Prognostic Analysis Of Primary Glioblastoma With IDHwt/MGMTunmet Type Based On 4D-LFQ Proteomics And Phosphoproteomics

BACKGROUND: Brain tumors are currently the seventh leading cause of cancer death in China,with glioblastoma multiforme(GBM)being the most malignant type among them.The molecular diagnosis of GBM with IDH wild-type IDH1/IDH2 and MGMT promoter non-methylation(IDHwt/MGMTunmet)has no effective treatment.Although patients with this type of GBM theoretically have the worst prognosis,there are still patients who have a better prognosis.Therefore,further typing studies of this subtype of GBM are necessary to improve its classification,facilitate clinical disease monitoring,and enable targeted therapy.OBJECTIVE:By utilizing 4D label-free quantification(LFQ)proteomics and 4D-LFQ phosphoproteomics,we were able to screen for prognostic markers in IDHwt/MGMTunmet GBM.This study lays the foundation for further refinement of molecular subtyping in GBM.METHODS: 1.We included a total of 315 patients who were diagnosed with the IDHwt/MGMTunmet type of GBM between January2016 and May 2021.We collected basic patient information,clinical data,and pathological characteristics and divided the patients into two groups based on Overall Survival(OS)– the short survival group(STS)with OS less than 12 months and the long survival group(LTS)with OS more than 36 months.We then compared the differences in clinical and pathological information between the two groups and constructed a multifactorial Cox regression model to determine the clinical prediction model of previous cases.Our aim was to identify clinical and pathological factors that are related to the prognostic impact of this type of GBM.2.We randomly selected fresh tissue specimens from 12 patients in the short survival group(STS)and 13 patients in the long survival group(LTS)with GBM.We analyzed these specimens for their overall histological profile,differentially expressed proteins,differentially phosphorylated peptides,and predicted differential functional kinases.We also screened them for six candidate prognostic markers using4D-LFQ proteomics and 4D-LFQ phosphoproteomics by mass spectrometry.3.Our study involved several methods for validation and analysis,including immunohistochemical validation of paraffin sections from patients with IDHwt/MGMTunmet GBM,construction of prognostic curves using immunohistochemical scores and subject operating characteristic curves(ROC),assessment of marker specificity and sensitivity,and pan-cancer analysis of candidate markers.We also conducted association analysis of tumor stemness and heterogeneity,tumor immunity,kinase prediction,and pathway analysis.Results: 1.The results of a one-way Cox regression analysis revealed that several factors had an impact on patient survival.Specifically,the presence or absence of tumor involvement in the lateral ventricles(Hazard Ratio(HR)= 3.335,95% confidence interval1.803-6.617)and corpus callosum(HR = 3.228,95% confidence interval1.794-5.808)were significant predictors of survival.Additionally,the number of lesions present(HR = 1.474,95% confidence interval1.160-1.874)and the ability of the surgical approach to achieve total or near-total excision(HR = 0.554,95% confidence interval 0.440-0.699)also had an impact on patient survival.A multifactorial Cox proportional risk regression analysis was conducted,and the results indicated that lesion involvement of the lateral ventricles was the most significant predictor of overall survival,with an HR of 2.785(P = 0.004).2.Proteomics analysis of samples from STS and LTS groups identified a total of 7613 proteins.Among them,216 proteins were up-regulated and 284 were down-regulated in the LTS group.These differentially expressed proteins were mainly involved in tumor immunity and stimulation response.Phosphorylation proteomics of the samples from both groups identified a total of 9531 phosphorylation sites,and the differential phosphorylation sites in the LTS group had 464 sites up-regulated and 299 sites down-regulated.These differentially phosphorylated proteins were mainly involved in tumor immune cell response,immune checkpoint,phosphorylation signaling regulation,and synaptic signaling.Four kinases were screened for differential kinases,among which PIKFYVE,SGK,and WSTF were functionally upregulated in the LTS group,while CASK was downregulated.DLGAP1 was involved in synaptic function related to both histological functions and might be involved in neuron-GBM network function and affect invasion.The results of the combined dual-omics analysis,literature-based,and transcriptomics database prognostic analysis identified candidate protective factors CCDC6,DLGAP1,EMILIN3 and risk factors ATM,KRT7,PDGFRA.3.The immunohistochemical validation of candidate validation proteins showed that DLGAP1 was highly expressed in the long survival group(LTS)(P < 0.05),while there was no significant difference in expression between the two groups for the remaining proteins.DLGAP1 was able to differentiate the short survival group(STS)from the other survival groups with a survival-dependent area under the ROC curve(AUC)of up to 0.72.High expression of DLGAP1 was positively correlated with IDH1 mutation frequency in GBM(P < 0.05).Additionally,high expression of DLGAP1 in overall glioma and low-grade glioma was positively correlated with prognosis.Co-expression analysis revealed DLGAP1 involvement in pathway functions related to synaptic signaling and tumor invasion.Conclusions: 1.The involvement of the lateral ventricle is the most significant clinical adverse factor for IDHwt/MGMTunmet GBM.The overall differences in protein expression levels and differentially phosphorylated sites between LTS and STS groups can effectively distinguish between these two groups and predict prognosis;2.DLGAP1 may be a prognostic marker for this subtype and may regulate the development of GBM in this subtype by modulating neuronal synaptic activity.