Acquired temozolomide resistance in MGMT-deficient glioblastoma cells is associated with regulation of DNA repair by DHC2

Abstract

The acquisition of temozolomide resistance is a major clinical challenge for glioblastoma treatment. Chemoresistance in glioblastoma is largely attributed to repair of temozolomide-induced DNA lesions by O6-methylguanine-DNA methyltransferase (MGMT). However, some MGMT-deficient glioblastomas are still resistant to temozolomide, and the underlying molecular mechanisms remain unclear. We found that DYNC2H1 (DHC2) was expressed more in MGMT-deficient recurrent glioblastoma specimens and its expression strongly correlated to poor progression-free survival in MGMT promotor methylated glioblastoma patients. Furthermore, silencing DHC2, both in vitro and in vivo, enhanced temozolomide-induced DNA damage and significantly improved the efficiency of temozolomide treatment in MGMT-deficient glioblastoma. Using a combination of subcellular proteomics and in vitro analyses, we showed that DHC2 was involved in nuclear localization of the DNA repair proteins, namely XPC and CBX5, and knockdown of either XPC or CBX5 resulted in increased temozolomide-induced DNA damage. In summary, we identified the nuclear transportation of DNA repair proteins by DHC2 as a critical regulator of acquired temozolomide resistance in MGMT-deficient glioblastoma. Our study offers novel insights for improving therapeutic management of MGMT-deficient glioblastoma.

Keywords: CBX5; DHC2; XPC; acquired TMZ resistance; glioblastoma.

Figure 1

Analysis of the correlation of DHC2 expression to MGMT expression and TMZ treatment responses in GBM patients. (A) DYNC2H1/DHC2 expression across glioma tissues of different grades and healthy tissues from the tumour Glioma French-284 database. ns = no significance. (B) DYNC2H1/DHC2 expression across different GBM subtypes in the Glioblastoma TCGA-540 database. (C) qRT-PCR analysis of DYNC2H1 mRNA expression levels in 21 primary GBMs without TMZ treatment and 21 recurrent GBMs with standard TMZ treatment. GAPDH expression was used for normalization. **P < 0.01. (D) Progression-free (progrfree) survival analysis of GBM patients without MGMT promoter methylation using the Glioblastoma TCGA-540 database. (E) Progression-free survival analysis of GBM patients with MGMT promoter methylation using the Glioblastoma TCGA-540 database. (F) Co-expression analysis of DYNC2H1 and MGMT mRNA expression using the Glioblastoma TCGA-provisional database. (G and H) Representative images of immunohistochemical analysis of DYNC2H1 and MGMT in six paired primary and recurrent GBM tissues. Scale bars = 100 μm. H&E = haematoxylin and eosin.

Figure 2

DHC2 mediates TMZ-induced DNA damage repair of MGMT-deficient GBM cells. (A) Wild-type (WT) or CRISPR/Cas9-mediated DYNC2H1/DHC2 knockout (−/−) U87 cells were treated with 200 μM TMZ for 24 h, then the TMZ-contained culture media were replaced with common culture media and the TMZ-treated GBM cells were allowed to repair for indicated time points, and followed by western blotting analysis of cellular lysates for the γH2AX expression, using GAPDH as loading control. pre = pretreatment, indicates U87 cells without TMZ treatment. (B) The same experimental processes were conducted in shNC (normal control shRNA-transduced) and shDHC2 (DHC2 knockdown shRNA-transduced) GBM1 cells (400 μM TMZ for treatment), western blotting analysis of γH2AX expression and GAPDH served as the internal loading control. (C and D) DYNC2H1DHC2(WT) or DYNC2H1/DHC2(−/−) U87 cells (C) and shNC or shDHC2-transduced GBM1 cells (D) were treated with the indicated concentration of TMZ (200 μM for U87, 400 μM for GBM1) for 24 h and recover for 6 h with TMZ washing out. Immunofluorescence staining of γH2AX foci and the percentage of cells containing >10 γH2AX foci in 10 random microscopic fields was calculated. Scale bar = 5 μm. **P < 0.01, ***P < 0.001, ns = no significance. (E and F) Representative images of subcutaneous xenografts in nude mice derived from GBM1 shNC and shDHC2 cells (E), or U87 DHC2(WT) and DHC2(−/−) cells (F) with or without TMZ treatment (intraperitoneal injection of 20 mg/kg TMZ in saline) on the 42nd day, shNC-vehicle or DHC2(WT)-vehicle group served as the control and the comparison symbol above the bar represented the statistical results compared with the control group. *P < 0.05; **P < 0.01; ***P < 0.001; ns = no significance. (G) T₂-weighted MRI of intracranial xenografts (arrows) in mice bearing DHC2(WT) or DHC2(−/−) U87 cells before and after treatment with vehicle alone or TMZ. (H) Brain images, haematoxylin and eosin staining and immunohistochemical analyses for DHC2 and γH2AX expression in sections of representative intracranial tumour-bearing mice from each treatment arm. Scale bar = 1 mm in haematoxylin and eosin, 25 μm in immunohistochemistry images. (I) Survival curve of mice with DHC2(WT) and DHC2(−/−) U87 cell-derived intracranial xenografts treated with vehicle or TMZ. (J) Schematic representation of the therapeutic schedule followed. In each treatment course, 20 mg/kg TMZ/vehicle was injected intraperitoneally into mice, and 10⁹ units of shDHC2/shNC adenovirus were injected intratumourally. (K) Survival curve of mice harbouring the U87 cell-derived intracranial xenografts and subjected to the therapeutic schedule in J.

Figure 3

Proteomic analysis to identify potential regulatory partners of DHC2 and subsequent validation of targets. (A) Schematic representation of protocol followed for proteomic analysis. shNC and shDHC2 U87 cells were treated with 200 μM TMZ or dimethyl sulphoxide (DMSO) for 1 week, the nuclear fractions were harvested for proteomic analysis, experimental procedures of mass spectrometry proteomic were described in ‘Materials and methods’ section. (B) Schematic of DHC2 downstream analysis. First, to select the related proteins with TMZ-treatment response in GBM cells, DEPs were identified by meeting either of the following criteria: (i) with a significant fold change of >2 in the U87-TMZ group compared with the U87-DMSO group (U87 shNC-TMZ/U87 shNC-DMSO > 2); or (ii) only present in the U87-TMZ group but not in the U87-DMSO group (indicated proteins related to TMZ treatment). A total of 184 DEPs were identified in this study. Next, we sorted proteins correlated with DHC2 expression among these 184 DEPs. We filtered the 184 DEPs with >1.5-fold changes in the U87-TMZ group compared with the DHC2 knockdown U87-TMZ group (U87 shNC-TMZ/U87 shDHC2-TMZ >1.5; also indicated DEPs downregulated in DHC2 knockdown U87-TMZ group compared with U87-TMZ group), and a total of 45 DEPs were identified as candidate regulatory proteins of DHC2. Finally, we analysed the TCGA GBM-540 DYNC2H1/DHC2 mRNA co-expression dataset and set the cut-off value to a Pearson score >0.6 to identify the DEPs that most strongly correlated with DHC2 expression. Both XPC and CBX5 were then identified. (C) Immunofluorescence staining to detect the localization of XPC and CBX5 in DHC2(WT) and DHC2(−/−) U87 cells after 200 μM TMZ treatment or DMSO treatment for 1 week. Scale bar = 20 μm. (D) Western blotting analysis of XPC and CBX5 expression in cytoplasmic (C) and nuclear (N) fractions of DHC2(WT) and DHC2(−/−) U87 cells after TMZ or DMSO treatment for 1 week. GAPDH and Histone H3 served as the internal loading controls. (E) Western blotting analyses of XPC, CBX5, MGMT and γH₂AX expression in six primary and recurrent GBM tissues. GAPDH served as the internal loading control. (F and G) Representative images of immunohistochemical staining of XPC and CBX5 in six paired primary and recurrent GBM tumours. Patient 1: MGMT-deficient (F), Patient 2: MGMT-positive (G). Scale bar = 100 μm.

Figure 4

DHC2 mediates nuclear transportation of XPC and CBX5 into the nucleus. (A and B) Immunofluorescence staining to examine co-localization of XPC, CBX5 and DHC2 in U87 cells (A) and GBM1 cells (B) after indicated TMZ treatment (200 μM for U87, 400 μM for GBM1) for 48 h. Scale bar = 20 μm. (C) Schematic representation of the domains in full-length DHC2 protein with amino acid (aa) residue numbering; the stem (aa 1–1650) and stalk (aa 2881–3169) domains are as indicated. Five fragments of DHC2 were constructed: #1 (aa 1–360), #2 (aa 359–979), #3 (aa 980–1584), #4 (aa 1582–2052) and #5 (aa 2053–2515), with His and Flag tags. (D) The five fragments of DHC2 were each transfected into U87 cells followed by treatment with 200 μM TMZ for 48 h. Empty plasmid transfected U87 cells with the same treatment served as the normal control (NC). Cell lysates were immunoprecipitated (IP) with anti-His antibody-conjugated magnetic beads and immunoblotted with anti-His, anti-XPC and anti-CBX5 antibodies. (E) qRT-PCR for analysis of XPC and CBX5 mRNA expression in DHC2(WT) and DHC2(−/−) U87 cells after 200 μM TMZ treatment for 1 week. GAPDH expression was used for normalization. ***P < 0.001. (F and G) FLAG-tagged ubiquitin was transfected into U87 cells, in the presence or absence of 200 µM TMZ treatment for 48 h, followed by treatment with 10 mM MG132 or vehicle for 6 h. Cellular lysates were harvested and immunoprecipitated (IP) with anti-FLAG antibody, and the ubiquitin-conjugated XPC (F) and CBX5 (G) were subsequently detected by immunoblotting (IB) using the indicated antibodies.

Figure 5

Both XPC and CBX5 were involved in repair of TMZ-mediated DNA damage. (A) U87 and GBM1 cells were transfected with siRNA using Lipofectamine™ 2000. Cellular lysates were harvested at 72 h after transfection, followed by immunoblotting analysis for the efficiency of XPC and CBX5 knockdown by RNA interference. siNC = normal control siRNA. (B) Immunoblotting for XPC and CBX5 in U87 DHC2(WT) and DHC2(−/−) cells demonstrating effective overexpression (oe). NC = normal control; empty plasmid transfected. (C) GBM1 and U87 cells with XPC or CBX5 knockdown were treated with TMZ (200 μM for U87, 400 μM for GBM1) for 24 h and allowed to recover for 6 h. Immunofluorescence staining of γH2AX foci, and the percentage of cells containing >10 γH2AX foci in 10 random microscopic fields was calculated. Scale bar = 5 μm. ***P < 0.001. (D) U87 DHC2(WT) and DHC2(−/−) cells with XPC or CBX5 overexpression were treated with 200 μM TMZ for 24 h and allowed to recover for 6 h. Immunofluorescence staining of γH2AX foci and percentage of cells containing >10 γH2AX foci in 10 random microscopic fields was calculated. Scale bar = 5 μm. ***P < 0.001, ns = no significance. (E) GBM1 and U87 cells were treated with the indicated concentration of TMZ (200 μM for U87, 400 μM for GBM1) for 24 h. Immunofluorescence staining of γH2AX foci and XPC shows co-localization. Scale bar = 10 μm. (F) Immunofluorescence staining to examine co-localization of XPC and CBX5 in U87 and GBM1 cells after TMZ treatment (200 μM for U87, 400 μM for GBM1) for 48 h. Scale bar = 20 μm. (G) U87 cells were exposed to 200 µM TMZ or DMSO for 48 h, and cell lysates were immunoprecipitated (IP) with anti-XPC or anti-CBX5 antibody and immunoblotted with the indicated antibodies.

Figure 6

A mechanistic model for DHC2-mediated acquired TMZ resistance in MGMT-negative GBM cells. In MGMT-positive GBM cells, DHC2 is low or not expressed. MGMT mediates TMZ-induced DNA damage repair and results in TMZ-resistance. In MGMT-deficient GBM cells upon TMZ treatment, DHC2 is upregulated and mediated nuclear transportation of DNA repair proteins XPC and CBX5, further contributes to DNA damage repair and TMZ-resistance. Once DHC2 is depleted in MGMT-deficient GBM cells, XPC and CBX5 are degraded via a proteasome-dependent pathway, and result in persistent DNA damage and improvement of TMZ sensitivity.