© The Author 2007. Published by Oxford University Press.
ARTICLES |
The 1p-Encoded Protein Stathmin and Resistance of Malignant Gliomas to Nitrosoureas
Affiliations of authors: Surgical and Molecular Neuro-oncology Unit (TTBN, TP, XJL, OA, DWK, JKP), Flow Cytometry Facility (DM), and Office of the Clinical Director (NOJ), National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD; Neuro-oncology Branch, National Cancer Institute, Bethesda, MD (SP, WZ, YK, JCZ, HAF); Dana-Farber/Brigham and Women's Cancer Center, Boston, MA (PYW); Brigham and Women's Hospital, Boston, MA (UDG, PMB); Proteomics Core Facility, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD (WWW, RFS)
Correspondence to: John K. Park, MD, PhD, 35 Convent Dr, MSC 3706, Bethesda, MD 20892 (e-mail: parkjk{at}ninds.nih.gov).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsMaterials and MethodsResultsDiscussionReferencesNotes |
|---|
Background: Malignant gliomas are generally resistant to all conventional therapies. Notable exceptions are anaplastic oligodendrogliomas with loss of heterozygosity on chromosome 1p (1p+/–). Patients with 1p+/– anaplastic oligodendroglioma frequently respond to procarbazine, 1-(2-chloroethyl)-3-cyclohexyl-l-nitrosourea, and vincristine. Because the underlying biologic basis for this clinical finding is unclear, we evaluated differentially expressed 1p-encoded proteins in 1p+/– and 1p+/+ malignant glioma cell lines and then examined whether their expression was associated with outcome of patients with anaplastic oligodendroglioma.
Methods: We used a comparative proteomic screen of A172 (1p+/–) and U251 (1p+/+) malignant glioma cell lines to identify differentially expressed 1p-encoded proteins, including stathmin, a microtubule-associated protein. 1p+/– and 1p+/+ anaplastic oligodendroglioma specimens from 24 patients were assessed for stathmin expression by immunohistochemistry. The relationship between stathmin expression and clinical outcome was assessed with Kaplan–Meier analyses. RNA inhibition and cDNA transfection experiments tested effects of stathmin under- and overexpression, respectively, on the in vitro and in vivo resistance of malignant glioma cells to treatment with nitrosourea. For in vivo resistance studies, 36 mice with intracranial and 16 mice with subcutaneous xenograft tumor implants were used (one tumor per mouse). Flow cytometry was used for cell cycle analysis. Immunoblotting was used to assess protein expression. All statistical tests were two-sided.
Results: Decreased stathmin expression in tumors was statistically significantly associated with loss of heterozygosity in 1p (P<.001) and increased recurrence-free survival (P<.001). The median recurrence-free survival times for patients with tumors expressing low, intermediate, or high stathmin levels were 45 months (95% confidence interval [CI] = 0 to 90 months), 17 months (95% CI = 10.6 to 23.4 months), and 6 months (95% CI = 1.7 to 10.3 months), respectively. Expression of stathmin was inversely associated with overall survival of nitrosourea-treated mice carrying xenograft tumors. Median survival of mice with stathmin+/– tumors was 95 days (95% CI = 68.7 to 121.3 days) and that of mice with stathmin+/+ tumors was 64 days (95% CI = 58.2 to 69.8 days) (difference = 31 days, 95% CI = 4.1 to 57.9 days; P<.001, log-rank test). Nitrosoureas induced mitotic arrest in malignant glioma cells, and this effect was greater in cells with decreased stathmin expression.
Conclusions: Loss of heterozygosity for the stathmin gene may be associated with improved outcomes of patients with 1p+/– anaplastic oligodendroglioma tumors.
Prior knowledge Patients with anaplastic oligodendroglioma with loss of heterozygosity on chromosome 1p (1p+/–) frequently respond to combination chemotherapy with procarbazine, 1-(2-chloroethyl)-3-cyclohexyl-l-nitrosourea, and vincristine. Study design Molecular study in malignant glioma cell lines, xenograft tumor mouse model systems, and human tumor specimens from patients with survival information available. Contribution Decreased expression of stathmin, a microtubule-associated protein whose gene is located on chromosome 1p, was associated with loss of heterozygosity in 1p and with increased recurrence-free survival. Implications Loss of heterozygosity of stathmin may be associated with outcome of patients with 1p+/– anaplastic oligodendroglioma. Limitations Although stathmin haploinsufficiency appears to contribute to the nitrosourea sensitivity of 1p+/– gliomas, other factors may also be involved. The immunohistochemical analysis was only semiquantitative, and tumor samples could be reproducibly classified into three expression categories, not two as would be predicted from gene copy number.
|
Malignant gliomas are composed of neoplastic cells that histologically resemble astrocytes, oligodendrocytes, and ependymal cells. The glioma subtypes of these normal glial elements are astrocytomas, oligodendrogliomas, and ependymomas, respectively. As a result of their rapid growth and infiltrative spread within the brain, malignant gliomas are generally resistant to all conventional therapies, including surgery, radiation, and chemotherapy, and patients have median survivals of 1–3 years (1,2).
In contrast to most malignant gliomas, a subset of tumors with loss of heterozygosity for chromosome 1p can be exquisitely sensitive to chemotherapy with procarbazine, 1-(2-chloroethyl)-3-cyclohexyl-l-nitrosourea (CCNU), and vincristine. In the initial report of this response (3), the relative risk of recurrence or death after chemotherapy for patients with anaplastic oligodendroglioma tumors containing two intact 1p alleles (1p+/+) was 4.3 times higher than that of patients with tumors containing only one intact 1p allele (1p+/–) (P = .001). Other studies (4–10) have confirmed these results, particularly the association between loss of heterozygosity at 1p and increased time to tumor recurrence (i.e., recurrence-free survival). An association between loss of heterozygosity at 1p and chemosensitivity has also been reported for the astrocytoma and mixed oligoastrocytoma subtypes of gliomas (9–13). To date, no explanation for these clinical findings has been reported.
We hypothesized that the haploinsufficiency of a 1p-encoded chemoresistance protein contributes to the chemosensitivity of patients with 1p+/– malignant gliomas to treatment with procarbazine, CCNU, and vincristine. Predictions of this hypothesis include the decreased expression of such a protein in 1p+/– tumors relative to that in 1p+/+ tumors and an association between decreased expression levels of this protein and increased recurrence-free survival. We tested this hypothesis by identifying candidate proteins that were encoded on 1p, including the microtubule-associated protein stathmin, and experimentally decreasing or increasing their expression to determine whether specific and appropriate changes in resistance to procarbazine, CCNU, and/or vincristine could be detected and whether a mechanism for the interaction between a protein and the appropriate drug(s) could be identified.
|
Materials and Methods |
|---|
|
TopAbstractContext and CaveatsMaterials and MethodsResultsDiscussionReferencesNotes |
|---|
Cell Culture
A172, Hs683, T98G, U87, and U251 human glioma cell lines were obtained from American Type Culture Collection (Manassas, VA) and maintained in Dulbecco's modified Eagle medium (Mediatech; Herndon, VA) supplemented with 10% fetal bovine serum (Mediatech), 100 IU penicillin, and streptomycin (Mediatech; 100 µg/mL). All cell lines were incubated at 37 °C in an atmosphere of 5% CO2 and 95% air.
Cytogenetic Analysis
DNA samples were prepared from the A172, Hs683, T98G, U87, and U251 human glioma cell lines and applied to Genechip Human Mapping 10K arrays (Affymetrix; Santa Clara, CA) according to the manufacturer's recommendations. Chromosome copy numbers were estimated with Affymetrix Chromosome Copy Number Tool version 1.1, which compares intensity values with build-in distributions that are specific for a single-nucleotide polymorphism or genotype and that are derived from a reference set of 145 normal individuals. These comparisons result in single-point P values for each single-nucleotide polymorphism that represent the probability that the copy number is equal to 2. To increase the confidence of detected alterations and to reduce noise, contiguous points analysis (14) was also performed. The contiguous points analysis is based on the assumption that the greater the number of contiguous single-nucleotide polymorphisms that indicate similar gains or losses of particular genes, the greater the likelihood that the chromosomal region containing those individual genes is gained or lost, respectively. A contiguous points analysis P value, which is calculated by the Affymetrix Chromosome Copy Number Tool and outputted as a "meta P value," is a surrogate probability for a given stretch of genome to have normal DNA copy number. The algorithm implemented by use of the Affymetrix Chromosome Copy Number Tool is described in detail by Huang et al. (14).
Two-Dimensional Polyacrylamide Gel Electrophoresis and Protein Identification
For two-dimensional polyacrylamide gel–based analysis, A172 and U251 cells were harvested with hypotonic lysis buffer (0.5% Nonidet P-40, 10 mM Tris, 5 mM MgCl2, 10 mM KCl, protease inhibitor mixture [product P8340; 1:500 dilution], and phosphatase inhibitors [2 mM sodium orthovanadate and 2 mM sodium fluoride]; all from Sigma, St Louis, MO), and the mixture was centrifuged at 1000g for 5 minutes at 4 °C to yield nuclear and nonnuclear protein fractions. The A172 and U251 nonnuclear protein fractions were precipitated with a combination of 44% methanol and 11% chloroform; resuspended in 7 M urea, 2 M thiourea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, and 10 mM Tris (pH 8.8) (all from Sigma); and labeled with Cy3- and Cy5-CyDye DIGE Fluor minimal dyes (GE Healthcare; Piscataway, NJ), respectively, for 30 minutes at 4 °C. A pooled sample of A172 and U251 nonnuclear protein fractions was similarly labeled with Cy2-CyDye DIGE Fluor minimal dye (GE Healthcare) for use as an internal control. Cy3-, Cy5-, and Cy2-labeled samples (50 µg each) were mixed together with 1% Bio-Lyte 3/10 Ampholyte (Bio-Rad; Hercules, CA), 40 mM dithiothreitol (Sigma), and 0.002% bromphenol blue (Sigma) and isoelectrically focused on an immobilized pH-gradient strip (24 cm; pH 3–10) (GE Healthcare) with an IPGphor apparatus (GE Healthcare). After focusing, strips were equilibrated for 15 minutes in a solution containing 6 M urea, 30% glycerol (Sigma), 2% sodium dodecyl sulfate (SDS; Sigma), 50 mM Tris–HCl (pH 8.8), 0.002% bromphenol blue, and dithiothreitol (10 mg/mL) followed by a second 15-minute equilibration with iodoacetamide (25 mg/mL; Sigma) instead of dithiothreitol. Strips were briefly rinsed in 1x SDS–polyacrylamide gel electrophoresis (PAGE) buffer and applied to a 12.5% polyacrylamide gel (GE Healthcare) for electrophoresis at a constant 1–2 W per gel overnight. Samples were prepared and then electrophoresed in triplicate. A Typhoon scanner in fluorescence mode (GE Healthcare) was used to obtain Cy3, Cy5, and Cy2 images of the gels at a resolution of 100 µm. Gels were then fixed in 30% ethanol and 7.5% acetic acid (both from Sigma) for 2 hours, followed by Sypro Ruby (Invitrogen; Carlsbad, CA) staining overnight for total protein visualization. Protein spots were analyzed with DeCyder software (GE Healthcare), and gel plugs containing proteins with Cy3:Cy5 labeling ratios of less than 0.67 were identified and processed with the automated Spot Handling Workstation (GE Healthcare). After trypsin (GE Healthcare) digestion of the gel plugs, peptides were extracted and analyzed with an ABI 4700 matrix-assisted laser desorption ionization–tandem time-of-flight mass spectrometer (MALDI-TOF/TOF MS; Applied Biosystems; Foster City, CA), operating in reflectron positive-ion mode. We searched in the Swiss-Prot human database with GPS Explorer software (Applied Biosystems) to identify the peptide mass data that we acquired by use of MALDI-TOF/TOF MS. A group of peptides were considered to be a positive match to a database protein if the confidence interval (CI) for the match, as determined by GPS Explorer, was at least 95% and if the position of the spot on the two-dimensional gel reflected approximately the theoretical isoelectric point and molecular weight of the specified protein (see Table 1).
|
Immunoblot Analysis
Whole-cell extracts of A172 cells, U251 cells, and cell lines derived from either A172 or U251 cells were subject to one- or two-dimensional PAGE, and protein bands or spots, respectively, were transferred to polyvinyl difluoride membranes (Millipore; Billerica, MA). Rabbit anti-stathmin (Calbiochem; San Diego, CA; 1:10000 dilution), mouse anti–
-actin (Sigma; 1:40000 dilution), mouse anti-securin (Abcam; Cambridge, MA; 1:1000 dilution), mouse anti–O6-methylguanine-DNA methyltransferase (MGMT) (Lab Vision; Fremont, CA; 1:500 dilution), mouse anti–
-enolase (Novus Biologicals; Littleton, CO; 1:500 dilution), and rabbit anti–DJ-1 (Chemicon; Temecula, CA; 1:5000 dilution) antibodies were used at the dilutions indicated. Anti-rabbit and/or anti-mouse alkaline phosphatase–conjugated antibodies (Promega; Madison, WI; 1:7500 dilution) were used as secondary antibodies. Blots were developed with 9-H-(1,3-dichloro-9,9-dimethylacridin-2-one-7-yl) phosphate (Invitrogen) and scanned with an FLA-5100 laser–based scanner (Fujifilm; Stamford, CT). Protein bands and spots were quantitated with MultiGauge version 2.3 imaging software (Fujifilm). For coimmunoprecipitation experiments, cell extracts were incubated with mouse anti-p55cdc antibody (Santa Cruz Biotechnology; Santa Cruz, CA; 1 µg/400 mg of extract) for 2 hours at 4 °C, and then protein G-agarose beads (Santa Cruz Biotechnology; 15 µL) were added for an additional hour. Beads were washed, boiled in XT sample buffer (Bio-Rad) for 5 minutes, and centrifuged at 16000g for 5 minutes at 25 °C. Supernatants were subjected to one-dimensional SDS–PAGE, and protein bands were transferred to polyvinyl difluoride membranes and probed with rabbit anti-p55cdc (Santa Cruz Biotechnology; 1:500 dilution), anti-MAD2 (Novus Biologicals; 1:3000 dilution), or anti-cdc27 (Santa Cruz Biotechnology; 1:300 dilution) antibodies. The subsequent secondary antibody and signal visualization procedures used were as described above.
For two-dimensional immunoblot analysis of U251 cells treated with CCNU (HNZ Portlink; ON, Canada), cells were incubated with 10 µM CCNU or 0.005% dimethyl sulfoxide (DMSO; Sigma) vehicle control for 4 hours at 37 °C in an atmosphere of 5% CO2/95% air. Cell extracts were prepared as described above for separation of proteins on two-dimensional polyacrylamide gels. Rather than differentially labeling the proteins and analyzing both lysates on a common gel, each lysate was analyzed on its own gel and the protein spots were transferred to separate polyvinyl difluoride membranes. The membranes were probed with rabbit anti-stathmin antibodies (Calbiochem; 1:10000 dilution). Further procedures for visualization of stathmin proteins were identical to those described above for one-dimensional immunoblot analysis. Immunoblot images were pseudocolored (DMSO-treated extracts = green; CCNU-treated extracts = red) and combined by use of Photoshop version 7.0 (Adobe; San Jose, CA).
Fluorescent In Situ Hybridization and Immunofluorescence Histochemistry
Brain tumor specimens were obtained by neurosurgical resection in accordance with Brigham and Women's Hospital (Boston, MA) and National Institutes of Health (NIH, Bethesda, MD) institutional review board–approved human tissue collection protocols. Written informed consent for research usage of tumor tissues was obtained from each patient. Specimens were fixed in paraformaldehyde, embedded in paraffin, sectioned with a microtome, mounted on glass slides (Fisher Scientific; Waltham, MA), and stained with hematoxylin and eosin (Fisher Scientific) for histologic analysis. Those specimens satisfying the diagnostic criteria for an anaplastic oligodendroglioma of World Health Organization grade III (2) were selected for use in the study. Fluorescent in situ hybridization was performed on 5-µm paraffin-embedded sections with the LSI-1p36/LSI-1q25 Dual-Color Probe Set (Vysis; Downers Grove, IL) according to the manufacturer's recommendations. The LSI 1p36 probe is approximately 400 kilobases long and hybridizes to a region that extends from a point near the SHGC57243 locus, through the TP73 and EGFL3 genes, to a point telomeric to the EGFL3 locus. The LSI 1q25 probe is approximately 620 kilobases long and hybridizes from a point telomeric to the WI-6848 locus, through the ABL2 and ANGPTL1 genes, to a point near the SHGC-1322 locus.
Immunofluorescence histochemistry was performed on 5-µm paraffin-embedded sections with rabbit anti-olig2 polyclonal antibodies (generous gift from J. Alberta and C. Stiles; 1:10000 dilution) to identify tumor cells (15) and either goat anti-stathmin polyclonal antibodies (Santa Cruz Biotechnology; 1:100 dilution) or mouse anti-MGMT monoclonal antibodies (1:50 dilution), as previously described (15,16). The corresponding secondary antibodies used were Alexa Fluor 594 (red fluorophore)–conjugated chicken anti-rabbit IgG and Alexa Fluor 488 (green fluorophore)–conjugated chicken anti-goat or anti-mouse IgG antibodies (all Invitrogen; 1:500 dilution), respectively. Anti-stathmin antibody–stained sections were also counterstained with 300 nM 4',6-diamidino-2-phenylindole (Sigma) to label cell nuclei. Images were acquired with an LSM 510 confocal laser scanning microscope (Carl Zeiss MicroImaging; Thornwood, NY). The overall expression level of stathmin in each tumor was determined by first identifying the cells within each field with maximal stathmin staining intensity, which were commonly neurons. All olig2-positive cells within each field were then scored on a scale of I–III for stathmin staining intensity in which 0%–33% is level I, 34%–66% is level II, and 67%–100% is level III (the maximal intensity is 100%). The median stathmin staining intensity score (i.e., level I, II, or III) of the tumor cells in three separate fields of each specimen was determined, and that score was assigned to the specimen as a whole. A consensus was reached by three independent graders who were blinded to the 1p status of the tumor and clinical outcomes of the patients treated with procarbazine, CCNU, and vincristine. A three-tier scoring system was used because the three independent graders could not reach a consensus on all specimens with a two-tier scoring system (high versus low stathmin expression). For determination of tumor MGMT status, specimens in which more than 20% of the olig2-positive cells also expressed MGMT were considered to be positive for MGMT. This cutoff point was set a priori as described previously by McLendon et al. (16). The MGMT status of the tumors was determined by individuals who were blinded to the 1p status of the tumors and clinical outcomes of the patients.
Transfections of U251 and A172 Cells
To generate U251 cells that could be induced to express decreased levels of stathmin (termed U251-STMNi cells) and U251 cells that could be used in experiments as transfection and induction controls for the U251-STMNi cells (termed U251-LacZi cells), U251 cells were first transfected with pERV3 (Stratagene; La Jolla, CA) by use of Lipofectamine 2000 (Invitrogen), according to manufacturer's recommendations. The pERV3 vector contains the neomycin resistance gene, and stably transfected cells were selected on the basis of their resistance to G418 (Invitrogen), a neomycin analog. G418-resistant cells were then transfected with pEind-RNAi (17), which contains the hygromycin resistance gene as well as either the stathmin-specific short-hairpin RNA (shRNA) oligonucleotide sequence 5'-AGTTGTTGTTCTCTTCTATTGCCTTCTGATTGGTCAGAAGGCAATAGAAGAGAACAACAACT-3' or the LacZ-specific shRNA oligonucleotide sequence 5'-CTACACAAATCAGCGATTTCGAAAAATCGCTGATTTGTGTAG-3' downstream of a modified U6 promoter sequence. Cells resistant to both G418 and hygromycin B (Invitrogen) were subcloned by limiting dilution and assessed for stathmin expression before and after incubation with muristerone A (Invitrogen), which enhances the activity of the modified U6 promoter and leads to the transcription of a downstream shRNA oligonucleotide sequence, by immunoblotting with anti-stathmin antibodies as described above. To generate A172 cells that overexpress stathmin (STMN), -enolase (ENOA), or DJ-1 (PARK7) (termed A172-STMN, A172-ENOA, or A172-PARK7 cells, respectively), A172 cells were transfected with plasmid vectors containing the cytomegalovirus promoter and the corresponding cDNAs (Origene Technologies; Rockville, MD) by use of FuGENE 6 (Roche Diagnostics; Indianapolis, IN), according to manufacturer's recommendations. Transgene expression was confirmed with immunoblotting as described above.
Cell Cytotoxicity Assays
We cultured 2 x 102 U251-STMNi or U251-LacZi cells in 96-well plates (Corning; Acton, MA), and cells were incubated with 10 µM muristerone A, which induced transcription of the intended shRNA, or with the ethanol control the next day. After 2 days of incubation, cells were treated with 0.005% DMSO vehicle control, procarbazine (Sigma Tau Pharmaceuticals; Gaithersburg, MD), vincristine (Sigma), temozolomide (Drug Synthesis and Chemistry Branch, National Cancer Institute [NCI]; Frederick, MD), or CCNU for 4 days, as indicated. Cells were counted by use of Alamar Blue (Biosource; Camarillo, CA) according to manufacturer's recommendations. Briefly, during the last 3 hours of the assay, the existing medium was replaced with fresh medium containing 10% Alamar Blue. The Alamar Blue is converted from a blue to red color by viable cells, and the degree of color conversion, as monitored with a SpectraMax Plus 384 microplate spectrophotometer (Molecular Devices; Sunnyvale, CA), is indicative of the number of viable cells. Six replicate wells were used for each condition, and experiments were carried out in triplicate. Proliferation assays for A172-derived cells were similar to the assays for U251-derived cells, except that cells were cultured at a density of 5 x 102 cells per well and treated with chemotherapy drugs 1 day after plating.
Cell Cycle Analysis
U251-STMNi cells were incubated with muristerone A, which induces transcription of a stathmin-specific shRNA, or ethanol control for 3 days and then treated with 10 µM CCNU or 0.005% DMSO vehicle control. On the second day after CCNU or DMSO treatment, cells were pulsed labeled with 10 µM bromodeoxyuracil (Sigma) for 2 hours and harvested. Cells were then stained with fluorescein isothiocyanate–coupled anti-bromodeoxyuracil antibody (1:10 dilution) and propidium iodide (5 µg/mL) to determine the proportion of cells in the G2 and M phases of the cell cycle (both from BD Biosciences; San Jose, CA) or with 300 nM 4',6-diamidino-2-phenylindole, anti–phospho-histone H3 antibody (at Ser-10) (Chemicon; 1:200 dilution) ,and Alexa Fluor 488 goat anti-rabbit IgG (Invitrogen; 1:2000 dilution) to determine the proportion of cells in the M phase of the cell cycle. Stained cells were analyzed with a Vantage SE flow cytometer and CellQuest software (BD Biosciences). The percentage of cells in G2 phase of the cell cycle was determined by subtracting the percentage of cells in M phase from that in G2 and M phases. All cell cycle–related experiments were performed in triplicate.
Murine Tumor Models
For intracranial tumors, severe combined immunodeficient mice (NCI, NIH, Bethesda, MD) were anesthetized with a combination of ketamine (150 mg/kg) and xylazine (10 mg/kg) and secured in a stereotactic apparatus (Stoelting Co; Wood Dale, IL). Each mouse was then injected in its right striatum with 1 x 106 U251-STMNi cells (n = 18 mice) or 1 x 106 U251-LacZi control cells (n = 18 mice) that had been treated previously with 10 µM muristerone A for 2 days. Three days later, these mice were injected intraperitoneally with either 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) (Bristol-Myers Squibb; New York, NY) in 10% ethanol (Sigma) at a dose of 15 mg/kg (n = 9 for mice with U251-STMNi cell tumors and n = 9 for mice with U251-LacZi control cell tumors) or vehicle (10% ethanol) (n = 9 for mice with U251-STMNi cell tumors and n = 9 for mice with U251-LacZi control cell tumors). The total number of mice in this study was 36; each mouse had only one tumor. BCNU, rather than CCNU, was used because it is more soluble than CCNU and thus less likely to precipitate in the ethanol vehicle. Mice were observed daily for death or signs of distress.
For subcutaneous tumors, 5 x 106 U251 cells were injected subcutaneously as a single bolus into the left dorsum of 7-week-old male athymic nude mice (NCI; n = 24 mice). After 4 days, mice bearing a single tumor that was more than 20 mm3 (n = 20 mice) were randomly divided into three groups. Lentiviruses containing either LacZ shRNA (n = 9 mice)– or stathmin shRNA (n = 9 mice)–encoding sequences were injected intratumorally at 2.5 µg/100 µL for 3 consecutive days. Tumor-bearing mice in the third group (n = 2 mice) had no lentiviral injections. Two days after the final injection, one mouse from each of the two virus injections groups, as well as one mouse with an uninjected tumor, were killed by use of carbon dioxide narcosis, followed by pneumothorax induction, and tumors were removed for immunohistochemical analysis of stathmin expression as described above. The remaining mice in the LacZ and stathmin shRNA groups were randomly assigned to intraperitoneal administration of either BCNU (15 mg/kg) (n = 4 mice from the LacZ shRNA group; n = 4 mice from the stathmin shRNA group) or vehicle (10% ethanol) (n = 4 mice from the LacZ shRNA group; n = 4 mice from the stathmin shRNA group). Tumor length (D) and width (d) were measured every 3 or 4 days for 7 weeks with a caliper. Tumor volume (V) was calculated as V = 4/3
x (d/2)2 x D/2 (18). All procedures were in accordance with NIH Animal Care and Use Committee protocols.
Synthesis and Modification of Six-Histidine–Tagged Human Recombinant Stathmin
A full-length human stathmin cDNA (Origene Technologies) with the native stop codon replaced by 5'-CACCACCATCACCATCATTAA-3' (encoding for six C-terminal histidines plus a stop codon) was subcloned into the pIX 2.0 bacterial expression vector (Qiagen; Valencia, CA), which also contains a copy of the ampicillin resistance gene. Competent BL21(DE3) Escherichia coli cells (Novagen; San Diego, CA) were transformed with the six-histidine–tagged stathmin cDNA–containing plasmid, according to manufacturer's recommendations, and cultured on ampicillin-containing agar plates (KD Medical; Columbia, MD) at 37 °C overnight. Ampicillin-resistant clones were selected and expanded in Luria–Bertani broth (KD Medical) containing ampicillin (Novagen) at 50 µg/mL. Once cultures were in the logarithmic phase of growth, as determined by an optical density measurement at 600 nm of 0.6 with a SpectraMax Plus 384 microplate spectrophotometer, 0.5 mM isopropyl--D-thiogalactopyranoside (Novagen) was added to induce bacterial synthesis of the recombinant protein. When cultures reached an optical density of 1.2, bacteria were centrifuged at 6000g at 4 °C for 15 minutes. The bacterial pellet was lysed with X-Tractor buffer (Clontech; Mountain View, CA) to release the recombinant proteins, and the resulting lysate was clarified by centrifugation at 12000g at 4 °C for 15 minutes. The recombinant six-histidine–tagged stathmin was purified from this clarified lysate with TALON CellThru resin (Clontech), an immobilized metal affinity chromatography resin for purifying polyhistidine-tagged proteins. The purity of the six-histidine–tagged stathmin was confirmed by analysis after one-dimensional SDS–PAGE.
Purified six-histidine–tagged stathmin in 10 mM sodium phosphate (Sigma) buffer (pH 7.4) containing 150 mM sodium chloride (Sigma) was chemically modified by incubation with 0.005% DMSO vehicle control, 10 µM CCNU, or 10 µM temozolomide (a DNA-alkylating agent with no protein-carbamoylating activity) at 37 °C for 4 hours.
Reverse-Phase High-Performance Liquid Chromatography of Six-Histidine–Tagged Stathmin
Reverse-phase high-performance liquid chromatography (RP-HPLC) separation of native and chemically modified (as described above) six-histidine–tagged stathmin was performed with a nonporous reverse-phase column (Beckman Coulter; Fullerton, CA) at 50 °C with a flow rate of 0.75 mL/min and a 0%–100% gradient of solvent B (0.08% trifluoroacetic acid in acetonitrile) in solvent A (0.1% trifluoroacetic acid in water). Fractions were measured for absorbance at 214 nm and collected in 96-well plates every 0.25 minutes.
Fourier Transform–Ion Cyclotron Resonance Determination of Protein Molecular Weight
Intact protein (native and chemically modified six-histidine–tagged stathmin) mass spectra were obtained by use of electrospray on a Fourier transform–ion cyclotron resonance (FT-ICR) mass spectrometer (Thermo Scientific; Waltham, MA). RP-HPLC protein fractions (as described above) were syringe infused at 5 µL/min into the FT-ICR mass spectrometer, and ions generated from the positive electrospray were directed through a heated metal capillary, skimmer, and multiple ion guides into the FT-ICR ion cell. The charge states of the analyte were determined from the spacings of isotopic peaks. The molecular weights of analytes were calculated from charge-state deconvoluted spectra, as previously described (19).
Tubulin Depolymerization Assay
Tubulin depolymerization was used as a functional test of stathmin activity before and after chemical modification. The assay was performed as previously described (20). Briefly, 50 µL of bovine brain tubulin (Cytoskeleton; Denver, CO; 3 mg/mL) was loaded into the wells of a transparent 96-well plate (Corning) at 37 °C and allowed to polymerize at 37 °C to a steady state, as indicated by a stabilization of absorbance at 340 nm from a SpectraMax Plus 384 microplate spectrophotometer. Purified human recombinant stathmin (Calbiochem) in a solution of 10 mM phosphate buffer (pH 7.4), 150 mM sodium chloride, and 1mM EDTA was pretreated for 4 hours at 37 °C with 0.005% DMSO vehicle control, 10 µM CCNU, 10 µM BCNU, or 10 µM temozolomide and then added to the polymerized microtubules. Microtubule depolymerization was monitored by measuring the decrease in absorbance at 340 nm. Microtubule depolymerization with untreated stathmin was used as a positive control. The results are representative of three independent experiments, all with similar results.
Statistical Analysis
Spearman correlation (with statistical significance assessed by Pitman's permutation test) was used to determine the association between stathmin expression levels (with values of I, II, or III) and 1p status (1p+/+ coded as 1 and 1p+/– coded as 0) (21). Pitman's approach to determining statistical significance was selected because it handles discrete (i.e., noncontinuous) data better. Kaplan–Meier analyses that investigated the relationship of stathmin expression, 1p status, and MGMT expression of tumors to the progression-free survival times of patients were performed with SPSS version 14.0. P values were obtained with log-rank tests. Cell cytotoxicity assays were analyzed with two-way analysis of variance (ANOVA) models with dose level and stathmin expression as factors. P values reported are those for the effect of stathmin expression. When P values obtained from ANOVA models were considered to be statistically significant (i.e., P<.05), an analysis of drug sensitivity differences associated with stathmin expression differences was also performed for individual drug doses with Welch's t test. Cell cycle studies were analyzed with Student's t test. Adherence to the assumptions of the normal distribution was examined graphically, and little evidence of skewness or fold-dependent variance was present despite the large range in fold measures. Kaplan–Meier analyses of the intracranial tumor data were also performed with SPSS, and P values were obtained with log-rank tests. For the subcutaneous tumor experiments, natural logarithm–transformed tumor volumes were analyzed with a multivariable ANOVA (MANOVA) approach over time with BCNU and STMNi virus as treatment effects. Fisher's exact test was used to examine the association between MGMT expression and 1p status. For all statistical tests, a P value of less than .05 was considered to be statistically significant. Statistical analyses were performed with SPSS version 14.0 (SPSS; Chicago, IL) and JMP 5.1 (SAS Institute Inc; Cary, NC). All statistical tests were two-sided.
|
Results |
|---|
|
TopAbstractContext and CaveatsMaterials and MethodsResultsDiscussionReferencesNotes |
|---|
Differential Proteomic Screening of 1p+/+ and 1p+/– Malignant Glioma Cell Lines
We carried out a comparative proteomic screen of A172 (1p+/–) and U251 (1p+/+) malignant glioma cells to identify differentially expressed 1p-encoded proteins. The proteomic screen consisted of a quantitative comparison of proteins expressed by A172 and U251 cells with two-dimensional SDS–PAGE, followed by the identification of individual differentially expressed proteins with mass spectroscopy. The A172 malignant glioma cell line was selected as the 1p+/– cell line because it has large regions of loss of heterozygosity in 1p that are similar to those observed in sporadic gliomas (Fig. 1). The U251 malignant glioma cell line was selected as the 1p+/+ cell line because it is heterozygous for 1p (Fig. 1, A) and more closely resembles A172 in the expression of proteins considered to be important in glioma tumorigenesis (p53, p21, MDM-2, p16, and pRB) than Hs683, T98G, and U87 (22,23).
|
A172 and U251 cellular proteins labeled with Cy3- and Cy5-CyDye DIGE Fluor minimal dyes, respectively, were separated on a common two-dimensional SDS polyacrylamide gel (Fig. 2, A). Twenty-nine protein spots on the gel had Cy3/Cy5 labeling ratios of less than 0.67, as determined by the DeCyder software program, and gel plugs corresponding to these spots were excised and processed for identification of the contained proteins by MALDI-TOF/TOF MS. When different isoforms of the same protein were found in separate spots, mean expression levels for the isoforms were determined. By this method, 18 unique proteins were identified, three of which are encoded on 1p—i.e.,
-enolase, stathmin, and DJ-1 protein (Table 1). Of these three differentially expressed 1p-encoded proteins, the locus for stathmin (i.e., STMN1 or OP18) is the only one that is located between microsatellite markers D1S482 and D1S513, which is the minimal chromosomal region most commonly deleted in 1p+/– clinical glioma specimens (4,24–26). By quantitative immunoblot analysis, we determined that the level of stathmin expression in A172 cells was 50% of that in U251 cells (Fig. 2, B and C), consistent with the 1p status of the cells. Stathmin was therefore selected for further detailed study.
|
Association of Stathmin Expression With 1p Loss of Heterozygosity and Recurrence-Free Survival
To determine the clinical significance of differential stathmin expression, 24 specimens of surgically resected anaplastic oligodendrogliomas were double-labeled with antibodies against stathmin and antibodies against olig2; anti-olig2 was used to distinguish oligodendroglioma tumor cells from nonneoplastic neurons, astrocytes, and vascular and inflammatory cells (15,27). Stathmin expression in each tumor was scored as level I, II, or III (Fig. 3, A–C, respectively). Of the 13 1p+/– specimens, nine had level I stathmin expression and four had level II expression. Of the 11 1p+/+ specimens, five had level II expression and six had level III expression (Fig. 3, E). Stathmin levels of the 1p+/– specimens were statistically significantly lower than those of the 1p+/+ specimens (Spearman correlation coefficient = 0.76 and Pitman's test P<.001).
|
Patients were stratified according to the stathmin level (I–III) in their tumors, and their recurrence-free survival times were examined. The median recurrence-free survival for patients with tumors with stathmin level I (i.e., 0%–33% of maximal stathmin expression) was 45 months (95% CI = 0 to 90 months), for those with level II (i.e., 34%–66% of maximal stathmin expression) was 17 months (95% CI = 10.6 to 23.4 months), and for those with level III (i.e., 67%–100% of maximal stathmin expression) was 6 months (95% CI = 1.7 to 10.3 months) (log-rank test; P<.001) (Fig. 3, F). Median recurrence-free survival times for patients stratified according to 1p status were 24 months (95% CI = 17.3 to 30.7 months) for those with 1p+/– tumors and 6 months (95% CI = 3.4 to 8.6 months) for those with 1p+/+ tumors (log-rank test; P<.001) (Fig. 3, G).
Stathmin Expression Levels and In Vitro and In Vivo Chemoresistance of Malignant Glioma Cell Lines
To examine the functional consequences of changes in stathmin expression, a U251-derived stable cell line (termed U251-STMNi) carrying inducible U6 promoter and stathmin-specific shRNA–encoding oligonucleotide sequences (17,28) was established. Muristerone A enhances the activity of the inducible U6 promoter and thus induces the transcription of stathmin-specific shRNAs in U251-STMNi cells; these shRNAs inhibit the expression of stathmin. Noninduced U251-STMNi cells had approximately wild-type levels of stathmin expression (U251-STMN+/+), but cells incubated with 10 µM muristerone A expressed approximately 50% of the wild-type stathmin level (U251-STMN+/–) (Fig. 4, A). These results reflect the stathmin levels present in 1p+/+ and 1p+/– malignant glioma cells, respectively (Fig. 2, C). Higher doses of muristerone A (>15 µM) further reduced the level of stathmin expression in these cells and also inhibited cell proliferation. U251-STMN+/– and U251-STMN+/+ cells did not differ statistically significantly in their resistance to procarbazine (ANOVA P = .328), vincristine (ANOVA P = .196), or temozolomide (ANOVA P = .454), an imidazotetrazine DNA-alkylating agent that is used widely in the treatment of malignant glioma patients (29) (Fig. 4, B–D). U251-STMN+/– cells were, however, statistically significantly less resistant to the nitrosourea CCNU than U251-STMN+/+ cells over a large range of CCNU doses (ANOVA P<.001 for dose by cell type interaction) (Fig. 4, E). The viabilities of U251-STMN+/– and U251-STMN+/+ cells treated with 10 µM CCNU were 63% and 97% of control untreated cells, respectively (difference = 34%, 95% CI = 25% to 42%). After treatment with 20 µM CCNU, the viabilities of U251-STMN+/– and U251-STMN+/+ cells were 18% and 69% of control untreated cells, respectively (difference = 50%, 95% CI = 45% to 56%). To rule out nonspecific shRNA effects and muristerone A toxicity as explanations for this difference, U251 cells stably transfected with a muristerone A–inducible LacZ shRNA construct (termed U251-LacZi) were also generated. Treatment of U251-LacZi cells with 10 µM muristerone A did not statistically significantly alter their stathmin expression (Fig. 4, A) or CCNU resistance (ANOVA P = .150) (Fig. 4, F).
|
Transfection of A172 cells with stathmin cDNA increased their resistance to CCNU (ANOVA P<.001 for dose and cell type interaction) (Fig. 4, G). Transfection of A172 cells with cDNAs for
-enolase (ENOA) or DJ-1 (PARK7), the two other 1p-encoded proteins identified in our differential proteomic screen (Table 1), had no effect on their CCNU resistance (ANOVA P = .902 and .784, respectively) (Fig. 4, G). To determine the in vivo effects of altering stathmin levels, U251-STMN+/– or U251-STMN+/+ cells were stereotactically implanted into the right striata of severe combined immunodeficient mice. Three days after implantation, mice were injected intraperitoneally with either the nitrosourea BCNU or vehicle (10% ethanol) as control. The median overall survival of mice in the U251-STMN+/– group treated with BCNU was 95 days (95% CI = 68.7 to 121.3 days), that of mice in the U251-STMN+/– group treated with vehicle was 69 days (95% CI = 65.7 to 72.3 days), that of mice in the U251-STMN+/+ group treated with BCNU was 64 days (95% CI = 58.2 to 69.8 days), and that of mice in the U251-STMN+/+ group treated with vehicle was 63 days (95% CI = 60.1 to 65.9 days) days (n = 9 mice for each of the four groups). Overall survival of mice in the U251-STMN+/– group treated with BCNU was statistically significantly longer than that in the three other groups (difference with vehicle-treated U251-STMN+/– group = 26 days, 95% CI = –0.5 to 53.5 days, log-rank test P<.001; difference with U251-STMN+/+ group treated with BCNU = 31 days, 95% CI = 4.1 to 57.9 days, log-rank test P<.001; and difference with U251-STMN+/+ group treated with vehicle = 32 days, 95% CI = 5.5 to 58.4 days, log-rank test P = .001). The overall survivals of the vehicle-treated U251-STMN+/– group, the BCNU-treated U251-STMN+/+ group, and the vehicle-treated U251-STMN+/+ group did not statistically significantly differ from each other (log rank test P = .653) (Fig. 4, H).
Stathmin Expression, Nitrosourea Treatment, and Cell Cycle Progression
To increase understanding of the effects of decreased stathmin expression on nitrosourea resistance, we examined the progression of U251-STMNi cells through the G2 and M phases of the cell cycle as a function of stathmin level. The proportion of cells in G2 phase was not affected by decreasing stathmin expression by 50% after muristerone A treatment (Student's t test P = .925), but the proportion increased 3.9-fold after CCNU treatment (Student's t test P = .004). When stathmin expression was decreased by 50% and cells were treated with CCNU, the proportion of cells in G2 phase increased 3.7-fold (Student's t test P = .020), but this increase was not statistically significantly different from the increase associated with treatment with CCNU alone (Student's t test P = .831). The proportion of cells in M phase of the cell cycle was increased 1.5-fold by decreased stathmin expression, although the increase was not statistically significant (Student's t test P = .097), and it increased 4.2-fold with CCNU treatment (Student's t test P = .003). The combination of decreased stathmin expression and CCNU treatment, however, increased the proportion of cells in M phase by 7.2-fold (Student's t test P<.001), which exceeds both the calculated sum (5.7-fold) and calculated product (6.3-fold) of the individual increases (Fig. 5, A).
|
An increase in the M-phase population when cell proliferation is decreased (Fig. 4, E) is consistent with the cells undergoing mitotic arrest. To investigate this possibility further, the formation of p55cdc, Mad2, and cdc27 protein–containing complexes—an indicator of spindle checkpoint activation and mitotic arrest—was examined (30). Lysates of U251-STMN+/+ and U251-STMN+/– cells, each with and without CCNU treatment, were immunoprecipitated with anti-p55cdc antibodies and then immunoblotted with anti-p55cdc, anti-Mad2, or anti-cdc27 antibodies. Equivalent amounts of p55cdc were precipitated from each sample, but the highest levels of Mad2 and cdc27 were coprecipitated from lysates of CCNU-treated U251-STMN+/– cultures (Fig. 5, B). This result is indicative of increased p55cdc–Mad2–cdc27 complex formation and therefore increased mitotic arrest in CCNU-treated U251-STMN+/– cells. Furthermore, immunoblot analyses of whole-cell lysates found that lysates of CCNU-treated U251-STMN+/– cultures also had the highest level of total securin (Fig. 5, B), another indicator of spindle checkpoint activation (30). Thus, U251-STMNi cells with decreased stathmin expression undergo a higher level of mitotic arrest in response to CCNU treatment than those with wild-type levels of stathmin expression.
Stathmin Function and Nitrosourea Treatment
Because nitrosoureas, such as CCNU and BCNU, are customarily regarded as DNA-alkylating agents but can also carbamoylate proteins, particularly on N-terminal amino groups and 
-amino groups of lysine side chains (31), we investigated whether stathmin was carbamoylated in the presence of CCNU. When cellular protein extracts of DMSO (i.e., vehicle)-treated control cells and cells treated with 10 µM CCNU were compared with two-dimensional immunoblot analysis, stathmin isoforms from CCNU-treated cells had lower isoelectric points than those from untreated cells (Fig. 6, A). This observation is consistent with CCNU-mediating carbamoylation of stathmin, particularly on lysine residues (Fig 6, B).
|
To confirm that CCNU mediates the chemical modification of stathmin, recombinant C-terminal, six-histidine–tagged stathmin proteins were treated with 0.005% DMSO (vehicle control), 10 µM CCNU, or 10 µM temozolomide (a DNA-alkylating agent with no protein-carbamoylating activity). We predicted that CCNU-mediated addition of cyclohexylcarbamoyl groups to lysines in stathmin would increase its hydrophobicity and molecular mass. RP-HPLC analysis of DMSO-treated stathmin or temozolomide-treated stathmin detected only a single protein species (Fig. 6, C, and data not shown, respectively). FT-ICR mass spectrometry analysis of DMSO-treated stathmin revealed a molecular mass of 18433.92 Da. The calculated average mass of six-histidine–tagged stathmin was 18434.62 Da. After CCNU treatment of stathmin and RP-HPLC analysis, many stathmin protein species were found, including a native stathmin peak and multiple new broadly spread stathmin peaks with higher hydrophobicity than the native peak (Fig. 6, C). The molecular masses of the proteins in the four most prominent new peaks, as determined by FT-ICR mass spectrometry, exceed that of native stathmin by one to four increments of 125 Da, the molecular mass of one cyclohexylcarbamoyl group.
Because the best characterized primary biologic role of stathmin is the regulation of microtubule dynamics, we next examined the effect of cyclohexylcarbamoyl modification of stathmin on this function. As free tubulin in solution polymerizes and forms microtubules of increasing length, the optical density of the solution increases. If the microtubules subsequently depolymerize and decrease in length, the optical density of the solution decreases. Addition of stathmin to a solution of polymerized microtubules induces its depolymerization. By 1 minute after the addition of either 0.005% DMSO- or 10 µM temozolomide-treated stathmin to a polymerized tubulin solution, the optical density of the solution decreased 50%. In contrast, 50% decreases in optical density were achieved by 14 and 15 minutes after the addition of stathmin pretreated with 10 µM CCNU or 10 µM BCNU, respectively (Fig. 6, D). Addition of 0.005% DMSO, 10 µM CCNU, 10 µM BCNU, or 10 µM temozolomide to polymerized microtubules had no depolymerizing effect in the absence of stathmin. Addition of 0.005% DMSO, 10 µM CCNU, 10 µM BCNU, or 10 µM temozolomide concurrently with stathmin also had no inhibitory effect on stathmin-mediated microtubule depolymerization. The two-dimensional immunoblot, RP-HPLC, FT-ICR mass spectrometry, and microtubule depolymerization assay results indicate that nitrosoureas can carbamoylate and inhibit the microtubule-depolymerizing function of stathmin.
Decreased Stathmin Expression, Nitrosourea, and Tumor Growth
We next assessed the therapeutic potential of combining RNA interference to reduce stathmin expression with nitrosourea treatment by use of a xenograft mouse model. Wild-type U251 cells were subcutaneously implanted in the left flank regions of nude mice to generate one solid tumor per mouse that could subsequently be monitored by direct measuring. Four days after implantation, tumors were injected on 3 consecutive days with lentiviruses containing DNA constructs for the constitutive expression of shRNAs directed against either stathmin (STMN shRNA) or LacZ (LacZ shRNA) as a control. Two days after the final injections, a randomly selected mouse from each of the two groups was killed, and tumors were excised and examined for stathmin expression by use of immunohistochemistry. Tumors injected with STMN shRNA had a qualitatively lower level of stathmin expression than tumors injected with LacZ shRNA tumors, which had wild-type levels (Fig. 7, A). The remaining mice were randomly assigned to treatment with either BCNU or drug vehicle, resulting in four study groups (STMN shRNA and BCNU, n = 4 mice; STMN shRNA and drug vehicle, n = 4 mice; LacZ shRNA and BCNU, n = 4 mice; and LacZ shRNA and drug vehicle, n = 4 mice). Tumor size was measured over time, and the study was terminated 49 days after initial tumor cell implantation because of the large sizes of the tumors in the LacZ shRNA group. Neither BCNU nor STMN shRNA treatment alone statistically significantly altered tumor growth compared with LacZ shRNA and drug vehicle control treatment (MANOVA P = .124 and P = .306, respectively). Combination treatment with BCNU and STMN shRNA, however, statistically significantly decreased tumor growth compared with control treatment (MANOVA P = .005). At 49 days after initial tumor implantation, the mean tumor volumes of the control and combination (i.e., BCNU and STMN shRNA) treatment groups were 2312 mm3 (95% CI = 1044 to 3580 mm3) and 391 mm3 (95% CI = 48 to 734 mm3), respectively (difference = 1921 mm3, 95% CI of difference = 354 to 3487 mm3).
|
|
Discussion |
|---|
|
TopAbstractContext and CaveatsMaterials and MethodsResultsDiscussionReferencesNotes |
|---|
We have identified stathmin as a candidate 1p chemoresistance protein for malignant gliomas and found that decreased stathmin expression in anaplastic oligodendroglioma tumors was statistically significantly associated with loss of heterozygosity of 1p and increased progression-free survival of patients after chemotherapy treatment. In vitro and in vivo experiments using RNA interference demonstrated that tumor cells with decreased stathmin expression were less resistant to nitrosourea treatment than those with wild-type levels of stathmin expression. Biochemical experiments demonstrated that stathmin was carbamoylated and functionally inhibited by nitrosoureas and revealed a direct mechanism of interaction between stathmin and nitrosoureas. We found also that the decrease in stathmin function mediated by the combination of nitrosourea treatment and RNA interference was effective at suppressing the growth of experimental tumors by more than 80%.
Stathmin was initially identified as a protein that is overexpressed in acute leukemia (32) and as a ubiquitously expressed, neuron-enriched phosphoprotein (33). Stathmin also appears to participate in the regulation of microtubule dynamics and cell cycle progression (34). For microtubule polymerization, mitotic spindle assembly, and entry of cells into mitosis, stathmin must be inactivated by phosphorylation on serine residues; for microtubule shortening, disassembly of the mitotic spindle, and progression of cells out of mitosis, stathmin must be activated by dephosphorylation of those serine residues (35).
To establish whether decreased stathmin expression, 1p loss of heterozygosity, and clinical outcomes are associated, we analyzed a series of anaplastic oligodendroglioma specimens, the glioma subtype in which loss of heterozygosity at 1p occurs most frequently. Indeed, lower stathmin expression was found in 1p+/– specimens than in 1p+/+ specimens. Decreased stathmin expression in tumors was also statistically significantly associated with increased recurrence-free survival after treatment with procarbazine, CCNU, and vincristine. Sequence analysis of the stathmin gene locus showed that no tumors examined carried stathmin mutations or polymorphisms (data not shown); this result also supports the association between low stathmin expression and increased recurrence-free survival. If this association is confirmed and validated in large cohorts of patients, immunohistochemical assessment of stathmin expression might be used when fluorescent in situ hybridization analysis for 1p status is unavailable for technical reasons.
We found that reduced stathmin expression decreased the resistance of tumor cells to CCNU treatment, but not to treatment with procarbazine, vincristine, or temozolomide, indicating that stathmin expression is associated with resistance to nitrosoureas, not resistance to all drugs. Overexpression of stathmin, but not of two other proteins encoded on 1p, -enolase or DJ-1, increased resistance to CCNU, indicating that resistance is not associated with the overexpression of randomly selected 1p-encoded proteins. Thus, these in vitro results support the specificity of the interaction between CCNU and stathmin. A xenogeneic intracranial glioma model provided in vivo confirmation of the in vitro results, in that nitrosourea treatment resulted in statistically significantly longer survival in mice implanted with tumors with reduced stathmin expression than it did in mice implanted with tumors with wild-type levels of stathmin expression.
Why are malignant glioma cells with decreased stathmin expression more sensitive to nitrosoureas? Many cell types require a threshold level of functional stathmin to progress through mitosis (36–38). In K562 leukemia cells, for example, mitosis is arrested when the level of stathmin mRNA is reduced by 90% (38). Decreasing stathmin protein expression to 90% of its normal level in U251-STMNi cells also caused growth arrest. A mechanism by which large reductions in stathmin levels may arrest mitosis is through activation of the spindle checkpoint. During the transition from metaphase to anaphase, stathmin induces the poleward movement of kinetochore-attached spindle microtubules by increasing the catastrophe frequency at their minus ends (39–41). When stathmin levels fall below a critical threshold, the tension normally generated at kinetochores by this stathmin-mediated spindle shortening may be insufficient and result in spindle checkpoint activation (42,43). In U251-STMNi cells, a 50% reduction in stathmin expression did not by itself appreciably affect mitotic progression and cell proliferation. However, additional stathmin inactivation in these cells with nitrosourea-mediated carbamoylation may reduce functional stathmin levels to below a critical threshold and induce checkpoint activation, mitotic arrest, and decreased proliferation, as demonstrated in this study. Although nitrosoureas can potentially carbamoylate all cellular proteins, the high lysine content of stathmin (i.e., 15.5%) renders it a ready substrate. Additional studies are required to further characterize this mechanism.
Although our results indicate that decreased expression of stathmin is associated with 1p loss of heterozygosity and with more favorable clinical outcomes in patients with anaplastic oligodendrogliomas, it has been reported by others (44) that decreased expression of MGMT is associated with more favorable clinical outcomes in patients with glioblastomas, another malignant glioma subtype. We therefore examined the expression of MGMT in anaplastic oligodendroglioma specimens and found no association between MGMT protein expression and 1p status (Fisher's exact test P = .80) or progression-free survival (log-rank test P = .841) (Supplementary Fig. 1, A–C; available online). These results are consistent with those of previously published studies (16,45). Furthermore, we found no in vitro changes in MGMT expression associated with alterations in stathmin expression (Supplementary Fig. 1, D; available online). Consequently, the role of MGMT expression in the favorable outcomes of patients with 1p+/– anaplastic oligodendroglioma tumors remains to be determined.
The frequent loss of heterozygosity of 1p in oligodendroglial tumors has prompted an extensive search for tumor suppressor genes in this region (46,47). Although loss of a putative 1p-encoded tumor suppressor gene is likely to confer a growth advantage to tumor cells, decreased stathmin expression did not appear to have such an effect. There is also no evidence that decreased stathmin expression is associated with tumor initiation. It is therefore unlikely that stathmin is a tumor suppressor. The allelic loss of the stathmin gene is likely a fortuitous side effect of its proximity to an as yet unidentified tumor suppressor gene or of the frequent loss of the entire 1p arm when deletions involving 1p occur (3,46,47).
One limitation of our study is that comparative proteomic screening methods to identify differentially expressed proteins, although more direct, are less comprehensive than gene expression microarray methods. Stathmin haploinsufficiency appears to contribute to the chemosensitivity of 1p+/– gliomas, but there are likely other contributing factors as well. An additional limitation is the use of a three-tier system, rather than a two-tier system, for grading the stathmin expression levels of the clinical tumor specimens. If 1p+/– tumor cells have one stathmin gene copy, if 1p+/+ tumor cells have two stathmin gene copies, and if all copies are nonmutated and active, one would expect stathmin expression levels to be binary (i.e., the expression levels of 1p+/+ tumors should be twice that of 1p+/– tumors) and readily distinguishable. We did not, however, find that all tumors could be classified into one of two groups. Although 15 of the 24 tumors clearly had low (n = 9) or high (n = 6) stathmin expression, nine of the tumors had intermediate stathmin expression. We therefore devised a three-tier grading system that allowed us to reach a consensus on the stathmin expression grades of all the tumors. Possible explanations include the semiquantitative nature of immunohistochemical labeling and the intratumoral genetic heterogeneity of clinical specimens. Tumors that are predominantly composed of 1p+/– neoplastic cells can also contain a minority of 1p+/+ neoplastic cells, and vice versa. In sufficient quantities, the 1p+/+ cells within a predominantly 1p+/– tumor could cause the tumor to have an intermediate rather than low level of stathmin expression.
Our studies demonstrate specific and statistically significant in vitro and in vivo changes in nitrosourea sensitivity in response to experimental manipulations of stathmin expression, a direct mechanism of interaction between stathmin and nitrosoureas, and a mouse treatment paradigm combining nitrosoureas with RNA interference–mediated decreases in stathmin expression. The challenge is to apply these findings to the development of an effective therapy for the patients with treatment-resistant 1p+/+ gliomas.
|
NOTES |
|---|
|
TopAbstractContext and CaveatsMaterials and MethodsResultsDiscussionReferencesNotes |
|---|
T.-T. B. Ngo, T. Peng, and X.-J. Liang contributed equally to this work.
This work was supported by the NINDS, NCI, and NHLBI Intramural Research Programs of the NIH. T. Peng was supported by the Clinical Research Training Program, NIH. O. Akeju was supported by the Howard Hughes Medical Institute—NIH Research Scholars Program.
P. Y. Wen is a member of the Speakers' Bureau for Schering-Plough, the makers of temozolomide. He is currently conducting research sponsored by Amgen, Genentech, Novartis, Schering, Celgene, Glaxo Smith Kline, and Astra Zeneca. The other authors declare that they have no financial conflicts of interest.
The authors had full responsibility for the design of the study, the collection of the data, the analysis and interpretation of the data, the decision to submit the manuscript for publication, and the writing of the manuscript.
We thank H. Gainer, T. Friedman, C. Thiele, R. Youle, S. Landis, and G. Park for helpful discussions and critical reading of the manuscript; H. Jaffe for assistance with protein analysis; C. Stiles and J. Alberta for providing anti-olig2 antibodies; J. Richman, A. Kellogg, M. Gravell, and R. Bailey for technical assistance; D. Sackett for advice on the tubulin assays; L. Doherty and D. Gigas for assistance with patient data collection; and K. Pike for stathmin gene sequence analysis.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsMaterials and MethodsResultsDiscussionReferencesNotes |
|---|
(1) Central Brain Tumor Registry of the United States. Statistical report: primary brain tumors in the United States, 1997-2001. Central Brain Tumor Registry of the United States, Chicago, IL (2004).
(2) Kleihues P, Cavenee WK, eds. Tumours of the nervous system: pathology and genetics. In: World Health Organization Classification of Tumours (2000) Lyon (France): IARC Press.
(3) Cairncross JG, Ueki K, Zlatescu MC, Lisle DK, Finkelstein DM, Hammond RR, et al. Specific genetic predictors of chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas. J Natl Cancer Inst (1998) 90:1473–9.
(4) Hashimoto N, Murakami M, Takahashi Y, Fujimoto M, Inazawa J, Mineura K. Correlation between genetic alteration and long-term clinical outcome of patients with oligodendroglial tumors, with identification of a consistent region of deletion on chromosome arm 1p. Cancer (2003) 97:2254–61.[CrossRef][Web of Science][Medline]
(5) Ino Y, Betensky RA, Zlatescu MC, Sasaki H, Macdonald DR, Stemmer-Rachamimov AO, et al. Molecular subtypes of anaplastic oligodendroglioma: implications for patient management at diagnosis. Clin Cancer Res (2001) 7:839–45.
(6) van den Bent MJ, Kros JM, Heimans JJ, Pronk LC, van Groeningen CJ, Krouwer HG, et al. Response rate and prognostic factors of recurrent oligodendroglioma treated with procarbazine, CCNU, and vincristine chemotherapy. Dutch Neuro-oncology Group. Neurology (1998) 51:1140–5.
(7) van den Bent MJ, Looijenga LH, Langenberg K, Dinjens W, Graveland W, Uytdewilligen L, et al. Chromosomal anomalies in oligodendroglial tumors are correlated with clinical features. Cancer (2003) 97:1276–84.[CrossRef][Web of Science][Medline]
(8) Walker C, du Plessis DG, Joyce KA, Fildes D, Gee A, Haylock B, et al. Molecular pathology and clinical characteristics of oligodendroglial neoplasms. Ann Neurol (2005) 57:855–65.[CrossRef][Web of Science][Medline]
(9) Cairncross G, Berkey B, Shaw E, Jenkins R, Scheithauer B, Brachman D, et al. Phase III trial of chemotherapy plus radiotherapy compared with radiotherapy alone for pure and mixed anaplastic oligodendroglioma: Intergroup Radiation Therapy Oncology Group Trial 9402. J Clin Oncol (2006) 24:2707–14.
(10) van den Bent MJ, Carpentier AF, Brandes AA, Sanson M, Taphoorn MJ, Bernsen HJ, et al. Adjuvant procarbazine, lomustine, and vincristine improves progression-free survival but not overall survival in newly diagnosed anaplastic oligodendrogliomas and oligoastrocytomas: a randomized European Organisation for Research and Treatment of Cancer phase III trial. J Clin Oncol (2006) 24:2715–22.
(11) Schmidt MC, Antweiler S, Urban N, Mueller W, Kuklik A, Meyer-Puttlitz B, et al. Impact of genotype and morphology on the prognosis of glioblastoma. J Neuropathol Exp Neurol (2002) 61:321–8.[Web of Science][Medline]
(12) Kujas M, Lejeune J, Benouaich-Amiel A, Criniere E, Laigle-Donadey F, Marie Y, et al. Chromosome 1p loss: a favorable prognostic factor in low-grade gliomas. Ann Neurol (2005) 58:322–6.[CrossRef][Web of Science][Medline]
(13) Ino Y, Zlatescu MC, Sasaki H, Macdonald DR, Stemmer-Rachamimov AO, Jhung S, et al. Long survival and therapeutic responses in patients with histologically disparate high-grade gliomas demonstrating chromosome 1p loss. J Neurosurg (2000) 92:983–90.[Web of Science][Medline]
(14) Huang J, Wei W, Zhang J, Liu G, Bignell GR, Stratton MR, et al. Whole genome DNA copy number changes identified by high density oligonucleotide arrays. Hum Genomics (2004) 1:287–99.[Medline]
(15) Ligon KL, Alberta JA, Kho AT, Weiss J, Kwaan MR, Nutt CL, et al. The oligodendroglial lineage marker OLIG2 is universally expressed in diffuse gliomas. J Neuropathol Exp Neurol (2004) 63:499–509.[Web of Science][Medline]
(16) McLendon RE, Herndon JE 2nd, West B, Reardon D, Wiltshire R, Rasheed BK, et al. Survival analysis of presumptive prognostic markers among oligodendrogliomas. Cancer (2005) 104:1693–9.[CrossRef][Web of Science][Medline]
(17) Gupta S, Schoer RA, Egan JE, Hannon GJ, Mittal V. Inducible, reversible, and stable RNA interference in mammalian cells. Proc Natl Acad Sci U S A (2004) 101:1927–32.
(18) Michieli P, Mazzone M, Basilico C, Cavassa S, Sottile A, Naldini L, et al. Targeting the tumor and its microenvironment by a dual-function decoy Met receptor. Cancer Cell (2004) 6:61–73.[CrossRef][Web of Science][Medline]
(19) Reid GE, McLuckey SA. ‘Top down’ protein characterization via tandem mass spectrometry. J Mass Spectrom (2002) 37:663–75.[CrossRef][Web of Science][Medline]
(20) Shelanski ML, Gaskin F, Cantor CR. Microtubule assembly in the absence of added nucleotides. Proc Natl Acad Sci U S A (1973) 70:765–8.
(21) Good P. Permutation tests: a practical guide to resampling methods for testing hypotheses. (1994) New York: Springer-Verlag.
(22) Kitange GJ, Templeton KL, Jenkins RB. Recent advances in the molecular genetics of primary gliomas. Curr Opin Oncol (2003) 15:197–203.[CrossRef][Web of Science][Medline]
(23) Law ME, Templeton KL, Kitange G, Smith J, Misra A, Feuerstein BG, et al. Molecular cytogenetic analysis of chromosomes 1 and 19 in glioma cell lines. Cancer Genet Cytogenet (2005) 160:1–14.[CrossRef][Web of Science][Medline]
(24) Felsberg J, Erkwoh A, Sabel MC, Kirsch L, Fimmers R, Blaschke B, et al. Oligodendroglial tumors: refinement of candidate regions on chromosome arm 1p and correlation of 1p/19q status with survival. Brain Pathol (2004) 14:121–30.[Web of Science][Medline]
(25) Husemann K, Wolter M, Buschges R, Bostrom J, Sabel M, Reifenberger G. Identification of two distinct deleted regions on the short arm of chromosome 1 and rare mutation of the CDKN2C gene from 1p32 in oligodendroglial tumors. J Neuropathol Exp Neurol (1999) 58:1041–50.[Web of Science][Medline]
(26) Iuchi T, Namba H, Iwadate Y, Shishikura T, Kageyama H, Nakamura Y, et al. Identification of the small interstitial deletion at chromosome band 1p34-p35 and its association with poor outcome in oligodendroglial tumors. Genes Chromosomes Cancer (2002) 35:170–5.[CrossRef][Web of Science][Medline]
(27) Lu QR, Park JK, Noll E, Chan JA, Alberta J, Yuk D, et al. Oligodendrocyte lineage genes (OLIG) as molecular markers for human glial brain tumors. Proc Natl Acad Sci U S A (2001) 98:10851–6.
(28) Akeju O, Peng T, Park JK. Short hairpin RNA loop design for the facilitation of sequence verification. Biotechniques (2006) 40:154, 156, 158.
(29) Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med (2005) 352:987–96.
(30) Weaver BA, Cleveland DW. Decoding the links between mitosis, cancer, and chemotherapy: the mitotic checkpoint, adaptation, and cell death. Cancer Cell (2005) 8:7–12.[CrossRef][Web of Science][Medline]
(31) Cheng CJ, Fujimura S, Grunberger D, Weinstein IB. Interaction of 1-(2-chloroethyl)-3-cyclohexyl-1-nitrosourea (NSC 79037) with nucleic acids and proteins in vivo and in vitro. Cancer Res (1972) 32:22–7.
(32) Hanash SM, Strahler JR, Kuick R, Chu EH, Nichols D. Identification of a polypeptide associated with the malignant phenotype in acute leukemia. J Biol Chem (1988) 263:12813–5.
(33) Sobel A, Boutterin MC, Beretta L, Chneiweiss H, Doye V, Peyro-Saint-Paul H. Intracellular substrates for extracellular signaling. Characterization of a ubiquitous, neuron-enriched phosphoprotein (stathmin). J Biol Chem (1989) 264:3765–72.
(34) Cassimeris L. The oncoprotein 18/stathmin family of microtubule destabilizers. Curr Opin Cell Biol (2002) 14:18–24.[CrossRef][Web of Science][Medline]
(35) Rubin CI, Atweh GF. The role of stathmin in the regulation of the cell cycle. J Cell Biochem (2004) 93:242–50.[CrossRef][Web of Science][Medline]
(36) Larsson N, Marklund U, Gradin HM, Brattsand G, Gullberg M. Control of microtubule dynamics by oncoprotein 18: dissection of the regulatory role of multisite phosphorylation during mitosis. Mol Cell Biol (1997) 17:5530–9.[Abstract]
(37) Larsson N, Melander H, Marklund U, Osterman O, Gullberg M. G2/M transition requires multisite phosphorylation of oncoprotein 18 by two distinct protein kinase systems. J Biol Chem (1995) 270:14175–83.
(38) Luo XN, Mookerjee B, Ferrari A, Mistry S, Atweh GF. Regulation of phosphoprotein p18 in leukemic cells. Cell cycle regulated phosphorylation by p34cdc2 kinase. J Biol Chem (1994) 269:10312–8.
(39) Manna T, Thrower D, Miller HP, Curmi P, Wilson L. Stathmin strongly increases the minus end catastrophe frequency and induces rapid treadmilling of bovine brain microtubules at steady state in vitro. J Biol Chem (2006) 281:2071–8.
(40) Mitchison TJ, Salmon ED. Poleward kinetochore fiber movement occurs during both metaphase and anaphase-A in newt lung cell mitosis. J Cell Biol (1992) 119:569–82.
(41) Waters JC, Mitchison TJ, Rieder CL, Salmon ED. The kinetochore microtubule minus-end disassembly associated with poleward flux produces a force that can do work. Mol Biol Cell (1996) 7:1547–58.[Abstract]
(42) Li X, Nicklas RB. Mitotic forces control a cell-cycle checkpoint. Nature (1995) 373:630–2.[CrossRef][Medline]
(43) Stern BM, Murray AW. Lack of tension at kinetochores activates the spindle checkpoint in budding yeast. Curr Biol (2001) 11:1462–7.[CrossRef][Web of Science][Medline]
(44) Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med (2005) 352:997–1003.
(45) Watanabe T, Nakamura M, Kros JM, Burkhard C, Yonekawa Y, Kleihues P, et al. Phenotype versus genotype correlation in oligodendrogliomas and low-grade diffuse astrocytomas. Acta Neuropathol (Berl) (2002) 103:267–75.[CrossRef][Medline]
(46) Reifenberger J, Reifenberger G, Liu L, James CD, Wechsler W, Collins VP. Molecular genetic analysis of oligodendroglial tumors shows preferential allelic deletions on 19q and 1p. Am J Pathol (1994) 145:1175–90.[Abstract]
(47) Bello MJ, Vaquero J, de Campos JM, Kusak ME, Sarasa JL, Saez-Castresana J, et al. Molecular analysis of chromosome 1 abnormalities in human gliomas reveals frequent loss of 1p in oligodendroglial tumors. Int J Cancer (1994) 57:172–5.[Web of Science][Medline]
Manuscript received October 4, 2006; revised February 8, 2007; accepted March 5, 2007.
Related Article in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 577.
This article has been cited by other articles:
|
|
 |
|
  S. Singer, M. Malz, E. Herpel, A. Warth, M. Bissinger, M. Keith, T. Muley, M. Meister, H. Hoffmann, R. Penzel, et al. Coordinated Expression of Stathmin Family Members by Far Upstream Sequence Element-Binding Protein-1 Increases Motility in Non-Small Cell Lung Cancer Cancer Res., March 15, 2009; 69(6): 2234 - 2243. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Y. Wen and S. Kesari Malignant Gliomas in Adults N. Engl. J. Med., July 31, 2008; 359(5): 492 - 507. [Full Text] [PDF]
|
|
|
|
|
|
X.-J. Liang, Y. Choi, D. L. Sackett, and J. K. Park Nitrosoureas Inhibit the Stathmin-Mediated Migration and Invasion of Malignant Glioma Cells Cancer Res., July 1, 2008; 68(13): 5267 - 5272. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||