The Study About The Effect And Mechanism For Deubiquitinating Enzyme USP22 On The Proliferation, Migration, Invasion And Stemness In Malignant Glioma

Malignant glioma represented by glioblastoma multiforme(GBM) has been regarded as one of the most common intracranial tumors. Despite of advancement of comprehensive therapeutic strategies including surgery, radiation and chemotherapy,the median survival of patients was hardly improved, still only approximately 14 months. GBM are usually featured by their rapid proliferation,migration,invasion,relapse and resistance to radiation and chemotherapy. Presently,the “cancer stem cell” hypothesis has been widely accepted, which indicates that the obstinacy should be authentically attributed to a small group of side populations in the center of tumor called “glioma stem cells(GSCs)”,who demonstrate many features similar to neural stem cells like neurosphere formation,self-renewal and differentiation. Conventional therapeutic strategies including surgery, radiation and chemotherapy that mainly target tumor bulk comprised of differentiated glioma cells,have undesirable ability to eliminate glioma stem cells, so the subsequent relapse is inevitable. In conclusion, it is of great significance to explore specific GSC-targeted therapy to improve the prognosis of GBM patients.In recent years, the Polycomb gene family has become a hot seat in the research sphere of glioma stemness. These genes can form two complexes:polycomb repressor complexes 1(PRC1) represented by Bmi1 and polycomb repressor complexes 2(PRC2)based on Ezh2, which modulate the transcription of multiple downstream genes by epigenetic regulation. It is critical for stem cells that their capacity of self-renewal and differentiation is rigidly controlled and soundly orchestrated. Epigenetic perturbations by misregulation of polycomb genes are probable to be the cause for transcriptional deregulation of tumor suppressor genes that leads to unscheduled activation of stemness-related pathways, promoting the establishment of cancer stem cells and progression of tumors.It worth noting that the experimental team led by Glinsky has found one "death-from-cancer" genetic profile characterized by close relation with tumor stemness and importance in predicting invasion and metastasis of many kinds of tumors, which consists of 11 numbers including polycomb gene Bmi1 and the noval cancer stem cell marker USP22.Belonging to ubiquitin specific protein hydrolase(USP) family that comprises of the largest subtype of deubiquitinating enzymes(Dubs), USP22 plays an important role in regulating cell proliferation, migration, invasion and immune environment. On the mechanism of transcriptional regulation, USP22 can participate in gene expression by epigenetic modification. On one hand,as a member of the SAGA transcription complexes, USP22 can participate in the transcriptional regulation of target genes by deacetylation. On the other hand, USP22 can directly make histone ub H2 A and ub H2 B deubiquitinated, thereby regulating the target gene transcription. On the mechanism of post- translational modification, USP22 can remove the ubiquitin tag from target proteins to prevent their degradation and enhance the stability of target proteins. Some target proteins have been identified such as SIRT1, TBP1, FUBP1 and RCAN1. Previous study has found USP22 is overexpressed in a series of malignant tumors, such as gastric cancer, colon cancer, prostate cancer and it can promote tumor proliferation, invasion, metastasis and inhibit tumor apoptosis. According to related study, USP22 can participate in the regulation of multiple cancer promoting signal transduction in different tumors, such as JAK-STAT signal pathway, androgen receptor(AR) signaling pathway, FAK(focal adherence kinase) pathway and Bmi1 mediated INK4a/ARF and AKT signal pathway. At present, individual reports have maintained that compared with normal brain tissues, USP22 expression is more enriched in tumor tissues,which is positively correlated to tumor grade and nagetively correlated to overall survive time. Meanwile,knockdown of USP22 promotes human brain glioma apoptosis and induces arrest in G1/S phase. However,apart from cellcycle and apoptosis,whether USP22 can affect glioma proliferation,migration and ivasion remains elusive. More importantly,as a cancer stemness related marker, whether USP22 can regulate glioma stemness should worth more intensive study. As a deubiquitinase, oncogenic potential of USP22 should largely depend on its stabilization of a series of oncoprotein. Hence, the function of USP22 on stenmess maintenance should be also closely related with stabilization of stemness-associated substrates.As the center of "death-from-cancer" genetic profile, Bmi1 can regulate the expression of many downstream genes by epigenetic modification. In malignant glioma, studies have demonstrated that Bmi1 can not only promote the malignant biological behavior of differentiated glioma cells, reduce the sensitivity of radiotherapy and chemotherapy, but can maintain cell survival and self-renewal of glioblastoma stem cells, prevent glioma stem cells from apoptosis and the immune system attacks. Knockdown of Bmi1 can significantly reduce the potential of colony formation and tumorigenicity of glioma. So reduction of Bmi1 levels in malignant glioma is of great significance for the eradication of glioma stem cells. Previous studies have shown that Bmi1 is a small molecular protein with a short half-life in vivo, whose post-translational level is regulated by E3 ubiquitin ligase mediated proteolysis. In fact, related studies have mentioned that E3 ubiquitin ligase β-Trcp and Spop can influence Bmi1 expression by enhancing Bmi1 ubiquitination, indicating high expression of Bmi1 is largely dependent on the stabilizing role of the deubiquitinating enzymes, so it is crucial for exploration of the deubiquitinating enzymes for stability of Bmi1 in malignant glioma.It has been found that USP22 and Bmi1 are often overexpressed together in many malignant tumors, especially with a positive expressive correlation in gastric cancer. Besides, USP22 has also been mentioned to regulate the expression of Bmi1 in colon cancer and non-small-cell lung cancer, which further indicates a close relation between them. Through access to the Human protein atlas database, we also found that both USP22 and Bmi1 are overexpressed in malignant glioma. In consideration of unclear mechanism for Bmi1 stabilization in malignant glioma and above-mentioned evidences on tight relation between USP22 and Bmi1,a hypothesis was proposed that USP22 may affected cancer stenmess by stabilization of Bmi1, which account for Bmi1 enrichment in malignant glioma.This article can be divided into five parts. In the first part, we comfirmed that compared with normal brain tissues, USP22 was more enriched in tumor tissues and positively correlated to tumor grade by immunohistochemical staining assay of 30 clinical glioma samples(15 WHO II grade glioma and 15 WHO IV grade glioma) combined with investigation of Human protein atlas data base. Survival analysis indicated an adverse correlation between USP22 expression and overall survival time of glioma patients. In the second part,we conducted related research on the biological functions of USP22 in U251 cells and it revealed that knockdown of USP22 attenuated cell proliferation, migration and invasion. In the third part,we transformed differentiated U251,U87 and GBM-HSF cells into undifferentiated condition in serum-free culture and judged tumor stemness according to the capacity of neurosphere formation. We focused on whether USP22 can affect the stemness of above-mentioned cells and found that knockdown of USP22 can all inhibit the capacity of neurosphere formation in U251,U87 and GBM-HSF cells. Then we tested the transcriptional alteration of stemness-associated genes in adherent USP22 knockdown U251 cells and GBM-HSF neurospheres and found that USP22 depletion led to transcriptional inhibition of several stemness-associated genes in both cell lines. By western blot, we detected USP22 protein level under conditions of differentiation and undifferentiation respectively and found USP22 was markedly enriched in the neurosphere condition, compared with the level in differentiation condition. In the fourth part, we tried to explain the mechanism for oncogenic effect of USP22 in glioma cells, especially for cancer stenmess maintenace. To verify our hyperthesis, the expressive and regulatory interactions between USP22 and Bmi1 were intensively studied. Firstly, through immunohistochemistry of 30 clinical glioma samples and immunofluorescence in HEK293 FT and U251 cells,we found USP22 and Bmi1 co-localized to the nucleus, which beared a high fitting degree in intranuclear distribution. Besides, compared with low-grade glioma, both of them display enhanced expression level in GBM with a positive correlation between their expressions. Then we again confirmed that UPS affected Bmi1 protein expression. Through co-immunoprecipitation and immunoblot assay, we found in HEK293 FT cells exogenous USP22 could directly interact with Bmi1, and thus enhance the stability of Bmi1,which was dependent on the activity of deubiquitinating enzyme. In the U251 cells, knockdown of endogenous USP22 elevated the transcription level of Bmi1, but downregulated the protein level of Bmi1 to different extents, which also provided indirect evidence that USP22 plays a role in stabilization of Bmi1 at the post-translational level. When it came to blockade of endogenous Bmi1,the transcription level of USP22 was enhanced dramatically, indicating that Bmi1 may be involved in the transcriptional repression of USP22. Additionally,considering the important effect of PRC on glioma stenmess regulation, we detected the expression of PRC2 numbers in USP22 knockdown U251 cells and found no obvious alteration compared with negative control. In the fifth part,we carried on a preliminary inquiry of the concrete mechanism for glioma stemness regulation mediated by USP22 and Bmi1. Through the relevant gene expression microarray analysis, we found that USP22 and Bmi1 could jointly influence the transcriptional level of many downstream target genes, almost of which were associated with extracellular matrix, neuroendocrine and cell development. PART I:Study of USP22 protein expression in clinical glioma samples with different grade and the relation with prognosisObjective: Confirmation of USP22 overexpression in glioma and explore the relation with prognosis to highlight the oncogenetic potential of USP22.Methods:Comparison of USP22 protein expression in normal brain tissues, low grade glioma(LGG) and high grade glioma(HGG) in 30 clinical glioma samples(15 WHO II grade glioma and 15 WHO IV grade glioma) by immunohistochemistry. Validation the immunohistochemical result by investigation of Human protein atlas database(http://www.proteinatlas.org/). Explore the relation between USP22 expression and prognosis with survival analysis. Results: Immunohistochemical staining assay showed increased USP22 expression from normal tissues, low grade glioma to high grade glioma, which was consistent with USP22 expression of glioma in Human protein atlas database. Survival analysis demonstrated a negative relation between USP22 expression and overall survival time.Conclusion: USP22 was more enriched in tumor tissues and positively correlated to tumor grade and negatively correlated to prognosis, which indicated a close relation between USP22 and malignant behaviors of glioma. PART II:Effect of USP22 on malignant glioma cell proliferation, migration and invasionObjective: To examine the effect of USP22 on malignant biological behaviors of glioma cells.Methods: We constructed two U251 cell lines transfected with USP22 sh RNA by lentivirus infection,and proved the validity of these two sh RNA with Western blot. Then we examined the potential of U251 cell proliferation by WST-1 method and did a comparison of proliferation ability with control group in 24 h, 48 h and 72 h respectively and drew the proliferation curve. Then we examined the potential of U251 cell migration by wound healing test and compared the ability of migration with control group in 18 h and 36 h respectively. Finally, we examined the potential of U251 cell invasion by transwell assay and compared the ability of invasion with control group after incubation for 72 h.Results: In the first 24 h, there was no obvious statistical difference in the cell proliferation rate compared with control group. But after 48 h and 72 h, cell proliferation of USP22 sh RNA transfected group was inhibited with significant statistical difference compared with the control group(p<0.001). By wound healing assay,we found cell migration of USP22 sh RNA transfected group was also inhibited after 18 h and 36 h with significant statistical difference compared with the control group(p<0.001). By Transwell assay,we also found cell invasion of USP22 sh RNA transfected group was attenuated after 72 h with significant statistical difference compared with the control group(p<0.001).Conclusion: USP22 plays an important role in maintaining cell proliferation, migration and invasion of U251 cells. PART III: Study about effect of USP22 on glioma stemnessObjective: To explore effect of USP22 on stemness of different glioma cell lines according to the capacity of neurosphere formation.Methods: Lentiviruses were packed and U251 and U87 cells were infected to establish USP22 stably knockdown cell lines. We also constructed inducible lentivirus USP22 sh RNA expression system in GBM-HSF cells. The transfection efficiency was detected by Real-time quantitative PCR and Western blot. In serum-free medium, U251, U87 and GBM-HSF cells were cultured into neurospheres. Through observation and statistical analysis, we researched on the influence of USP22 on the ability to form neurospheres in the three cell lines. Detecting the alteration of stemness-associated gene transcription under conditions of differentiation and neurosphere respectively by Real-time quantitative PCR. Examining the alteration of USP22 protein level under conditions of differentiation and neurosphere respectively by western blot.Results: The efficiency of USP22 sh RNA in U251 and U87 cells was satisfactory. In GBM-HSF cells, after doxycycline treatment, USP22 m RNA levels decreased by about 75%, protein expression also significantly reduced. In serum-free medium, U251, U87 and GBM-HSF cells are all able to form neurospheres. And in these three kinds of cells, inhibition of USP22 expression led to reduction of neurosphere diameter with significant statistical difference(p<0.001) compared with control group. We tested the transcriptional alteration of stemness-associated genes in adherent USP22 knockdown U251 cells and GBM-HSF neurospheres. We found that USP22 depletion led to transcriptional inhibition of several stemness-associated genes in both cell lines. By western blot, we detected USP22 protein level under conditions of differentiation and neurosphere respectively and found USP22 was markedly enriched in the stem cell culture condition, compared with the level in the differentiation culture condition.Conclusion: 1.USP22 is one of the markers of glioma stem cells, with higher expression in glioma stem cells compared with differentiated glioma cells. 2. USP22 is essential for sustaining the stemness of malignant glioma cells, the mechanism can be partly explained by the transcriptional regulation of certain stemness-associated genes by USP22. PART IV: Study on the mechanism of correlated expression of USP22 and Bmi1Objective: Exploration of the mechanism for the effect of USP22 on malignant behaviors and intensive study on expressive and regulative correlation between USP22 and Bmi1.Methods: The correlation between the expression of USP22 and Bmi1 was Observed and analyzed by immunohistochemistry of 30 clinical glioma samples collected from departmartment of neurosurgery in Changzheng Hospital. In HEK293 FT cells, exogenous USP22 and Bmi1 were cotranfected and the correlation between USP22 and Bmi1 in the subcellular localization and expression level was detected by immunofluorescence staining. In U251 cells, the correlation between endogenous USP22 and Bmi1 in subcellular localization and expression level was detected by immunofluorescence staining. In HEK293 FT cells, exogenous Bmi1 was transfected and exogenous Bmi1 protein level after proteasome inhibitor treatment was detected by western blot. In HEK293 FT cells, GFP-Bmi1 and exogenous wild type Flag-USP22/enzyme activity mutant Flag-USP22C185 S were cotransfected and the effect of exogenous USP22 on protein level of exogenous Bmi1 was detected by Western blot. In USP22 knockdown U251 cells, Bmi1 m RNA and protein levels were detected by Real-time quantitative PCR and Western blot. In HEK293 FT cells, direct interaction between USP22 and Bmi1 was examined by coimmunoprecipitation. In Bmi1 knockdown U251 cells, USP22 transcription level was detected by Real-time quantitative PCR.Results: Through analyzing the results of immunohistochemistry for USP22 and Bmi1 in 30 clinical glioma samples, we found a weak/medium expression level of both USP22 and Bmi1 in low grade glioma and a significantly higher expression level in GBM samples with a strong positive correlation of USP22 and Bmi1 expression. Through forced overexpression of exogenous USP22 and Bmi1,we observed a positive correlation between exogenous USP22 and Bmi1 in subcellular localization and expression level by immunofluorescence. In U251 cells, a similar correlation was also observed between endogenous USP22 and Bmi1. In HEK293 FT cells, exogenous Bmi1 protein level was affected by ubiquitination level and exogenous overexpression of wild type Flag-USP22 increased the expression of GFP-Bmi1, whereas overexpression of enzyme activity mutant Flag-USP22C185 S hardly affected the Bmi1 protein level. In U251 cells, depletion of endogenous USP22 expression led to transcriptional activation of Bmi1, but the protein level of Bmi1 had a downward trend in different degree, especially in the USP22 sh RNA1 group, compared with control group. In HEK293 FT cells, exogenous overexpression of wild type Flag-USP22 rather than Flag-USP22C185 S enzyme activity mutant directly interacted with exogenous GFP-Bmi1 by co-immunoprecipitation assay. In Bmi1 knockdown U251 cells, the transcription level of USP22 was increased. Finally, the protein level of the several PRC2 family members were detected by Western blot in U251 cells. We found the endogenous USP22 expression inhibition hardly affected the protein level of PRC2 complex subunits like Ezh2, Suz2 and the H3K27me3.Conclusion: 1.USP22 and Bmi1 co-localized to the nucleus and there was a high fitting degree between their subcellular localization. 2. Compared with low-grade glioma, both of them display enhanced expression level in GBM. The expression level of USP22 and Bmi1 is positively correlated not only in clinical samples but also in GBM cell lines. 3. The stability of Bmi1 protein is indeed regulated by ubiquitination level. 4. In the absence of transcriptional interference, the protein level of Bmi1 can be stabilized by USP22, and the stabilization is achieved by the direct interaction with each other that is depend on the deubiquitinating enzyme activity of USP22. 5. On one hand, endogenous USP22 can suppress the transcription levels of Bmi1; on the other hand USP22 can maintain the stability of Bmi1 protein in U251 cells. 6. The positive correlation between the expression of USP22 and Bmi1 can be explained by the mechanism of Bmi1 regulation of USP22 transcription in U251 cells. 7. USP22 hardly affects the protein level of PRC2 complex subunits like Ezh2, Suz2 and the H3K27me3. PART V: Preliminary study of the molecular mechanism for glioma stemness regulation mediated by USP22 and Bmi1Objective: Research on the gene profile that is influenced by USP22 and Bmi1 and explore its relation to glioma stemness.Methods:In U251 cells transfected with USP22 sh RNA1, Bmi1 sh RNA and scramble sh RNA, we identified genes jointly influenced by USP22 and Bmi1 with Agilent gene expression microarray to analyze related signal pathway of these genes. In these genes, we further screened these had clear biological functions in glioma by searching the Pubmed literature database,then randomly selected 8 genes for validation of microarray results by real-time quantitative RCR.Results: Compared with the negative control group, after inhibition of the expression of endogenous USP22 and Bmi1, there were respectively 2151 and 1488 genes occurred more than 2 fold alteration on the level of transcription, and 413 of which are jointly influenced by USP22 and Bmi1. Through related signal pathway enrichment analysis,we found that these 413 genes were mainly associated with extracellular matrix, neuroendocrine, immune regulation and cell differentiation pathway. We further screened the genes that had clear biological functions in glioma and found that these genes are also involved in above-mentioned pathways. We randomly selected 8 genes and found most of these genes were associated with glioma stemness maintenance. Chip analysis revealed that transcriptional level of all these 8 genes were suppressed to some extent in either USP22 or Bmi1 knockdown U251 cells. The result of Real-time quantitative PCR was consistent with that of chip analysis.Conclusion: 1. In glioma cells, USP22 and Bmi1 influence the expression of many genes in either a direct or indirect manner, many of which are jointly influenced by both USP22 and Bmi1, indicating a high correlation between USP22 and Bmi1 in transcriptional regulation. 2. Genes jointly influenced by USP22 and Bmi1 are mainly associated with extracellular matrix, neuroendocrine, immune regulation and cell differentiation pathway. 3. Through validation of the microarray results, USP22 and Bmi1 play a role in maintaining the transcription of many stemness-associated genes.