The Role Of MAGED4B Gene In The Proliferation,Migration And Apoptosis Of Glioma Cells And Its Imaging Phenotype

Part 1 MAGED4 B promotes glioma progression via inactivation of the TNF-α-induced apoptotic pathway by down-regulating TRIM27 expression[Background] Glioma,originated from glial cells,is the most prevalent primary malignant tumor in the brain.The main treatments for glioma include surgical resection,radiotherapy and chemotherapy.However,the therapeutic effect of current treatment is not ideal,with poor prognosis and high mortality.Studies have shown that effective tumor biomarkers are valuable for early detection,early diagnosis and early treatment of tumor patients,and provide information to predict the prognosis of patients.The unbalanced ratio of tumor cells proliferation and death is a key factor in tumorigenesis,in which the inhibition of apoptosis plays an important role.Most antitumor drugs in the clinic work by inducing apoptosis.Therefore,it has important clinical value and significance to develop tumor markers and to explore their apoptosis-related molecular mechanism in glioma.MAGED4 B belongs to the family of the melanoma-associated antigen（MAGE）.Many MAGE proteins are highly expressed in tumors,and drive tumor progression through various mechanisms.Recently,some studies have shown that MAGED4 B is highly expressed in various types of cancer,including lung cancer,oral squamous cell carcinoma,breast cancer,liver cancer,and etc.However,there are few studies about the expression of MAGED4 B in gliomas and the potential role of MAGED4 B in glioma apoptosis remains to be elucidated.[Objective]1.To know the expression of MAGED4 B in glioma tissues and glioma cell lines,,and to explore the relationship between MAGED4 B and patient prognosis by analyzing the correlation between MAGED4 B and the clinicopathological parameters.2.To explore the effects of MAGED4 B expression levels on the proliferation,invasion,migration and apoptosis of glioma cells and to study the underlying apoptosis-related molecular mechanisms.[Methods] 1.The cancer tissues and the adjacent tissues were collected from glioma patients.The expression levels of MAGED4 B were measured using real-time quantitative polymerase chain reaction（q PCR）,Western blotting（WB）and immunohistochemistry（IHC）.The relationship between the expression levels of MAGED4 B and the clinicopathological parameters was analyzed.And the relationship between MAGE4 B and the prognosis of glioma patients was studied based on Kaplan-Meier survival curve method.2.The U87 cell line with stable knockdown of MAGED4 B was selected after sh-MAGED4 B lentivirus infection.Glioma cells overexpressing MAGED4 B was constructed using transient transfection.The effects of MAGED4 B expression on the proliferation of glioma cells was evaluated by CCK-8 test,colony formation test,EDU test and flow cytometry.Cell scratch assay and Transwell migration assay were used to explore the effect of MAGED4 B on glioma cell migration and invasion.Flow cytometry and TUNEL experiment were used to analyze the effect of MAGED4 B on glioma cell apoptosis.The effects of MAGED4 B on the growth of glioma cells in vivo was studied in intracerebral-grafted glioma model in nude rat.The effects of MAGED4 B on the genesis and progression of glioma cells was investigated by Co-Immunoprecipitation（Co-IP）assay,immunofluorescence and WB assay.[Results] 1.In 8 pairs of fresh gliomas and their corresponding paracancerous tissues,the levels of m RNA and protein expression of MAGED4 B in para-cancerous brain tissues were lower than those in the corresponding cancer tissues（P < 0.05）.There was a significant correlation between the expression levels of MAGED4 B and the age,tumor grade,tumor diameter and Ki-67 expression of glioma patients based on the analysis results of the clinical data of 170 glioma patients（P < 0.05）.COX regression analysis showed that high tumor grade（P = 0.001）and high MAGED4 B expression（P = 0.006）were independent risk prognostic factors of glioma patients.Kaplan-Meier survival curve shown that glioma patients with high MAGED4 B expression were associated with poor prognosis,and the difference was statistically significant（P = 0.002）.The above results suggested that MAGED4 B may promote the occurrence and development of gliomas.2.The U87 cell line with stable knockdown of MAGED4 B was selected after sh-MAGED4 B lentivirus infection.Glioma cells overexpressing MAGED4 B was constructed using transient transfection.The results showed that overexpression of MAGED4 B promoted the proliferation,migration and invasion of glioma cells,and inhibited the apoptosis of glioma cells.In addition,glioma cells overexpressing MAGED4 B acquired chemoresistance against cisplatin.3.Nude rats with intracerebrally-grafted glioma model were used to observe the effect of MAGED4 B on tumor growth in vivo.The results showed that the grafted tumors in the brain grow slower in nude rats injected with MAGED4 B knocked-down glioma cells than that of control group.4.The interaction between TRIM27 and MAGED4 B was studied in U87 cells.There was an endogenous co-localization of MAGED4 B and TRIM27 in U87 cells.In glioma cells,the expression of TRIM27 was increased after MAGED4 B knockdown.In addition,proteins of TNF-α-induced apoptosis pathway（USP7,RIP1,Caspase8,Caspase3,Cleaved-caspase3）were up-regulated.[Conclusions] 1.The expression of MAGED4 B in gliomas was higher than that in normal tissues.And the expression of MAGED4 B was positively correlated with the glioma grade, tumor diameter,Ki-67 levels and the age of patients.In addition,the prognosis of glioma patients with high expression levels of MAGED4 B is less ideal compared with those with low expression levels of MAGED4 B.MAGED4B was a potentially independent prognostic factor for patients with glioma.2.Overexpression of MAGED4 B promoted the proliferation,migration and invasion of glioma cells and inhibited the apoptosis of glioma cells.It also increased the chemoresistance of glioma cells against cisplatin.3.MAGED4 B induced the inactivation of TNF-α apoptotic pathway by inhibiting the expression of TRIM27,which promoted the occurrence and progression of glioma.Part 2 Correlation between multimodal MRI and the expression of MAGED4 B in gliomas[Background]In the first part of the study,we found that MAGED4 B can be used as a tumor molecular marker of glioma.Therefore,it is of great clinical value to explore a non-invasive method to evaluate the expression of MAGED4 B.Magnetic resonance imaging（MRI）is a noninvasive approach to effectively evaluate patients with glioma.Compared with conventional MRI technology,advanced MRI technologies can better capture tumor changes at the microscopic level.Advanced MRI imaging techniques include diffusion weighted imaging（DWI）,magnetic resonance spectroscopy（MRS）and diffusion tensor imaging（DTI）.At present,there are no studies on the correlation between MAGED4 B and MRI imaging parameters.In order to identify imaging markers that can non-invasively evaluate the expression of MAGED4 B in gliomas,multimodal MRI technique was used to explore the relationship between the expression of MAGED4 B and imaging parameters.[Objective] 1.To analyze the correlation between the expression of MAGED4 B and the ADC parameters of DWI in gliomas with different pathological grades.To explore whether DWI can be used as a non-invasive predictive tool to evaluate the expression of MAGED4 B in gliomas.2.To analyze the correlation between the expression of MAGED4 B and the Cho/Cr,Cho/NAA and NAA/Cr of ¹H-MRS in gliomas with different pathological grades.To explore whether ¹H-MRS can be used as a non-invasive predictive tool to evaluate the expression of MAGED4 B in gliomas.[Methods] 1.Brain tumor samples from a total of 170 patients diagnosed with brain gliomas from January 2016 to June 2020 in our hospital were collected retrospectively.There were 65 females and 105 males.The median age of the patients was 53.00 years old,and the quartile interval was 42.75-64.00 years old.2.All patients received routine MRI and enhanced scan.On the basis of conventional MRI scan,DWI and ¹H-MRS sequences were checked.Among the 170 patients,166 patients underwent DWI sequence scanning and 67 patients underwent ¹H-MRS sequence scanning.The high-level MRI images were transferred to GEADW4.6 workstations and Functool software was used for post-line processing.The ADC values of tumor parenchyma and peritumoral parts in DWI images were measured respectively,and the ADC values of normal white matter areas on the opposite side was used for comparison,then the r ADC values of corresponding regions were obtained.For ¹H-MRS images,the tumor parenchyma and peritumoral parts were measured to obtain the ratio of Cho/Cr,Cho/NAA and NAA/Cr.The selected range of the peri-tumor part was within the 15 mm around the tumor parenchyma.3.The expression levels of MAGED4 B in 170 cases of gliomas were evaluated by immunohistochemical method.Nonparametric test was used to analyze the relationship between imaging parameters and pathological grade or MAGED4 B expression levels of gliomas.Spearman rank correlation was used to analyze the correlation between imaging parameters and glioma grade or MAGED4 B expression.[Results] 1.The r ADC values of tumor parenchyma and peritumoral parts were significantly different between low-grade and high-grade gliomas（P < 0.05）.The r ADC values of tumor parenchyma and peritumoral parts of high-grade gliomas were lower than those of low-grade gliomas.Spearman correlation analysis showed that the r ADC value of the tumor parenchyma and tumor peritumoral parts were negatively correlated with the glioma grade（P < 0.05）.2.The rADC value of MAGED4 B high expression group,regardless of tumor parenchyma or peri-tumor part,was significantly lower than that of MAGED4 B low expression group（P < 0.05）.Spearman correlation analysis indicated that the r ADC value of tumor parenchyma and peritumoral part was negatively correlated with the expression of MAGED4B（P < 0.05）.3.The Cho/Cr and Cho/NAA values of the high-grade group,regardless of tumor parenchyma or the peritumoral part,were higher than those of the low-grade group,while the NAA/Cr value of the high-grade glioma group was lower than that of the low-grade group（P < 0.05）.The Spearman correlation analysis showed that the values of Cho/Cr and Cho/NAA in the parenchyma and peritumoral parts of the tumor were positively correlated with the pathological grade of glioma,while the value of NAA/Cr was negatively correlated with the pathological grade of glioma（P < 0.05）.4.Compared with the MAGED4 B low-expression group,the Cho/Cr and Cho/NAA values of the MAGED4 B high-expression group were higher in the tumor parenchyma and the peritumoral part of the glioma（P < 0.05）.Spearman correlation analysis suggested that the values of Cho/Cr and Cho/NAA were obviously positively correlated with the expression of MAGED4 B in the tumor parenchyma or in the peritumoral parts（P < 0.05）.[Conclusions] 1.In the parenchyma and peritumoral tissues of gliomas,a higher tumor grade was tested with a lower value of r ADC,a higher value of the Cho/Cr and Cho/NAA,and a lower value of NAA/Cr.DWI and ¹H-MRS can be used as noninvasive methods to predict the pathological grade of gliomas.2.In the parenchyma and peritumoral tissues of gliomas,higher expression of MAGED4 B was associated with lower values of r ADC and higher values of Cho/Cr and Cho/NAA.DWI and ¹H-MRS might be used as noninvasive tools to estimate the expression levels of MAGED4 B.Part 3 The value of radiomics combined with clinical features in preoperative prediction of MAGED4 B expression in gliomas[Background] Radiomics is based on the studies of big data of medical images,and reflect the biological features of lesions from diffrent dimension and levels.Compared with traditional imaging studies,it can help analyze the inherent heterogeneity of tumors more comprehensively.Moreover,accurate predictive models can be established through the algorithms of machine learning.At present,radiomics has been widely used in clinical research,and has played an critical role in tumor diagnosis,differential diagnosis and prognosis evaluation of patients.Glioma is the most common tumor in human brain,which has high degree of malignancy,strong invasiveness,poor therapeutic effect,high postoperative recurrence rate and poor prognosis.The application of effective tumor molecular markers not only contributes to the early screening of tumors,but also helps to guide treatment options,to evaluate the therapeutic effects and to predict the prognosis of patients.However current methods of clinical diagnosis of molecular markers are invasive and expensive.Study results form part 1 suggested that MAGED4 B can be used as a tumor molecular marker of glioma.Overexpression of MAGED4 B in glioma promoted the occurrence and development of glioma.At present,there is no study on the prediction value of MAGED4 B expression in gliomas by MRI radiomics.Therefore it is valuable to explore a radiomics model that can non-invasively predict the expression of MAGED4 B in gliomas.[Objective] To construct a predictive model based on clinical and radiomic characteristics to estimate MAGED4 B expression in gliomas preoperatively.[Methods] To retrospectively collect 166 patients with pathological diagnosis of glioma from January 2016 to June 2020,and record the clinical and MRI image data of the patients.The incorporated MRI images were registered by SPM12 software.Three-dimensional manual segmentation of glioma images was performed using 3D slice software,and the pyradiomics package of Python software was applied to preprocess and extract features from T2-weighted imaging（T2WI）,contrast-enhanced T1-weighted imaging（CE-T1WI）and Apparent Diffusion Coefficient（ADC）of diffusion weighted imaging（DWI）.Based on ANOVA analysis and LASSO regression analysis,the extracted features were screened for dimensionality reduction,and factors with the most predictive value were retained.Logistic regression analysis was applied to construct a radiomics model based on the three individual MRI sequences and their combined sequences.The receiver operating characteristic（ROC）curve was used to determine the best radiomics model for predicting MAGED4 B expression.The univariate analysis method was used to screen out the clinical parameters that were significantly related to MAGED4 B levels.Based on clinical parameters,a clinical prediction model and a model integrating clinical and radiomics were constructed.ROC curve and decision curve analysis（DCA）were applied to evaluate the predictive efficacy of the model.The nomograms of the best model was draw,and the calibration curves were used to assess the accuracy of the model.[Results] 1.Through ANOVA analysis and LASSO regression analysis,we finally filtered 7 features from the ADC figures,and 7 features from the T2 WI sequences and 11 features from the CE-T1 WI sequences.Followed by union of the three sequences,we then filtered 10 features from the multiparametric MRI sequences.According to the ROC curve analysis,the areas under the ROC curves（AUC）of the CE-T1 WI,T2WI,ADC and the multi-parameter MRI sequence model were 0.928,0.889,0.895,and 0.946 in the training dataset,respectively,0.897,0.830,0.879,and 0.908 in the validation dataset,respectively,and 0.918,0.865,0.889,0.931,in the total patient dataset,respectively.In comparison,the predictive ability of the multiparametric MRI radiomics model was significantly better than that of the individual sequence model.2.Based on the clinical parameters,the clinical prediction model and the integrated model of clinical and multi-parameter MRI were constructed by logical regression analysis.According to the results of ROC curve analysis,the AUC values of the clinical model for the training group,validation group and all patient groups were 0.684,0.766,and 0.702,respectively.The AUC values of the integrated model for the training group,validation group and all patient groups were 0.963,0.976 and 0.954,respectively.Together with the previous results,we found that the predictive efficiency of the integrated model was better than that of clinical model and radiomics model.According to the results of DCA curve analysis,the integrated model had the highest clinical application value.The nomogram and the calibration curve indicated that the prediction probability of the integrated model for the high expression of MAGED4 B in gliomas was basically consistent with the actual probability.[Conclusions] 1.The intrinsic biological features of gliomas were extracted from multi-parameter MRI sequences by the method of radiomics analysis,and the corresponding radiomics prediction model was established.Compared with the traditional morphological model,this prediction model can be used to evaluate the expression status of MAGED4 B more objectively,comprehensively and accurately in gliomas.2.The nomogram of the clinical and radiomics integrated model was constructed through the method of radiomics in combination with clinical parameters.This model displayed high predictive efficiency and was of potential clinical value,and it might provide a noninvasive method to predict the expression level of MAGED4 B in gliomas.Part 4 MAGED4 B expression combined with radiomic parameters to predict the overall survival of patients with glioma[Background] Glioma is the most ordinary primary malignant tumor in the brain.Great progress has been obtained in clinical diagnosis and treatment of gliomas.However,the prognosis of patients is often not ideal due to the high degree of malignancy,strong invasion ability,and high recurrence rate of glioma.More specific,the 5-year survival rate of patients with high-grade gliomas is less than 10%.It is known that magnetic resonance imaging（MRI）technology can reflect the internal information of the diseased tissue non-invasively.Compared with biopsy and checking the diseased tissue,MRI technology can comprehensively reflect the disease information of the tissue.In addition,MRI-based radiomics methods can extract more in-depth information based on a large number of quantitative features about the inherent heterogeneity of tumors.At present,many studies have confirmed that the radiomics model constructed by MRI images can effectively predict the prognosis of tumor patients.Compared with the single MRI sequence and the single radiomics model,the integrated model combined with multi-parameter MRI radiomic characteristics and clinical prognostic factors is more effective in predicting the prognosis of patients.In this study,we found that MAGED4 B promoted the occurrence and development of gliomas,and glioma cells with high expression of MAGED4 B had stronger ability of growth,invasion and anti-apoptosis.Moreover,MAGED4 B can be used as an independent prognostic risk factor for glioma patients.And it is of great significance to incorporate the MAGED4 B expression level into the study of clinical characteristics.Therefore,construction of a predictive model that integrates MRI radiomics parameters and clinical features may provide a better method for predicting the prognosis of glioma patients.[Objective] To construct a predictive model that can evaluate the overall survival rate of glioma patients based on the radiomic parameters and the patient’s clinical characteristics including the expression level of MAGED4 B.[Methods] The clinical and MRI imaging data of 154 patients with glioma who had been pathologically diagnosed（from January 2016 to June 2020）was collected.Patients were preoperatively examined with T2 weighted imaging（T2WI）,contrast-enhanced T1 weighted imaging（CE-T1WI）,and diffusion-weighted imaging（DWI）.The DWI images were post-processed to obtain the corresponding apparent diffusion coefficient（ADC）maps.The three MRI images were registered using SPM12 software.3D images of glioma lesions were manually segmented using 3D slice software.The Python software was applied to preprocess and extract radiomic parameters from the three MRI images.ANOVA analysis and LASSO regression analysis were applied to dimensionally screen the extracted parameters,and the radiomic factors with the most prognostic value were finally selected by stepwise COX regression for model construction.Kaplan-Meier survival curves were used to confirm the predictive ability of the model.Independent prognostic factors were screened in clinical features by univariate and multivariate analysis.Finally,radiomic features and clinical features were integrated to construct the prognostic prediction model,and the prediction model nomogram was drawn.The ROC curve was used to evaluate different models and the calibration curve was used to confirm the accuracy of the integrated model.[Results] Through ANOVA analysis,LASSO regression analysis and stepwise COX regression analysis,11 radiomic features were finally selected from the multi-parameter MRI sequence,of which 4 were from the CE-T1 WI sequence,5 were from the ADC image,and 2 were from the T2 WI sequence.According to the results of Kaplan-Meier survival curves,we found that radiomic characteristics were significantly correlated with the overall survival（OS）of patients（P < 0.05）for the training set,validation set or total patient set.And the OS of patients in the high-risk group was lower than that of the low-risk group.Univariate and multivariate analysis of clinical characteristics showed that the tumor enhancement（P = 0.043）,pathological grade（P < 0.001）and MAGED4 B expression level（P = 0.047）were independent prognostic factors for the OS of glioma patients.The radiomics model,clinical model,and the integrated model of radiomics and clinical were constructed based on the radiomic features and clinical features,respectively.The ROC curves for patient’s 1-year or 3-year OS indicated that the predictive efficacy of the integrated model was higher than the single-approach model.The nomogram of the prediction model was plotted.The calibration curve suggested that the predicted probability of the integrated prediction model for the OS of patients with glioma was basically the same as the actual probability[Conclusions] In this study,the radiomic characteristics of multi-parameter MRI screened by the radiomics method were significantly correlated with the OS of patients with glioma（P < 0.05）.The clinical and radiomics integrated model which constructed based on both clinical features and radiomic features exhibited the best predictive value of OS compared with the single-approach model.The nomogram of the integrated model suggested that integrated model was a non-invasive and inexpensive means of assessing the OS of patients with glioma.