PSMC2/E2F1 Axis Promotes the Development and Progression of Glioma

Xuchen Qi Zhejiang University School of Medicine Sir Run Run Shaw Hospital IVF Centre Qin Lu Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University Dajiang Xie Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University Junhui Lv Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University Yiwei Liu Sir Run Run Hospital, School of Medicine, Zhejiang University Shuxu Yang Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University Yirong Wang Sir Sun Sun Hospital, School of Medicine, Zhejiang University Xueyong Zheng (  qxctougao@163.com ) Sir Run Shaw Hospital, School of Medicine, Zhejiang University


Results
PSMC2 is overexpressed in glioma tissues than normal tissues. Moreover, knockdown of PSMC2 can inhibit the proliferation, migration and arrest cell cycle in G2 phase of glioma cells. Additionally, PSMC2 knockdown promotes glioma cell apoptosis by increasing expression of caspase3, caspase8, IGFBP-1, while reducing expression of IGF-I, Survivin, TRAILR-4. In vivo ndings reveal that PSMC2 knockdown inhibit the tumorigenicity of glioma cells. Furthermore, downstream of PSMC2 is explored, identifying E2F transcription factor 1 (E2F1) as a potential target. Notably, E2F1 knockdown exhibits similar effects on the development of glioma with PSMC2, which could strengthen the inhibition effects of PSMC2 knockdown on glioma synergistically.

Conclusions
PSMC2 is closely associated with glioma development by targeting E2F1, and might be considered as a novel therapeutic target in patients with glioma.
Background Glioma is the most common primary craniocerebral malignancy resulting from the canceration of glial cells in the brain and spinal cord, which has the characteristics of high incidence, high recurrence, high mortality and low cure rate [1]. According to the World Health Organization (WHO) classi cation of central nervous system tumors, gliomas can be divided into four grades (I-IV), of which grade IV is also known as glioblastoma (GBM) [2]. In addition, molecular classi cation of gliomas based on gene expression has been proposed, including anterior nerve, neurological, classical and mesenchymal subtypes [3]. Although these different subtypes have been identi ed, effective targeted therapies for gliomas have not been developed in recent decades to improve prognosis, and most low-grade gliomas (I and II) inevitably relapse and develop into high-grade gliomas (III and IV) [4]. Currently, only 5.5% of patients can survive 5 years after diagnosis. Even with multimodal therapy including surgery, radiation, and chemotherapy, the overall median survival is still only 14.5-16.6 months, and the e cacy of treatment is still frustrating [5,6]. Lacking of effective therapeutic targets pose major challenges to prolong the survival and improve the quality of life of patients with glioma. Therefore, there is an urgent need to thoroughly understand molecular mechanism of glioma.
A large amount of evidence indicated that the 26S proteasome is involved in a lot of biological processes, such as cell cycle progression, apoptosis, metabolic regulation, and signal transduction [7][8][9]. Given its importance, the 26S proteasome is a multiple target for anti-cancer therapy [10]. Proteasome 26S subunit ATPase 2 (PSMC2) located in the genome 7q22.1-q22.3, is a key member of the 26S proteasome 19S regulatory subunit [11]. Partial genome deletion of PSMC2 was found in more than 3000 tumors, and PSMC2 is closely related to cancer cells, indicating that PSMC2 may be a potential target for cancer treatment [11]. Recently, the expression correlation and molecular mechanism of PSMC2 in various cancers have been explored. For example, Qin et al., proposed that in pancreatic cancer, PSMC2 expression is up-regulated, and promotes cell proliferation, inhibits apoptosis [12]. He et al. indicated that in colorectal cancer, PSMC2 is related to survival rate, and silencing of PSMC2 can regulate the biological activities of cancer cells [13]. Nevertheless, functional validation and mechanistic studies for PSMC2 in glioma is still lacking.
To the best of our knowledge, this study is the rst attempt to illustrate the potential carcinogenic activity of PSMC2 in glioma. First, clinical specimens were applied to IHC staining to reveal the differential expression of PSMC2 in glioma tissues and normal tissues. Subsequently, the correlation of PSMC2 expression with tumor characteristics in glioma patients was statistically analyzed to clarify that PSMC2 may be associated with the progression and development of glioma. More importantly, potential roles and mechanisms of PSMC2 in glioma cells in vitro and in vivo were explored. Notably, downstream of PSMC2 was investigated by RNA sequencing and IPA.

Results
Relationship between the expression of PSMC2 and glioma characteristics IHC analysis was performed to show that the expression of PSMC2 in the tumor sample was signi cantly higher than that in the normal sample (Fig. 1A). In addition, the mRNA level of PSMC2 was highly abundant in glioma cell lines including U87, U251, U373 and SHG-44 (Fig. 1B). The high expression of PSMC2 in glioma was further con rmed with reference to TCGA-GBM database (Fig. 1C). More importantly, based on the results of tissue microarray determination, the expression of PSMC2 may be associated with age (P < 0.01), recurrence (P < 0.05) and pathological stage (P < 0.01) of glioma (Table 1). In detail, the increased expression of PSMC2 was accompanied by the deepening of tumor malignancy in patients with glioma. In general, it can be known that PSMC2 may be related to the progression and development of glioma. group and shCtrl as negative control group. The green uorescence signal inside U87 and U251 cells was observed in > 80% cell popular, indicating successful transfection ( Figure S1). The successful knockdown of PSMC2 was veri ed by Western Blot (WB) analysis (Fig. 1D). Subsequently, MTT assay ( Fig. 2A), colony formation assay (Fig. 2B), and ow cytometry (Fig. 2C) were used to evaluate the proliferation, colony formation, apoptosis of U87 and U251 cells. Apparently, results indicated that the knockdown of PSMC2 dramatically inhibited cell proliferation (> 2 folds, P < 0.001) and promoted cell apoptosis (> 2 folds, P < 0.001). Moreover, human apoptosis antibody array (Fig. 2D) was used to detect the differential expression of apoptosis-related proteins induced by PSMC2 knockdown in U251 cells. The results showed that the protein expression of caspase3, caspase8, IGFBP-1 was signi cantly up-regulated, while expression of IGF-I, Survivin, TRAILR-4 was signi cantly down-regulated (P < 0.05). Taken together, we concluded that PSMC2 knockdown inhibited cell proliferation and promoted apoptosis of glioma cells.
Knockdown of PSMC2 suppressed tumor growth in vivo Furthermore, mice xenograft models were constructed by subcutaneous injection of U87 cells to further verify the role of PSMC2 in glioma. Results indicated that the tumor growth of glioma was signi cantly inhibited. Speci cally, smaller size and lighter weight of tumors were observed in shPSMC2 group compared with shCtrl group (P < 0.001) ( Fig. 3A-3C). In addition, bioluminescence imaging also proved that the bioluminescence intensity of shPSMC2 group was signi cantly lower than that of the shCtrl group (P < 0.05) (Fig. 3D, 3E), suggesting that proliferation ability of tumor tissue was weaker than that in negative group. Moreover, Ki67 expression in xenografts was also detected by IHC analysis, showing obvious lower expression in shPSMC2 group, which was consistent with the PSMC2-induced inhibition of tumor growth (Fig. 3F). In summary, knockdown of PSMC2 can inhibit the development of glioma cells, which was consistent with data from in vitro experiments.

Exploration of downstream molecular mechanism of PSMC2 in glioma cells
To further investigate the molecular mechanism of the effects of PSMC2 on glioma development, shPSMC2 (n = 3) and shCtrl (n = 3) transfected U251 cells were used for RNA sequencing analysis. In general, total of 2553 differentially expressed genes (DEGs) were identi ed based on |Fold Change| ≥ 2.0 and FDR < 0.05, including 1132 up-regulated DEGs and 1421 down-regulated ones (Fig. 4A, S2A and S2B). Subsequently, typical signaling pathway, as well as disease and function were performed based on Ingenuity Pathway Analysis (IPA) to determine gene enrichment induced by knockdown of PSMC2 ( Figure   S2C and S2D). As indicated, the DEGs had signi cant correlation with cell viability, cell cycle progression, cell invasion and malignant tumor development by analyzing the enrichment of DEGs in IPA disease and function ( Figure S2D). More importantly, considering our purpose to seek for tumor promotor factors, several down-regulated DEGs were selected for veri cation by qPCR and WB (Fig. 4B, 4C). Further combining the PSMC2 related interaction network predicted by IPA analysis, it was shown that E2F1 may be a potential downstream target of PSMC2 (Fig. 4D). Additionally, results of IHC indicated that E2F1 expression levels were obviously higher in glioma tissues than that in adjacent normal tissues (Fig. 4E). Indeed, the expression of E2F1 showed a similar pattern with PSMC2 in glioma tissues and cell lines ( Fig. 4F). Taken together, E2F1 was identi ed as a potential target of PSMC2 in the regulation of glioma.
Knockdown of E2F1 aggravated the inhibition of glioma cells by PSMC2 depletion To verify the hypothesis that PSMC2 regulates glioma development through E2F1, we investigated their functional roles in U251 cells. Herein, cells were divided into the following groups, including shCtrl (transfected with shCtrl, as negative control), shE2F1 (transfected with shE2F1 lentivirus for downregulating E2F1 expression) and shPSMC2 + shE2F1 (transfected with shE2F1 and shPSMC2 lentivirus for simultaneously down-regulating E2F1 and PSMC2 expression). First, shE2F1 (RNAi-3) with the highest knockdown e ciency was screened by qPCR and used in the following experiments ( Figure S3A). Fluorescence inside U251 cells transfected with shCtrl, shE2F1 and shPSMC2 + shE2F1 were observed by microscope and con rmed the successful transfection ( Figure S3B). Moreover, results of qPCR ( Figure  S3C) and WB ( Figure S3D) showed that E2F1 was down-regulated in the shPSMC2 group and PSMC2 also was down-regulated in the shE2F1 group, which indicated that there is a relationship between E2F1 and PSMC2 in U251 cells.

Discussion
In this study, the upregulation of PSMC2 in glioma was demonstrated by IHC and qPCR analysis.
Additionally, the data of in vitro experiments clari ed that knockdown of PSMC2 may inhibit the proliferation, migration and arrest cell cycle. In vivo ndings also revealed that PSMC2 knockdown inhibited the tumorigenicity of glioma cells. More importantly, PSMC2 knockdown promoted apoptosis of glioma cells by increasing expression of caspas3, caspase8, IGFBP-1, while reducing expression of IGF-I, Survivin, TRAILR-4.
Notably, apoptosis is a well-organized cellular mechanism that mediates programmed cell death through apoptosis-related protein cascades [14]. Therefore, the protein levels of apoptosis-associated factors were determined in PSMC2-transfected glioma cells.  [15,16]. Studies have shown that IGFBP-1 inhibited tumor activity by inducing apoptosis in human cancer, such as prostate cancer and breast cancer [17,18]. Furthermore, it was previously suggested that IGFBP-1 might be responsible for the reduced proliferative capacity of malignant glioma cells [19,20].
Zumkeller et al., proposed IGF-I was thought to play a pivotal role in the proliferation promotion of glioma [21]. Fenstermaker et al., supported that Survivin is an inhibitor of apoptosis protein that is highly expressed in human glioma cells [22]. Meusch et al., manifested that abnormalities of TRAIL and its receptor TRAIL-1, TRAILR-2, TRAILR-3, and TRAILR-4 appear to collectively promote the resistance of monocytes to TRAIL-induced apoptosis [23]. These ndings were supported by the fact that PSMC2 silencing induced abnormal protein abundance associated with cleavage apoptosis.
Moreover, downstream of PSMC2 was explored by RNA sequencing followed by IPA, and E2F1 was identi ed as a potential target. Genome-wide mapping studies have revealed that E2F transcription factor 1 (E2F1) with genes from hundreds of promoter regions is involved in numerous cellular pathways [24][25][26][27]. Including cell cycle control [28], apoptosis [29], senescence [30], and DNA damage response [31]. As the primary regulator of cell fate, the complex role of E2F1 has been extensively investigated [32,33]. Interestingly, recent data suggested that a certain correlation between the expression of this factor and cancer. Farra et al., proposed that the role of E2F1 in HCC is not a trivial aspect as it may have a signi cant impact on the development of new treatment options targeting E2F1 [34]. Additionally, Rouaud et al., reported the key role of E2F1 in controlling melanoma cell death and drug sensitivity [35]. Recently, Zhi et al., elucidated the new mechanism of ECT2 affecting the proliferation of glioma cells by regulating the expression of the deubiquitinating enzyme PSMD14 to affect the stability of E2F1 [36]. Strikingly, this study elucidated that knockdown of E2F1 has a similar effect on the development of glioma by detecting cell proliferation, apoptosis, cell cycle and migration and can also synergistically enhance the inhibitory effects of PSMC2 on glioma cells.

Conclusion
In conclusion, PSMC2 was closely associated with proliferation, apoptosis, cell cycle and migration of glioma cells by targeting E2F1, and might be considered as a novel prognostic indicator in patients with glioma.

Immunohistochemical (IHC) staining
Brain glioma tissue and paired normal tissue microarray chips were purchased from Shanghai Outdo Biotech Co. Ltd. (Cat. # HBraG180Su01). Meanwhile pathological characteristics of these clinical samples were collected. All patients signed the informed consent form.

Cells culture
Human glioma cell lines U87 and U251 were purchased from BeNa Technology. U87, U251 cells were cultured in CM1-1 medium containing 90% DMEM-H medium supplemented with 10% FBS. All culture medium was changed every 3 days and cells were humid maintained in a 37°C 5% CO 2 incubator.
Target gene knockdown cell models Short hairpin RNAs (shRNA) of human PSMC2 and E2F1 and related control sequence were designed by Shanghai Bioscienceres, Co., Ltd. for the knockdown experiment. The target sequences were inserted into BR-V-108 vector using the T4 DNA ligase enzyme (NEB). Plasmids were extracted by EndoFree Maxi Plasmid Kit (Tiangen) and quali ed plasmid was packaged with 293T cells. U251 and U87 cells at a density of 2×10 5 cells/ml were seeded in a six-well plate. 24 h later, cells were infected with 100 µL lentiviral vectors (1×10 7 TU/well) additive with ENI.S and polybrene (10 µg/ml, Sigma-Aldrich). After cultured at 37°C with a 5% CO 2 for 72 h, the uorescence was observed by microscope.
qPCR analysis U87 and U251 cells with shPSMC2, shE2F1, shPSMC2+shE2F1, and the corresponding negative controls grown in the 6-well plate were harvested with TRIZOL reagent (Invitrogen, Carlsbad, USA) and then provided by the manufacturer with a standard procedure to isolate the total RNA. Next, RNA samples were washed with ethanol and then dissolved in DEPC H 2 O. Reverse transcription was applied to M-MLV reverse transcriptase and RNase inhibitor (Promega Corporation, Madison, Wisconsin, USA). qPCR was performed using SYBR Master Mix (Takara, Japan), where GAPDH was used as the internal control.
Primer information was shown in the Table S1. Finally, the 2 −ΔΔCq method was used for quanti cation and dissolution curve was drawn.
WBanalysis U87 and U251 cells with shPSMC2, shE2F1, shPSMC2+shE2F1, and the corresponding negative controls grown in the 6-well plate were harvested with Radio Immunoprecipitation Assay (RIPA) buffer (Beyotime, Shanghai, China). Proteins were degenerated in Sodium dodecyl sulfate (SDS) sample buffer, followed by separating on SDS-polyacrylamide gelelectrophoresis (PAGE) gel using electrophoresis. After that, proteins were transferred onto polyvinylidene di uoride (PVDF) membrane, and incubated with primary antibodies (Table S2) overnight at 4°C. Membranes were then incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies IgG (Goat Anti-Rabbit) at room temperature for 1 h. Protein expression was then determined using ECL kit (Thermo Fisher Scienti c, NY, USA).
solution. Finally, the cells passing through the lter were photographed by an inverted uorescence microscope.

Human apoptosis antibody array
Differentially expressed apoptosis-related proteins in U251 cells with shPSMC2, and the corresponding negative controls were detected by human apoptosis antibody array (RayBio, Norcross, GA, USA). First, the membrane was placed in the 8-well trays provided in the kit. Incubations with sample and Biotinconjugated Anti-Cytokines should be performed overnight at 4°C. Overnight blocking and wash steps are useful for reducing background signal intensities even with completed membranes. Wash with Wash Buffer II, followed by repeating incubation with Streptavidin-HRP and chemiluminescent detection. Finally, Gelpro Analyzer software (Media Cybernetics, Rockville, MD, USA) was used to analyze protein expression levels.

Mice xenograft model
The animal experiments performed have been approved by the Animal Protection and Use Committee of Sir Run Run Shaw Hospital. Nude mice (BALB/c males, 4 weeks old) were purchased from Shanghai Lingchang Experimental Animal Co., Ltd (Shanghai, China) and were pathogen-free. Twenty mice were randomly divided into two groups (shCtrl group and shPSMC2 group) before the experiments. Meanwhile, 2 × 10 6 U87 cells with or without knockdown of PSMC2 suspended in PBS were injected into mice under the right axillary skin to construct a mouse xenograft model. Cultured for another 20 days after injection and collected data from 10 days after injection. Mice were injected with cells 10 days after starting to collect data twice a week, including animal weighing, measuring tumor length short diameter. The mice were nally sacri ced by injection of pentobarbital sodium, and the tumor was removed for photographing and weighing.
Repaired and blocked with citrate antigen after tumor tissue was removed from sacri cial mice. Antibody Ki67 (1:200, Abcam, USA, cat # ab16667) was added to shPSMC2 or shCtrl, respectively. After PBS elution, secondary antibody IgG (1:400, Abcam, USA, cat # ab6721) was incubated at room temperature. Tissue sections were rst stained with DAB and then with hematoxylin. Finally, images were collected and analyzed by optical microscope.

RNA sequencing analysis
First, RNA was extracted and tested for quality as described above. And then, RNA sequencing analysis was performed by Genechem (Shanghai, China). Library for RNA sequencing was constructed from TruSeq Stranded mRNA LT Sample Preparation Kit (Illumina, San Diego, California, USA) according to the instructions of manufacturer, and scan it by Affymetrix Scanner 3000 (Affymetrix, Santa Clara, California, USA). DEGs were determined between the two groups based on thresholds of |Fold Change| ≥ 2.0 and FDR < 0.05. IPA (Qiagen, Hilden, Germany) based on all DEGs to analyze rich functional annotations. Z-score ≥ 2 meant that the pathway was signi cantly activated; otherwise the pathway was signi cantly inhibited.

Statistical analysis
All experiments were performed in triplicate and data were shown as mean ± SDs. Statistical analyses and graphs were performed by GraphPad Prism 6.01 (Graphpad Software) and P value < 0.05 as statistically signi cant. The signi cance differences between groups were determined using the twotailed Student's t test or One-way ANOVA analysis. PSMC2 expression in glioma tissues and normal tissues revealed in IHC assay were analysis with Sign test. Mann-Whitney U analysis and Spearman rank correlation analysis were used while explaining the relationships between PSMC2 expression and tumor characteristics in patients with glioma.   The effects of PSMC2 knockdown on proliferation, apoptosis of glioma cells (A) MTT assay was performed to detect the cell proliferation rate of U87 and U251 cells transfected with shCtrl or shPSMC2.
(B) Colony formation assay was used to evaluate cell cloning capacity of U87 and U251 cells transfected with shCtrl or shPSMC2. (C) Flow cytometry analysis was performed to show the apoptotic cell percentage of U87 and U251 with or without PSMC2 knockdown (D) Human apoptosis antibody array was utilized to illustrate the regulation of the expression of apoptosis related proteins by PSMC2 knockdown. The data was expressed as mean ± SD (n ≥ 3), *P<0.05, **P<0.01, ***P<0.001.   independent experiments. The data was expressed as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001