 is one of the most somatic mutations regulating cell cycle and apoptosis (46), above all, TP53 is a key regulator of glycolysis (47). MYC stimulates cell growth and proliferation by causing selective gene expression amplification. Through its targets, MYC controls food uptake to produce ATP and essential cellular building blocks that increase cell mass, trigger DNA replication, and promote cell division. In tumors, defects in genetics and epigenetics checkpoints are silenced, and the metabolic activities of MYC9 that promote cell growth and proliferation are released(48). Janicek et al.discovered that the treatment of anti-gene TP53 and MYC could inhibit ovarian cancer line growth(49). Our findings are comparable with the findings of previous studies, implying that TP53 mutations and MYC gene expression amplification can accelerate the cell growth and proliferation of OV.
Then we focused on immune infiltration and checkpoint genes. The results showed that the infiltration abundance of 19 kinds of immune cells had a statistically significant difference between the GOSs High/Low group. The abundance of immune cell infiltration in GOSs(High )is higher than that in GOSs(Low), and this is consistent with Gajewsk and Zhang et al research conclusion(50–52). Increased infiltration of myeloid-derived MDSC, lymphocytes, macrophages, mast cells, neutrophils, and dendritic cells contributes to tumor growth, invasion, and metastasis, and all are negatively correlated with OS(53). Interestingly, there is a certain amount of significant correlation between the content of immune cells and the expression of key genes. The gene GLO1 in GOSs(High) had a negative correlation with the infiltrating abundance of cells, meanwhile, the gene H6PD in the GOSs Low group was significantly positively correlated with most immune cells. There were 15 higher checkpoint gene expressions in GOSs(High) than in GOSs(Low). Therefore, we classified the high components of the GOSs as immunoinflammatory phenotypes, characterized by increased infiltration of immune cells and increased expression of immune checkpoints. These subtypes were substantially associated with immunological activity, indicating the importance of immune control in tumor microenvironment(TME)(54). The genes associated with glycolysis and oxidative stress that can alter the TME have been linked to the formation and progression of OV(55).
In this study, to determine the prognostic value of GOSRGs, we built a prognostic model by LASSO regression. The analysis displayed that the risk score had a certain ability to forecast the OS of OV patients. The survival rate of OV patients in GOSs(High ) decreased compared to GOSs( Low) over time. This is our main discovery. In the risk score, there are five GOSRGs ( AKT1, ERBB2, GLO1, H6PD, and RB1). 4 GOSRGs(ERBB2, GLO1, H6PD, and RB1) differentially expressed in GOSs(High)and GOSs(Low). AKT1 and GLO1 were connected with better OS, whereas the other three genes, ERBB2, H6PD, and RB1 were correlated with poorer prognosis. And, 3 of them are selected as key genes (GLO1, H6PD, and RB1). The results showed that the key genes were associated with many malignancy related pathways in OV, especially can alter the TME (tumor microenvironment) pathways. The key genes were considerably enriched in myeloid leukocyte differentiation, cellular ketone metabolic process, myeloid cell differentiation, regulation of glucocorticoid metabolic process, the establishment of sister chromatid cohesion, cortisol biosynthetic process, tertiary alcohol biosynthetic process. Myeloid leukocyte differentiation and myeloid cell differentiation are leukemia cell triggering factors(56). Ketone metabolism process can suppress PKB/Akt signaling pathways, hence inhibiting tumor cell growth. Furthermore, ketone bodies can influence the expression of apoptosis-related proteins and promote tumor cell death. Ketone bodies can limit tumor cell invasion and migration via modulating transcription factor expression and the activity of extracellular matrix degradation enzymes(57). Glucocorticoids can promote tumor metastasis through ROR1(58). The establishment of sister chromatid cohesion is critical to genome stability, which is directly related to cells undergo cancer transformation.
The RB1 gene was identified to be the first tumor suppressor gene whose mutational inactivation causes human cancer. The inactivation of RB1 allows the production of downstream genes required for the cell cycle to continue into the S-phase and beyond(59). The genes that control RB1 inactivation are well-known oncogenes, such as CDK4 and Cyclin D1, so the genes that inhibit CDK4/6 activity are tumor suppressors(60). In previous studies on ovarian cancer research, patients with ovarian cancer who have Rb protein depletion may be more sensitive to platinum-based treatment and have a longer survival span(61). This is consistent with our research results that RB1 was correlated with a poorer prognosis of OV. Although RB1 was identified as a tumor suppressor gene, the mutation of RB1 may be the cause of OV. H6PD, which converts glucose 6-phosphate to 6-phosphogluconate, is implicated in the downregulation of NADPH in the endoplasmic reticulum lumen(62). H6PD increased in a variety of cancers, and early research suggested that H6PD promoted cancer cell proliferation, and H6PD knockdown inhibited breast cancer cell growth and migration. H6PD downregulation using short interfering RNAs in mouse bladder cancer cells was associated with lower glucose absorption and proliferation(63, 64). H6PD silencing aided in the reduction of gallbladder cancer invasion and migration(65). Furthermore, H6PD elevation is required for purine reprogramming metabolism and the formation of glioma stem cells(66). We found that H6PD was correlated with a poorer prognosis with OV, and this is consistent with previous research. Furthermore, we discovered a substantial positive relationship between H6PD and the majority of immune cells in the low GOSs group. It is unknown what the link is between glycolysis-oxidative stress state, H6PD, and immunological infiltration.
GLO1 belongs to the glyoxalase system. Tumor tissues with high metabolic rates showed increased expression of GLO1, whereas suppressing GLO1 expression caused cell death or injury(67). AP-2a, E2F4, NF-kB, Nrf2 stimulate GLO1 expression through the GlO1 promoter. Furthermore, multidrug resistance in cancer chemotherapy has been tightly connected to GLO1 overexpression(68, 69). These results reveal that elevated GLO1 can accelerate tumor apoptosis while lowering apoptosis in cancer cells via inactivating caspases treated with anticancer medicines. It could be a reversible consequence. However, in the study of Nokin M-J and Zender L (70, 71), GLO1 slicing boosted tumor development and metastasis in vivo, while GLO1 deficiency resulted in a considerable increase in tumor burden. It was also shown that GLO1- induced metastasis in cancer cells was linked to MEK/ERK/SMAD1 cascade activation. But in our prognostic model GLO1 was positively correlated with prognosis of OV.Will these seemingly contradictory data appear in OV? Immunohistochemistry assays were utilized in this investigation to confirm the link between GLO1 and ovarian cancer prognosis. In the validation experiment, ovarian cancer patients with low expression of GLO1 had a stronger prognosis than those with high expression of GLO1, and this was consistent with the research conclusions of Holsoda and et al(72).
Although this study may be useful for clinical guidance, some limitations should be acknowledged. The major defect in the study is that it entails intensive calculations. Regarding the sample size, data from TCGA and GEO is insufficient, and more data must be gathered. Furthermore, additional experimental and clinical research is required to verify our findings.
Despite its limitations, the study certainly adds to our understanding of glycolysis-oxidative stress in OV. Our findings reveal that glycolysis-oxidative stress plays an important role in determining the activities of numerous genes, biological behaviors, and immunological TME landscapes. We believe that GOSs can be utilized as predictive indicators of OV patients. In addition, our study might provide GLO1 as a new target for drug-resistant OV.