Glioblastoma (GB) is the highest grade and most common form of primary adult brain tumors. Despite surgical removal followed by concomitant radiation and chemotherapy with the alkylating agent temozolomide, GB tumors develop treatment resistance and ultimately recur. Impaired response to treatment occurs rapidly, conferring a median survival of just fifteen months. Thus, it is necessary to identify the genetic and signaling mechanisms that promote tumor resistance to develop targeted therapies to combat this refractory disease. Previous observations indicated that SGEF (ARHGEF26), a RhoG-specific guanine nucleotide exchange factor (GEF), is overexpressed in GB tumors and plays a role in promoting TWEAK-Fn14–mediated glioma invasion. Here, further investigation revealed an important role for SGEF in glioma cell survival. SGEF expression is upregulated by TWEAK-Fn14 signaling via NF-κB activity while shRNA-mediated reduction of SGEF expression sensitizes glioma cells to temozolomide-induced apoptosis and suppresses colony formation following temozolomide treatment. Nuclear SGEF is activated following temozolomide exposure and complexes with the DNA damage repair (DDR) protein BRCA1. Moreover, BRCA1 phosphorylation in response to temozolomide treatment is hindered by SGEF knockdown. The role of SGEF in promoting chemotherapeutic resistance highlights a heretofore unappreciated driver, and suggests its candidacy for development of novel targeted therapeutics for temozolomide-refractory, invasive GB cells.
Implication: SGEF, as a dual process modulator of cell survival and invasion, represents a novel target for treatment refractory glioblastoma. Mol Cancer Res; 14(3); 302–12. ©2016 AACR.
Glioblastoma (GB) is the most common form of primary adult brain tumors characterized by a poorly delineated tumor mass resulting from highly invasive cells. The problem of resistance to the standard antiproliferative treatment of concomitant radiotherapy with chemotherapy using the alkylating agent temozolomide is common, and actively invading cells survive the current therapeutic regimens. Glioma cells with the increased capacity for migration have a decreased expression of proapoptotic genes and are less sensitive to cytotoxic therapy–induced apoptosis (1-4); the knockdown of several proinvasive gene candidates in GB decreases glioma cell migration rate and sensitizes the cells to cytotoxic therapy and importantly, therapy directed at mediators of invasion has been shown to increase chemotherapeutic sensitivity (5, 6).
An increased capacity for cell survival results from the multifaceted regulation of pathways involved in promoting cell growth, replication and spread, and preventing apoptosis in response to cytotoxic insult (7). Treatment strategies of tumor irradiation and temozolomide administration in glioblastoma lead to the formation of DNA double strand breaks (DSB), either directly, or via mismatch repair conversion of O(6)-methylguanine adducts into DSBs, respectively (8). DSBs are primarily repaired through two mechanisms, homologous recombination (HR) and nonhomologous end-joining (NHEJ). HR repair makes use of a non-damaged homologous DNA template, and thus is characterized as an error-free mechanism, while NHEJ has no homologous strand for template use resulting in sequence errors near the break point (9).
DNA repair is initiated via sensing of DSBs by three kinases: ataxia telangiectasia mutated (ATM), ataxia telangiectasia and Rad3 related (ATR), and Chk2. Subsequently, the early phosphorylation of histone H2A.X (γH2A.X) by ATM occurs at damaged DNA foci and leads to the phosphorylation of mediator of DNA damage checkpoint protein 1 (MDC1), with subsequent chromatin remodeling and recruitment of DNA repair proteins (10). BRCA1 is one such key mediator of HR and NHEJ repair; after exposure to DNA-damaging agents BRCA1 is rapidly phosphorylated by ATM, ATR, and Chk2, and relocated to sites of replication forks with γH2A.X foci, where it recruits further proteins including BRCA2 and Rad51 to mediate strand exchange toward DNA repair and cell survival (9).
One key driver in GB that has been characterized to promote both cell invasion and cell survival is the transmembrane receptor fibroblast growth factor inducible-14 (Fn14). Fn14 is a member of the TNF receptor superfamily with one known ligand, the TNF-like weak inducer of apoptosis (TWEAK). Signaling through Fn14 by its cytokine ligand TWEAK activates Rac1, Akt, and NF-κB pathways, and has been shown to promote increased cell invasion and resistance to cytotoxic therapy-induced apoptosis (3, 4, 11).
Here we show that the src-homology 3 domain containing GEF (SGEF) promotes cell survival in response to temozolomide treatment. In GB tumors, SGEF has been shown to be significantly overexpressed, to be correlated with poor patient outcome, and to promote glioma cell migration (12). We report that SGEF expression is increased in a subset of temozolomide-resistant (TMZ-R) derived primary GB xenografts and is upregulated by TWEAK-Fn14 signaling via NF-κB activity. Moreover, levels of SGEF and Fn14 mRNA are positively correlated in GB tumor specimens. Depletion of SGEF impairs colony formation following temozolomide treatment and increases cell susceptibility to temozolomide-induced apoptosis. temozolomide treatment of glioma cells both leads to increased nuclear activation of SGEF and the SGEF dependent phosphorylation of the DNA damage repair protein BRCA1. SGEF and BRCA1 are found in a complex upon temozolomide treatment. SGEF may thus be an important mediator of pro-survival signaling in response to temozolomide-induced DNA damage.
Materials and Methods
Cell culture conditions
Human astrocytoma cell lines U87, U118, and T98G (ATCC), as well as primary glioblastoma xenograft cells (GBM and GBM TMZ-R lines) were maintained in DMEM (Gibco) supplemented with 10% heat-inactivated FBS (Gibco) at 37°C with 5% CO2. For all assays with TWEAK treatment, cells were cultured in reduced serum (0.5% FBS) for 16 hours before stimulation with recombinant TWEAK at 100 ng/mL in DMEM + 0.1% bovine serum albumin for the indicated times.
Antibodies, plasmids, reagents, and Western blot analysis
A polyclonal SGEF antibody was purchased from Sigma. A monoclonal tubulin antibody was purchased from Millipore. A polyclonal antibody for phospho-BRCA1 (Ser1524), and mAbs for BRCA1, cleaved PARP, phospho-histone H2A.X (Ser139), Histone H2A.X, Rabbit IgG (isotype control), Histone H3, and NF-κB p65 were purchased from Cell Signaling Technologies. Lipofectamine RNAiMax was purchased from Invitrogen. Human recombinant TWEAK was purchased from PeproTech. Human placental laminin and temozolomide were obtained from Sigma. In certain experiments, glioma cells were transiently transfected with either IκBα wild-type (IκBαWT) or IκBαS32/36A mutant (IκBαM) super-repressor expressing plasmids (Addgene) using the Effectene transfection protocol (Qiagen) for 24 hours prior to culture in reduced serum medium (0.5% FBS DMEM) for 16 hours with subsequent addition of TWEAK for 4 hours. Plasmids: pGEX4T-1-RhoG(15A) was obtained from Dr. Keith Burridge (University of North Carolina at Chapel Hill, Chapel Hill, NC).
For immunoblotting, cells were lysed in 2× SDS sample buffer (0.25 mol/L Tris-HCl, pH 6.8, 10% SDS, 25% glycerol) containing 10 μg/mL aprotinin, 10 μg/mL leupeptin, 20 mmol/L NaF, 2 mmol/L sodium orthovanadate, and 1 mmol/L phenylmethylsulfonyl fluoride. Protein concentrations were determined using the BCA assay (Pierce) with bovine serum albumin as a standard. Thirty micrograms of total protein were loaded per lane and separated by SDS PAGE. After 4°C transfer, the nitrocellulose (Invitrogen) was blocked with either 5% nonfat milk or 5% BSA in Tris-buffered saline, pH 8.0, containing 0.1% Tween 20 (TBST) prior to addition of primary antibodies and followed with peroxidase-conjugated anti-mouse IgG or anti-rabbit IgG. Protein was detected using SuperSignal West Dura Chemiluminescent Substrate (Thermo Scientific) with a UVP BioSpectrum 500 Imaging System. Densitometry was calculated via ImageJ software.
RNA isolation and quantitative reverse transcriptase-PCR
Total RNA was isolated as previously described (1). cDNA was synthesized from 500 ng of total RNA in a 20 μL reaction volume using the SuperScript III First-Strand Synthesis SuperMix Kit (Invitrogen) for 50 minutes at 50°C, followed by 85°C for 5 minutes. qPCR analysis of SGEF (sense: 5′-TGC TGA AAG GAC AAG GAA CA-3′; antisense: 5′-GTA GTT TTG ATA CAG GAC AGC ATT-3′) and histone H3.3 (sense: 5′- CCA CTG AAC TTC TGA TTC GC-3′; antisense: 5′-GCG TGC TAG CTG GAT GTC TT-3′) mRNA levels was conducted using SYBR green (Roche) fluorescence for detection of amplification after each cycle with LightCycler analysis software and quantified as described previously (3).
Biotinylated electrophoretic mobility shift assay
T98G glioma cells were plated at a density of 3 × 106 in 100 mm2 tissue culture dishes in normal growth medium. After 12 hours, cells were cultured under reduced serum (0.5% FBS) for an additional 16 hours before TWEAK (100 ng/mL) addition for 2 hours. Isolation of cell nuclear protein was carried out using the NE-PER kit (Pierce) according to the protocol of the manufacturer. Protein–DNA complexes were detected using biotin end–labeled double-stranded DNA 23-mer probes containing the NF-κB–binding sites within the SGEF promoter (NF-κB-SGEF wt target sequence: 5′- GTC TAG GAG GCA AAT CCC AGA AA -3′; NF-κB-SGEF mt target sequence: 5′- GTC TAG GAG CCA GAT CGC AGA AA -3′). The binding reactions were done using the LightShift kit (Pierce) according to the manufacturer's protocol. Where indicated, 200-fold molar excess of unlabeled NF-κB-SGEF wt oligonucleotides or anti-p65 antibody was included. The reaction products were resolved by gel electrophoresis and detected by chemiluminescence according to the manufacturer's protocol (Pierce).
Lentiviral vectors containing shRNA-targeting SGEF (shSGEF-12 and shSGEF-13) or control empty vector (control) were obtained from Open Biosystems (Fisher Scientific) and packaged for lentiviral production as described previously (12).
Clonogenic and apoptosis studies
Observations of colony-forming capacity following cytotoxic insult were performed as described (13). Briefly, T98G, U87, and U118 cells stably expressing either control or shRNA-targeting SGEF were treated with temozolomide (500 μmol/L). In certain experiments, cells were additionally transfected with siRNA targeting control luciferase or BRCA1 for 72 hours prior to the addition of temozolomide. Cells were trypsinized 24 hours post-temozolomide treatment and plated in triplicate in 6-well cell culture dishes at 250 cells per well. Colonies were allowed to grow until controls reached a 50 cell density (approximately 6–7 days) before being fixed briefly in a 10% (v/v) methanol, 10% (v/v) glacial acetic acid solution, stained with a 0.5% (w/v) crystal violet solution and washed with deionized water. Apparent colonies were recorded, and surviving fractions were determined relative to the nontreated control for each cell line.
For apoptotic studies, T98G and U87 control or shSGEF cells were treated with temozolomide (500 μmol/L) for 48 hours and whole cell lysates were analyzed for cleaved PARP and caspase-3 by Western blot analysis. For chromatin condensation studies, apoptotic cells were evaluated by nuclear morphology of DAPI-stained cells as described previously (14). Briefly, glioma cells were plated onto 10-well slides precoated with 10 μg/mL laminin. After 24 hours, cells were treated with temozolomide (500 μmol/L) for an additional 48 hours. Cells were fixed with 4% paraformaldehyde, stained with ProLong Gold Antifade Reagent with DAPI (Molecular Probes), and evaluated by nuclear morphology. Cells with condensed, fragmented chromatin were manually scored as apoptotic cells. At least three fields were evaluated per well and data reported as apoptotic cells/total cells × 100.
Nucleotide-free GEF pulldowns
RhoG activity was measured as previously described using a GST-ELMO-NT fusion protein (15). Affinity pulldowns of active SGEF bound to RhoG were performed using a nucleotide-free RhoG mutant (G15A) expressed and purified as described (16). Recombinant RhoG G15A-GST protein and GST-ELMO-NT were produced in Escherichia coli (BL21) cells. Cells were lysed in B-PER lysis buffer (Pierce) containing protease inhibitors and purified with glutathione sepharose beads (GE Healthcare).
Isolation of cell nuclear protein was performed according to Guilluy and colleagues (17). Briefly, 107 U87 cells were grown in 10-cm dishes before treatment with temozolomide (500 μmol/L) for the indicated times. Cells were washed in ice-cold PBS containing protease inhibitors and lysed in a 1 mL hypotonic solution [10 mmol/L HEPES (pH 7.9), 1.5 mmol/L MgCl2, 10 mmol/L KCl, and 0.5 mmol/L DTT (freshly added before use)]. Lysates were homogenized and centrifuged at approximately 300 × g for 5 minutes at 4 °C. Pellets were washed twice in 1.5 mL of a 30% (w/v) iodixanol solution and centrifuged at 10,000 × g at 4 °C. Supernatants were discarded and the pellets were resuspended in 300 mL Rho GEF buffer [20 mmol/L HEPES (pH 7.5), 150 mmol/L NaCl, 5 mmol/L MgCl2, 1% (v/v) Triton X-100, 1 mmol/L DTT with protease inhibitors]. Suspensions were sonicated briefly, centrifuged at 14,000 × g for 5 minutes at 4 °C, and the remaining nuclear fraction supernatants were quantified for protein concentration via BCA assay. Subsequently, equal amounts of total GST fusion protein were incubated with fresh nuclear protein lysate (1 mg) for one hour, and precipitated lysates were resuspended in 2× SDS buffer containing protease inhibitors and resolved with SDS-PAGE.
Subsequently, equal amounts of total GST fusion protein were then incubated with nuclear protein lysate (1 mg), and precipitated lysates were resolved with SDS-PAGE.
U87 cells were treated with 500 μmol/L temozolomide for the indicated times prior to lysis on ice in a buffer containing 10 mmol/L Tris-HCl (pH 7.4), 0.5% Nonidet P-40, 150 mmol/L NaCl, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L EDTA, 2 mmol/L sodium orthovanadate, 20 mmol/L sodium fluoride, 10 μg/mL aprotinin, and 10 μg/mL leupeptin. Isolation of cell nuclear protein was carried out using the NE-PER kit (Pierce) according to the manufacturer's protocol. Equivalent amounts of protein were precleared and immunoprecipitated from the nuclear lysates using either SGEF or BRCA1 antibodies as indicated, or a control isotype-matched antibody, and then washed with lysis buffer followed by TX-100 buffer [10 mmol/L HEPES (pH 7.4), 150 mmol/L NaCl, 2 mmol/L EDTA, 2 mmol/L EGTA, 20 mmol/L sodium fluoride, and 0.5% Triton X-100]. Samples were then resuspended in 1× LDS sample buffer containing DTT and boiled, separated by SDS-PAGE, transferred to nitrocellulose for 1 hour at 4°C, and then proteins were detected using SuperSignal West Dura Chemiluminescent Substrate (Thermo Fisher Scientific).
siRNA oligonucleotides specific for GL2 luciferase were described previously (18). BRCA1-specific siRNA target sequences are as follows: BRCA1-1 (5′-ACC ATA CAG CTT CAT AAA TAA-3′) and BRCA1-2 (5′- AAC CTA TCG GAA GAA GGC AAG-3′). Transient transfection of siRNA was performed using Lipofectamine RNAiMax. Cells were plated at 70% confluence in DMEM + 10% FBS without antibiotics and were transfected within 8 hours of plating. The siRNA and Lipofectamine were diluted separately in Opti-MEM (Thermo Fisher Scientific). After 5 minutes, the mixtures were combined and incubated for 20 minutes at room temperature to enable complex formation. siRNA oligonucleotides were transfected at 50 nmol/L and no cell toxicity was observed. Maximum inhibition of protein levels was achieved 48 to 72 hours after transfection.
Cell viability was determined via Alamar Blue. Briefly, glioma cells were plated in 96-well plates in quadruplicate for use as a standard curve of known cell counts or plated in replicates of 8 at 3,000 cells per well to monitor proliferation over 72 hours. After cell attachment of the standard curve, or 24 hours postattachment of experimental wells, cells were treated with 10% Alamar Blue (Trek Diagnostic Systems) for 5 hours at 37ºC. The absorbance was read at 560 and 595 nm and the cell viability was expressed as number of cells per well calculated relative to the standard curve for each line. Cell viability was repeatedly assessed daily over 72 hours.
Glioma cells were plated onto 10-well glass slides, precoated with 10 μg/mL laminin. Twenty-four hours later, cells were treated with 500 μmol/L temozolomide for 48 hours and fixed in 4% formaldehyde/PBS, permeabilized with 0.1% Triton X-100 dissolved in PBS, and incubated with antibodies to BRCA1 or phospho-H2A.X (Ser139). Slides were mounted with ProLong Gold Antifade Reagent with DAPI (Molecular Probes). Images were collected using a Zeiss LSM 510 microscope equipped with a 63× objective, ZEN 2009 image analysis software, and Adobe Photoshop CS3.
Gene expression analysis of SGEF and Fn14 correlation
Expression data generated using the Affymetrix U133 Plus 2.0 Array for 82 GB samples was downloaded from the REpository for Molecular BRAin Neoplasia DaTa (REMBRANDT) for the correlation of ARHGEF26 (SGEF) and TNFRSF12A (FN14; ref. 19). The expression level of ARHGEF26 (SGEF) was calculated as the median of the three relative expression intensity values for the three probe sets annotated for ARHGEF26 (SGEF). There was a high correlation of all three ARHGEF26 (SGEF) probes sets to each other (Pearson correlation of 0.90 to 0.97). The relative expression intensity of the probe set 218368_s_at was used for TNFRSF12A (FN14) as it is the only probe set annotated for TNFRSF12A (FN14). The Pearson product-moment correlation was calculated using the R software code as supplied in ref. (20).
Kinetics of histone H2A.X phosphorylation upon temozolomide treatment in GB cells was quantified using flow cytometer analysis. Phospho-H2A.X staining was performed as described previously (21) with minor modifications. Briefly, T98G NS Ctrl and T98G SGEF-12 cells were treated with 500 μmol/L temozolomide for 30 minutes, 2 hours, and 8 hours. Subsequently, cells were trypsinized and fixed in 70% ethanol overnight at 4°C. After washing with BSA-T-PBS (1% BSA/0.2% Triton X-100 in PBS), cells were incubated with phospho-H2AX antibody for 1 hour at room temperature. After washing with BSA-T-PBS, cells were incubated with the Alexa Fluor 547–conjugated secondary antibody and thereafter with propidium iodide (PI). After staining with PI, the stained cells (gated on the basis of forward and side scatter profiles) were analyzed on a BD FACSCalibur (BD Biosciences) and data were processed using the FlowJo program. Rabbit IgG was used as an isotype control to normalize for any nonspecific signal.
Statistical analyses were done using the two-sample t test. P < 0.05 was considered significant.
TWEAK-Fn14 signaling induces SGEF mRNA and protein expression via NF-κB
We previously reported that Fn14 signaling directs both proinvasive and prosurvival responses in GB tumors via Rac1 and NF-κB, respectively (3, 4, 12). We also described a role for the novel GEF, SGEF, in the promotion of Fn14-directed increased cell motility whereby Fn14 signaling enacted SGEF-required downstream RhoG and subsequently Rac1 activation (12). Of note, an analysis of 82 primary GB tumor specimens in the publicly available REMBRANDT dataset revealed a positive association between Fn14 and SGEF expression across the tissues (P < 0.001; Fig. 1A). We have previously shown that, similar to Fn14, SGEF expression was inversely correlated to patient survival among primary GB tumors and that SGEF protein expression is highly increased in GB clinical specimens (12). Thus, we sought to determine whether SGEF played an additional role in prosurvival signaling within GB cells. Given that there is a positive correlation between SGEF and Fn14 expression, we first analyzed whether Fn14 signaling played a role in the regulation of SGEF expression. SGEF expression is detected in T98G, A172, and U87 glioma cell lines, and minimally detected in U118 cells (Fig. 1B). Stimulation of glioma cells with the TWEAK ligand resulted in increased SGEF mRNA and protein levels with increased levels apparent within 2 hours of treatment, indicating that SGEF expression is inducible following TWEAK–Fn14 interaction. (Fig. 1C and D).
As NF-κB is an important promoter of cell survival in GB tumors (3, 4, 22), and Fn14 prosurvival signaling is dependent upon NF-κB upregulation of prosurvival gene transcripts (3), we next assessed whether the regulation of SGEF expression by TWEAK-Fn14 signaling required NF-κB. We analyzed the promoter sequence of SGEF and identified the presence of an NF-κB p65 consensus binding site at −2260 to −2238 base pairs upstream of the transcriptional start site including the 5′ UTR. Using an electrophoretic mobility shift assay with wild-type and mutant NF-κB p65 consensus sequence oligonucleotides from the SGEF promoter region, we assessed whether p65 NF-κB binds to the SGEF promoter following treatment with TWEAK. Electrophoretic mobility of SGEF wild-type but not mutant sequences shifted consequently to nuclear lysate binding; the addition of an anti-p65 antibody confirmed the role of p65 in the shift (Fig. 2 A). To further determine whether TWEAK-Fn14–driven increase in SGEF expression is dependent upon NF-κB, we either transiently transfected T98G glioma cells with plasmids expressing either control vector or IκBαM, an upstream super-repressor of NF-κB, or pharmacologically inhibited NF-κB activation via the cell permeable peptide inhibitor SN50 or control SN50M, and analyzed SGEF mRNA or protein levels following treatment with TWEAK. NF-κB inhibition either by IκBαM (Fig. 2B and C) or SN50 (Fig. 2D and E) resulted in diminished SGEF mRNA and protein expression, respectively, indicating that NF-κB is required for TWEAK-Fn14 induction of SGEF.
Depletion of SGEF impairs colony formation following temozolomide treatment and sensitizes cells to temozolomide-mediated apoptosis
As SGEF expression was upregulated by TWEAK-Fn14 in an NF-kB dependent fashion, we sought to assess whether SGEF was important in prosurvival signaling in GB. To determine the importance of SGEF protein in response to temozolomide, we utilized stable SGEF-depleted glioma cell lines, established via lentiviral-mediated transduction of either control (Ctrl) or shRNA-targeting SGEF (SGEF-12 & SGEF-13) expressing vectors in T98G (Fig. 3A), U87 and U118 (previously published ref. 12) glioma cells. Stable depletion of SGEF in T98G, U87, or U118 glioma cells did not alter proliferation or cell cycle (Supplementary Fig. S1). However, in glioma cells with stable depletion of SGEF, treatment with temozolomide for 48 hours followed by assessment for cellular apoptosis revealed that temozolomide-treated SGEF-depleted glioma cells showed increased chromatin condensation (Fig. 3B) as well as elevated cleaved PARP and cleaved caspase-3 by immunoblot analysis (Fig. 3C) in comparison with control temozolomide-treated cells. Validation of temozolomide-induced PARP cleavage in SGEF-depleted glioma cells was performed using a specific antibody to the “cleaved” form of PARP (Supplementary Fig. S2A). Therefore, knockdown of SGEF protein increases temozolomide-induced cytotoxicity. To further characterize the susceptibility to temozolomide of glioma cells with stable SGEF knockdown, we treated T98G, U87, and U118 glioma cells for 24 hours with temozolomide and measured colony growth formation. Cells depleted of SGEF by shRNA SGEF12 displayed significantly impaired colony formation after temozolomide treatment as compared with control temozolomide-treated cells (Fig. 3D). The depletion of SGEF in untreated glioma cells did not alter the number of colony formation as compared with control temozolomide-treated cells (Fig. 3D). Similar results were obtained for glioma cells depleted of SGEF expression by a second independent shRNA SGEF13 (Supplementary Fig. S2B). Together, these data indicate that SGEF protein function is important in the recovery response following temozolomide treatment.
Temozolomide treatment induces nuclear SGEF activity and SGEF-dependent BRCA1 activity and promotes SGEF in complex with BRCA1
Temozolomide treatment is known to result in the formation of double strand DNA breaks (DSB; ref. 8). The phosphorylation of histone (γH2A.X) is one of the earliest responses to DSB. γH2A.X is involved in the recruitment of and localization of DNA repair proteins and thus this phosphorylation is indicative of DNA damage DSB foci (23). SGEF contains two nuclear localization sequences (Fig. 4A; 24) and has previously been shown to be capable of nuclear localization, although the role of nuclear SGEF has not been described (25). We analyzed glioma cells over 24-hour treatment with temozolomide and assessed levels of H2A.X phosphorylation in control or SGEF-depleted lines (Fig. 4B). In both control and SGEF knockdown cells, increased phosphorylation of H2A.X is detected upon temozolomide treatment within 8 and 24 hours, indicating that SGEF does not play a role in preventing the formation of γH2A.X foci. However, SGEF knockdown cells do not show loss of γH2A.X foci beyond 24 hours, as measured by immunofluorescent staining of γH2A.X, indicating altered kinetics of DNA damage response subsequent to temozolomide-induced DSB (Supplementary Fig. S3).
We next examined whether SGEF plays a role in the coordinated response to DNA damage. We first assessed whether the activity of SGEF is altered following temozolomide treatment. U87 glioma cells treated for 24 hours with temozolomide were fractionated for nuclear lysates, in which SGEF activity was determined using RhoG G15A nucleotide–free mutant constructs (16). Treatment with temozolomide resulted in increased SGEF activity in the nucleus (Fig. 4C), further supporting a role for SGEF in the response to temozolomide treatment. Use of functional site prediction analysis suggests that SGEF contains two phosphopeptide domain motifs at amino acids 493–497 (ASKKF) and 741–745 (ASHLF; Fig. 4A), which can directly interact with the BRCT (carboxy-terminal) domain of BRCA1. BRCT domains are present in several DNA damage response proteins (26), and the phosphorylation of BRCA1 occurs in response to DNA-damaging agents (9). We therefore sought to determine whether SGEF is important in BRCA1 activation following temozolomide treatment. Glioma cells were treated for 24 hours with temozolomide and the phosphorylation of BRCA1 was assessed between control and SGEF-depleted lines (Fig. 4B). The depletion of SGEF prevented temozolomide-induced BRCA1 phosphorylation. We then assessed whether treatment with temozolomide induces complex formation between SGEF and BRCA1. U87 glioma cells treated for 24 hours with temozolomide were analyzed via immunoprecipitation of BRCA1 and immunoblot analysis of SGEF. Minimal SGEF and BRCA1 coimmunoprecipitated from nuclear lysates of untreated cells; however, temozolomide treatment induced complex formation between SGEF and BRCA1 within 30 minutes, and maximal interaction was observed at 8 hours (Fig. 4D), along with increased binding to phosphorylated H2A.X. Thus, SGEF may function in part to promote the BRCA1 response to DNA damage.
Depletion of BRCA1 impairs colony formation following temozolomide treatment, which is not enhanced by concomitant SGEF depletion
Depletion of SGEF impairs cell survival following temozolomide treatment (Fig. 3D). To assess whether SGEF may promote a divergent response to temozolomide treatment in addition to the promotion of a BRCA1-mediated response, we depleted BRCA1 via transient siRNA transfection (Fig. 5A) in control glioma cells or glioma cells stably depleted of SGEF and treated the cells with temozolomide. Depletion of BRCA1 impaired survival to temozolomide treatment similar to that observed after SGEF depletion alone. The combination of BRCA1 depletion in SGEF-depleted cells did not result in any further significant impairment of cell survival following temozolomide treatment (Fig. 5B). Given that there is no additive or synergistic effect of the dual knockdown of SGEF and BRCA1, the data support that SGEF works along the same pathway in concert with BRCA1 in the DNA damage repair response.
In this study, we report a relationship between SGEF and both therapeutic resistance and cell survival. SGEF protein expression is induced in glioma cells by TWEAK-Fn14 signaling and dependent upon NF-κB. NF-κB signaling has been well characterized to promote both GB cell invasion and survival (3, 4, 22, 27). Moreover, elevated or constitutive NF-κB activity has been demonstrated in gliomas and correlates with increasing tumor grade (27, 28). NF-κB signaling has specifically been shown to protect cells against the standard-of-care treatments in GB. The inhibition of IκBα phosphorylation prevents NF-κB activity and sensitizes glioma cells to radiation treatment (29) and NF-κB is an important player in promoting resistance to O6-alkylation (30), a temozolomide-induced DNA damage modification.
TWEAK-Fn14 signaling is one notable pathway in glioma that utilizes Rac1-dependent NF-κB activation to promote cell invasion and cytotoxic therapy resistance with enhanced cell survival (3, 4, 31). To date, SGEF has been largely associated with a role in promoting cell motility; SGEF has been shown to promote the invasive capacity of HPV-transformed tumor cells and actin cytoskeleton remodeling after salmonella infection (25, 32). In GB tumors, in addition to a role in the promotion of cell invasion, we have previously reported that SGEF is significantly overexpressed and is correlated with poor patient outcome (12). Here, we report the role of SGEF in promoting cell survival. We show that in glioblastoma, TWEAK binding to Fn14 fosters occupancy of the SGEF promoter by NF-κB, and that TWEAK-Fn14 upregulation of SGEF mRNA and protein expression is dependent upon NF-κB function. The link between Fn14 and SGEF in GB is further supported on the basis of mRNA expression analysis indicating a strong positive correlation in expression of the two genes among a panel of primary tumor specimens in the publicly available REMBRANDT dataset of 82 GB tumors (Fig. 1A). Of note, this correlation was not significant when brain tumors of all grades were considered (data not shown) but was highly statistically significant within GB tumors alone, indicating the relationship between SGEF and Fn14 may be specific to malignant progression.
Our data suggest an important role for SGEF in the response of glioma cells to temozolomide treatment. We showed that the shRNA-mediated depletion of SGEF does not affect cell proliferation or cell cycle (Supplementary Fig. S1), but does impair the ability of glioma cells to form colonies following temozolomide treatment and leads to glioma cell sensitization to temozolomide-induced cell death via apoptosis. SGEF is known to contain two nuclear localization sequences (24), and has been reported to localize in the nucleus of cells (25). Here, we show that temozolomide treatment of glioma cells induces nuclear activity of SGEF in a time-dependent fashion. Moreover, SGEF is known to facilitate guanine nucleotide exchange for the GTPase RhoG (33) and RhoG has been shown to contain a nuclear localization sequence (34), the significance of which remains unknown to date. Our data indicate that RhoG becomes active in the nucleus, similar to SGEF, in response to temozolomide treatment of glioma cells (Supplementary Fig. S4). Therefore, nuclear RhoG may play a role in the SGEF prosurvival response to temozolomide treatment.
Sequence analysis of SGEF revealed two BRCT-binding domains with the potential for binding BRCA1. Here, we demonstrate that SGEF is found in complex with BRCA1 following temozolomide treatment, and that phosphorylation of BRCA1 is dependent upon SGEF. Thus, the decreased capacity of temozolomide-induced BRCA1 phosphorylation in SGEF-depleted glioma cells may help explain the observed significantly impaired capacity of glioma cells to recover colony formation and the increased apoptosis in temozolomide-treated SGEF-depleted glioma cells. While our data showed that SGEF knockdown does not prevent γH2A.X foci formation as marked by H2A.X phosphorylation following temozolomide addition within 24 hours, longer assessment of the kinetics of DSB repair mechanism by immunofluorescent staining of γH2A.X showed defect in DSB repair in SGEF knockout cells beyond 36 to 72 hours (Supplementary Fig. S3). Interestingly, our data indicate that SGEF is required for BRCA1 recruitment to H2A.X foci following temozolomide treatment (Supplementary Figure 5). The determination of the specific functional site responsible for SGEF interaction with BRCA1, along with the mechanism of SGEF-BRCA1–mediated glioma cell survival will be the focus of future studies. We thus propose a scheme of TWEAK-Fn14–inducible SGEF mRNA and protein expression dependent upon NF-κB nuclear translocation and activity, whereby increased SGEF levels promote a prosurvival phenotype in the face of temozolomide treatment by promotion of BRCA1-γH2A.X DNA damage response activity (Fig. 6).
BRCA1 has been shown to transiently interact at sites of damage or stalled replication forks with the role of homologous recombination. BRCA1 also functions in NHEJ, and S- and G2–M-phase checkpoints; however, some reports suggest that BRCA1 preferentially promotes the error-free HR pathway for DNA repair over NHEJ to preserve chromosome stability (9). Cancers known to have a deficiency of the HR repair proteins BRCA1 and BRCA2 display particular sensitivity to PARP inhibition, a protein whose activity normally facilitates single-strand damage repair. When unrepaired, these single strands are converted to DSBs during replication, which are then unable to be corrected due to a nonfunctioning HR system (35). Interestingly in gliomas, the inhibition of HR via siRNA-mediated depletion of Rad51 or BRCA2 greatly sensitized glioma cells to temozolomide, the effect of which was enhanced by concurrent PARP inhibition (36). These studies further support the notion that targeting the modulation of BRCA1 activity as regulated by SGEF expression may enhance cell killing in GB tumors, and future studies will address the potential for synergistic lethality in targeting this axis in combination with other inhibitors of DNA repair.
Tumor modulation of DNA repair pathways has been described as one main avenue of glioblastoma resistance to temozolomide. Gene expression associated with promoter hypomethylation of the DNA repair protein O6-methylguanine-DNA methyltransferase (MGMT) allows for notable inherent tumor resistance in glioblastomas. However, epigenetic silencing of MGMT by promoter hypermethylation has been shown to occur in 30% to 60% of glioblastoma tumors, and it has been suggested that MGMT-deficient GB cells may be particularly susceptible to targeting HR in combination with additional DNA repair proteins including PARP inhibition (35). In addition, acquired temozolomide resistance has been demonstrated by silencing of DNA mismatch repair genes through treatment-induced mutations which then allows the tumor cell to escape a futile repair process otherwise leading to cell death, while the epigenetic silencing of DNA base excision repair genes may predict temozolomide sensitivity (35). Given the role of DNA repair pathways in the tumor cell response to temozolomide treatment, we further characterized SGEF expression in the setting of temozolomide chemoresistance using a panel of primary GB xenografts treated in vivo with temozolomide to derive matched parental and TMZ-R pairs (37). Interestingly, protein expression of SGEF was found to be higher in the resistant lines versus the parental line in a subset (8/17) of samples (Supplementary Fig. S6), suggesting in some GB tumors increased SGEF expression may result following exposure to temozolomide. It is unknown whether the high expression of SGEF in these primary tumors directly confers resistance to temozolomide; however, the study of these lines is currently under investigation. Moreover, it is unknown why SGEF protein expression is elevated in only a subset of the TMZ-R GB xenografts. It is possible that genetic heterogeneity of the patient-derived parental and temozolomide-resistant lines may play a role in perceived SGEF expression due to sampling bias, and this concern will also be addressed in future studies.
Interestingly, radial migration analysis of GBM14 and GBM14 TMZ-R primary xenografts revealed an elevated rate of cell migration in the GBM14 TMZ-R cells (data not shown). SGEF promotes cell migration and invasion in glioblastoma via activation of the Rho GTPase RhoG with subsequent RhoG-dependent activation of Rac1 and the formation of lamellipodia (12). Thus, NF-κB–mediated increased SGEF expression may be one mechanism that facilitates the increased cell motility of TMZ-R glioma cells. Indeed, increased invasive capacity has been previously reported in glioma as a response to cytotoxic therapy. For example, it has been shown that radiation of glioblastoma leads to the enhanced cell invasive potential via activation of the Rho–PI3K signaling pathway (38). In addition, the proinvasive integrins, αvβ3 and αvβ5, have been demonstrated to mediate a prosurvival response in glioma to radiation through integrin-linked kinase and the RhoB GTPase (39), thus implying overlapping roles for mediators of cell motility with the promotion of cell survival. Of note, there have also been multiple reports of increased cell invasiveness resulting from treatment with chemotherapeutic agents among several tumor types (40-42). Thus, the roles of invasion and survival are interconnected in the promotion of disease progression, and there is mounting evidence for overlap between these two processes (43). SGEF therefore presents a novel hub in the interrelated axes of tumor cell invasion and survival.
Despite advances in medical technology and treatment, GB prognosis has remained largely unchanged over the last several decades (44, 45). The ability of glioma cells to survive undeterred from current treatment strategies implies that new therapeutic avenues are necessary for treatment of this disease. There is accumulating evidence that combinatorial therapy that includes use of treatment modalities designed to hamper the DNA repair mechanisms of the cell may provide a significant added survival benefit to patients over the standard-of-care alone or when used in combination with inhibitors of other GB-deregulated pathways (46-49). Moreover, therapy aimed at mediators of invasion can also lead to increased chemotherapeutic sensitivity (5). Thus, pathways deregulated in GB that promote both temozolomide resistance and cell motility represent novel therapeutic targets in future drug design. Our data support a role for SGEF in both the promotion of cell invasion and cell survival signaling within GB tumors and provide a rationale for targeting this signaling axis. Interestingly, there has been a recent report of the RhoJ GTPase in promoting melanoma chemoresistance by suppressing DNA damage sensing pathways including the uncoupling of ATR from its downstream effectors with resulting decreased DNA damage–induced apoptosis (50). Thus, the role of GEFs and GTPases in chemoresistance via modulation of DNA repair mechanisms is an emerging field in which we have validated a role for SGEF in GB.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: S.P.F. Ensign, M.H. Symons, M.E. Berens, N.L. Tran
Development of methodology: S.P.F. Ensign, A. Roos, J.N. Sarkaria, J.C. Loftus, N.L. Tran
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.P.F. Ensign, A. Roos, I.T. Mathews, S. Tuncali, N.L. Tran
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.P.F. Ensign, A. Roos, I.T. Mathews, H. Dhruv, N.L. Tran
Writing, review, and/or revision of the manuscript: S.P.F. Ensign, I.T. Mathews, H. Dhruv, J.N. Sarkaria, M.H. Symons, J.C. Loftus, M.E. Berens, N.L. Tran
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.P.F. Ensign, H. Dhruv, S. Tuncali, N.L. Tran
Study supervision: S.P.F. Ensign, A. Roos, N.L. Tran
The authors thank Dr. Keith Burridge (University of North Carolina, Chapel Hill, NC) for the gift of the pGEX4T-1-RhoG(15A) plasmid and Danielle Quirino for experimental support.
This work is supported by NIH grants R01 CA130940 (to N.L. Tran), the ARCS Foundation Eller Scholarship and Science Foundation Arizona Fellowship (to S.P.F Ensign), and The Ben and Catherine Ivy Foundation (to M.E. Berens).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.