KIF15 promotes human glioblastoma progression under the synergistic transactivation of REST and P300

Abstract

Glioblastoma (GBM) is highly invasive and lethal. The failure to cure GBM highlights the necessity of developing more effective targeted therapeutic strategies. KIF15 is a motor protein to be involved in cell mitosis promotion, cell structure assembly and cell signal transduction. The precise biological function and the potential upstream regulatory mechanisms of KIF15 in GBM remain elusive. Here, we demonstrated that KIF15 was abnormally up-regulated in GBM and predicted poor prognosis of GBM patients. KIF15 promotes GBM cell proliferation, metastasis and cell cycle progression. REST could bind to KIF15 promoter and transactivate KIF15. Furthermore, REST interacts with P300 and depends on its histone acetyltransferase (HAT) activity to co-regulate KIF15 expression. Both REST and P300 were highly expressed in GBM and predicted poor prognosis of GBM patients alone or in combination with KIF15. The tumorigenic function of KIF15 in GBM was regulated by REST in vitro and in vivo and the combinational treatment of cell cycle inhibitor Palbociclib with P300 HAT inhibitor inhibited GBM xenografts survival more significantly. Our findings indicate that KIF15 promotes GBM progression under the synergistic transactivation of REST and P300. P300/REST/KIF15 signaling axis is expected to be served as a cascade of candidate therapeutic targets in anti-GBM.

Keywords: Acetylation; Glioblastoma (GBM); KIF15; P300; REST.

Figure 1

KIF15 is highly expressed in glioblastoma and promotes cancer cell proliferation and metastasis. (A) KIF15 is highly expressed in 12 kinds of different cancers based on Oncomine database, including Brain and CNS Cancer. (B) KIF15 expression is dramatically upregulated in glioblastoma compared to normal brain tissue in TCGA and GTEx databases. (C) KIF15 expression in three cases of brain tumors and the corresponding adjacent normal brain tissues and quantified presentation of its expression according to the IHC score. (D) The overall survival of GBM patients with KIF15 high and low expression was analyzed according to GEPIA databases. (E) The expression of KIF15 increases with the grade of glioma, and it's the highest in WHO grade IV glioblastoma in GEO datasets. (F) KIF15 expression in different glioblastoma cell lines was detected by western blot. (G) KIF15 expression was knocked down by its specific siRNAs and detected by western blot in U87MG and T98G cells. (H) Cell viability was detected by MTT assay after cells were cultured without transfection (Mock) or transfected with non-specific control (siCtrl), or KIF15 specific siRNAs respectively. (I-K) Colony formation assay (I), gap-closure assay (J) and transwell assay (K) of U87MG and T98G cells transfected with KIF15 specific siRNAs for 48 hours and the corresponding statistical analysis. (L) Western blot assay for expression of p-c-Raf, Erk1/2, p-Erk1/2, p-AKT (s473), AKT and β-actin in U87MG and T98G cells after KIF15 were knocked down. (M) Western blot assay for the expression of N-cadherin, Vimentin, Slug and Snail in U87MG and T98G cells after KIF15 was knocked down. The data represent the mean±SD of three independent experiments, and the level of significance was indicated by *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Mock: blank control. siCtrl: control siRNA.

Figure 2

Knockdown of KIF15 induces G0/G1 phase arrest and sensitizes glioblastoma cells to Palbociclib treatment. (A) Representative cell cycle plots of U87MG and T98G cells at 48 h after KIF15 knockdown. (B) The expressions of KIF15 and cell cycle-associated proteins were analyzed by western blot after KIF15 was knocked down for 48 h. (C) U87MG and T98G cells were transfected with KIF15 specific siRNAs or control siRNAs for 24 h, and then were treated by different concentration of Palbocilib for another 24 h. Cell viability was determined by MTT assay and IC50s were calculated. (D-E) Colony formation (D) and cell cycle assay (E) were performed respectively in U87MG and T98G cells with KIF15 knockdown and treatment of 10µM or 15µM Palbocilib sequentially. The corresponding quantified analysis was also performed. The data represents the mean±SD of three independent experiments, and the level of significance was indicated by *P<0.05, **P<0.01, ***P<0.001. NC: Negative control. PB: Palbociclib.

Figure 3

REST binds to KIF15 promoter and transactivates KIF15 to promote glioblastoma progression in vitro. (A) Schematic diagram of biotinylated KIF15 promoter probe was shown. (B) Biotin-labeled KIF15 promoter probe was synthesized by PCR. (C) Silver staining analysis of KIF15 promoter-binding proteins separated by SDS-PAGE. (D) The overlaps of the pulled-down proteins by KIF15 promoter probe in U87MG and T98G cells and the binding proteins at KIF15 promoter region in these two cells from ChIP-seq data. (E) Expressions of REST and KIF15 in U87MG and T98G cells after silencing REST with its specific shRNAs were measured by RT-PCR and Western blot. (F) The positive correlation between REST and KIF15 expression was analyzed in TCGA database based on their mRNA level. (G) REST was pulled down by KIF15 promoter probe in different GBM cells or in T98G cells with REST knockdown by DNA-pulldown assay. (H) The binding sequence recognized by REST in KIF15 promoter. (I) ChIP assay was performed using REST specific antibody or IgG to detect KIF15 promoter sequence from -402 to -287 (product length=116 bp) in different GBM cells or in T98G cells with REST knockdown. (J) The expressions of REST and KIF15 were determined by Western blot in U87MG and T98G cells after REST was overexpressed and/or KIF15 was knocked down. (K) MTT assay was performed in U87MG and T98G cells with REST overexpression and/or KIF15 knockdown. (L) Transwell assay was performed in U87MG and T98G cells with REST overexpression and/or KIF15 knockdown, and the invasive cells were photographed and calculated. The data represent the mean±SD of three independent experiments, and the level of significance was indicated by *P<0.05, **P<0.01, ***P<0.001.

Figure 4

REST promotes the growth of glioblastoma xenografts in mice partially by targeting KIF15. (A) The formed tumors were stripped and photographed at the end of experiment. (B-C) The mice of each group were sacrificed, and the tumor weight and tumor volume was measured respectively. (D) The tumor diameter was measured at interval of 2 days after cell injection for 10 days and the tumor volume was calculated. (E) The expressions of REST, KIF15, p-AKT, AKT, Snail, Slug, Cyclin D1, and β-actin in tumor xenografts were measured by western blot. (F) REST, KIF15, p-AKT, Slug, and Ki67 in tumor xenografts were detected by immunohistochemistry staining and representative images were shown. Scale bars=50µm. Original magnification: ×40. (G) Statistical analysis for the expression of REST, KIF15, p-AKT, Slug and Ki67 in U87MG xenografts based on IHC staining, n=5. Data are presented as means±SD, Two-tailed unpaired Student's T-test. The level of significance was indicated by *P<0.05, **P<0.01, ***P<0.0001.

Figure 5

P300 and its histone acetyltransferase activity were necessary for the transcriptional activation of KIF15 mediated by REST. (A) The binding peaks of P300 and REST at the same region of KIF15 promoter in GBM cells from ChIP-seq data in Cistrome Browser. (B) The predicted interactions between REST and other proteins, including P300, by STRING databases. (C) Immunofluorescence assay was used to detect the colocalization of REST and P300 in different glioblastoma cells. (D) The interaction between REST and P300 in different GBM cells was detected by co-immunoprecipitation. (E) The schematic image of P300 plasmids with different domains was shown. (F) Co-IP assay was used to detect the interaction between REST and P300 with different truncates in U87MG cells. (G) The interaction of P300 HAT domain (brown) with the potential sites (colored) of REST residues (150-156 amino acids) was predicted by HDOCK software. (H) Co-IP assay was performed to detect the interaction between P300 and REST after U87MG and T98G cells were treated with DMSO or C646 for 24 h. (I) The binding of P300 and REST in KIF15 promoter was analyzed by DNA pulldown assay with KIF15 promoter probe (-737/-38) in U87MG cells with different treatment. (J) P300 antibody or IgG was used to immunoprecipitate KIF15 promoter fragment (-402 to -287) in different GBM cells by ChIP assay. (K) P300 antibody, REST antibody or IgG was used to immunoprecipitate KIF15 promoter fragment (-402 to -287) in U87MG cells after transfection with P300, ΔP300 plasmids or treated with C646. (L) The expressions of P300 and KIF15 in U87MG cells transfected with P300, ΔP300 plasmids or treated with C646 were detected by western blot and RT-PCR. The data represent the mean±SD of three independent experiments. ΔP300: P300 histone acetyltransferase domain delete mutation.

Figure 6

REST synergizes with P300 to co-regulate KIF15 expression and glioblastoma cell malignancy. (A) The expressions of P300, REST and KIF15 were detected in U87MG and T98G cells with P300 overexpression and/or REST knockdown. (B-D) MTT (B), transwell (C), and cell cycle assay (D) was performed respectively in U87MG and T98G cells with P300 overexpression and/or REST knockdown. The corresponding quantified analysis was also performed. (E) The expressions of P300, REST and KIF15 were detected in U87MG and T98G cells with C646 treatment (10µM) and/or REST overexpression. (F-G) MTT assay (F) and transwell assay (G) was performed in U87MG and T98G cells treated with C646 and/or REST overexpression. The data represent the mean±SD of three independent experiments, and the level of significance was indicated by *P<0.05, **P<0.01, ***P<0.001.

Figure 7

C646 combined with Palbociclib produced synergistic anti-tumor effect in glioblastoma xenograft model. (A) The representative images of tumor xenografts from mice treated with indicated agents. (B-C) Tumor weight (B) and volume (C) was respectively measured or calculated after mice were sacrificed. (D) Tumor volumes of NYG mice bearing U87MG xenografts were measured at the 6th day after cell injection every two days. (E) The changes in body weight for mice bearing xenografts were recorded during the whole treatment process. (F) Blood Urea Nitrogen (BUN) levels of mice in each group were measured by the BUN kit (n=6). (G) The expressions of P300, KIF15, p-RB, CyclinD1, p-AKT, and Ki67 were analyzed by immunohistochemical staining in U87MG xenografts after receiving different treatments (n=5) and the representative images and the corresponding quantified analysis were also shown. Scale bars, 50µm. Data are presented as means±SD, Two-tailed unpaired Student's T-test. (H) The expressions of P300, REST, KIF15, p-AKT, Cyclin A2, Cyclin D1, CDK4, CDK6 and GAPDH from tumor tissue lysates were analyzed by Western blot. The data represent the mean±SD of three independent experiments, and the level of significance was indicated by *P<0.05, **P<0.01, ***P<0.001. PB: Palbociclib; Comb: Combination.

Figure 8

Schematic diagram of KIF15 regulate mechanism in GBM. We demonstrated that REST recruits P300 to co-anchor at the promoter region of KIF15 in GBM cells to synergistically regulate KIF15 transcription and subsequent tumor progression, during which the HAT activity of P300 has been proved necessity to potentiate KIF15 promoter activity.