Advertisement

Amplification of the STOML3, FREM2, and LHFP Genes Is Associated with Mesenchymal Differentiation in Gliosarcoma

      Gliosarcoma is a rare glioblastoma variant characterized by a biphasic tissue pattern with alternating areas that display either glial (glial fibrillary acidic protein–positive) or mesenchymal (reticulin-positive) differentiation. Previous analyses have shown identical genetic alterations in glial and mesenchymal tumor areas, suggesting that gliosarcomas are genetically monoclonal, and mesenchymal differentiation was considered to reflect the elevated genomic instability of glioblastomas. In the present study, we compared genome-wide chromosomal imbalances using array comparative genomic hybridization in glial and mesenchymal tumor areas of 13 gliosarcomas. The patterns of gain and loss were similar, except that the gain at 13q13.3-q14.1 (log2 ratio >3.0), containing the STOML3, FREM2, and LHFP genes, which was restricted to the mesenchymal tumor area of a gliosarcoma. Further analyses of 64 cases of gliosarcoma using quantitative PCR showed amplification of the STOML3, FREM2, and LHFP genes in 14 (22%), 10 (16%), and 7 (11%) mesenchymal tumor areas, respectively, but not in glial tumor areas. Results of IHC analysis confirmed that overexpression of STOML3 and FREM2 was more extensive in mesenchymal than in glial tumor areas. These results suggest that the mesenchymal components in a small fraction of gliosarcomas may be derived from glial cells with additional genetic alterations.
      Gliosarcoma is a rare variant of glioblastoma that constitutes approximately 2% of all glioblastomas.
      Histologically, these tumors are characterized by a biphasic tissue pattern, with alternating areas displaying glial [glial fibrillary acidic protein (GFAP)–positive] and mesenchymal (reticulin-positive) differentiation.
      Despite the presence of these two distinct types of differentiation, previous genetic analyses have shown that glial and mesenchymal tumor areas are usually genetically identical in terms of TP53 mutations, PTEN mutations, p16INK4a deletion, CDK4 amplification, and MDM2 amplification,
      • Biernat W.
      • Aguzzi A.
      • Sure U.
      • Grant J.W.
      • Kleihues P.
      • Hegi M.E.
      Identical mutations of the p53 tumor suppressor gene in the gliomatous and the sarcomatous components of gliosarcomas suggest a common origin from glial cells.
      • Reis R.M.
      • Konu-Lebleblicioglu D.
      • Lopes J.M.
      • Kleihues P.
      • Ohgaki H.
      Genetic profile of gliosarcomas.
      suggesting that gliosarcomas are genetically monoclonal. In two studies using conventional comparative genomic hybridization (CGH), although patterns of chromosomal imbalance were also largely similar at the genome-wide level in the glial and mesenchymal components, there were also gains and losses at several loci that were unique to either glial or mesenchymal tumor areas.
      • Actor B.
      • Cobbers J.M.
      • Buschges R.
      • Wolter M.
      • Knobbe C.B.
      • Reifenberger G.
      • Weber R.G.
      Comprehensive analysis of genomic alterations in gliosarcoma and its two tissue components.
      • Boerman R.H.
      • Anderl K.
      • Herath J.
      • Borell T.
      • Johnson N.
      • Schaeffer-Klein J.
      • Kirchhof A.
      • Raap A.K.
      • Scheithauer B.W.
      • Jenkins R.B.
      The glial and mesenchymal elements of gliosarcomas share similar genetic alterations.
      Therefore, it is unclear whether mesenchymal differentiation simply reflects the extensive genomic instability of glioblastomas, whether mesenchymal components are derived from glial cells with additional genetic alterations, or whether mesenchymal differentiation is caused by a mechanism similar to that involved in epithelial-mesenchymal transition (EMT) in epithelial neoplasms.
      • Mani S.A.
      • Guo W.
      • Liao M.J.
      • Eaton E.N.
      • Ayyanan A.
      • Zhou A.Y.
      • Brooks M.
      • Reinhard F.
      • Zhang C.C.
      • Shipitsin M.
      • Campbell L.L.
      • Polyak K.
      • Brisken C.
      • Yang J.
      • Weinberg R.A.
      The epithelial-mesenchymal transition generates cells with properties of stem cells.
      • Thiery J.P.
      Epithelial-mesenchymal transitions in tumour progression.
      In the present study, to explore the possibility that additional genetic alterations may lead to mesenchymal differentiation in gliosarcoma, we assessed genome-wide chromosomal imbalances in glial and mesenchymal tumor areas from the same gliosarcomas using array CGH.

      Materials and Methods

      Tumor Samples

      FFPE tissue samples from 64 cases of gliosarcomas and 10 primary (de novo) glioblastomas were obtained from the Department of Neuropathology, University Hospital Zurich, Zurich, Switzerland; the Institute of Neurology, University Hospital Frankfurt, Frankfurt, Germany; the Institute of Neuropathology, University Hospital Munster, Munster, Germany; the Institute of Neuroscience, Bordeaux, France; the Department of Neuropathology, University Hospital Rome, Rome, Italy; and the Department of Pathology, Gunma University, Gunma, Japan. The median ± SD age at histologic diagnosis of gliosarcomas was 59 ± 11 years (range, 32 to 82 years), and the sex ratio was 1.13 (male:female). Gliosarcomas from patients with recurrent disease were not included in this study.
      Gliosarcomas were diagnosed according to the 2007 World Health Organization classification.
      Histologically, tumors showed the typical biphasic pattern with alternating areas of glial and mesenchymal differentiation. The glial area was composed of anaplastic glial cells with GFAP expression. The mesenchymal component demonstrated bundles of spindle cells with malignant transformation and abundant connective tissue stained by reticulin, without GFAP expression.
      Survival data were also collected for 31 patients with gliosarcoma. Mean ± SD follow-up was 12.6 ± 8.5 months (range, 3.0 to 29.7 months), and 10 of 31 patients (32%) were still alive at the time this study was conducted.

      DNA Extraction

      We first selected 21 gliosarcomas in which mesenchymal and glial tumor areas were clearly recognized, and both areas were sufficiently large for manual microdissection to be practicable. DNA was separately extracted from these two tumor areas. To confirm that the areas had been correctly dissected, GFAP staining was performed on the same histologic sections from which tumor tissues were scraped for DNA extraction (Figure 1). Samples with a small amount of DNA (<1 μg) were amplified by whole genome amplification (WGA) using protocols we recently established.
      • Huang J.
      • Pang J.
      • Watanabe T.
      • Ng H.K.
      • Ohgaki H.
      Whole genome amplification for array comparative genomic hybridization using DNA extracted from formalin-fixed, paraffin-embedded histological sections.
      Eight cases were excluded from the study owing to lack of sufficient DNA even after WGA, and, therefore, 13 samples were available for array CGH (5 with WGA and 8 without WGA).
      Figure thumbnail gr1
      Figure 1A: GFAP IHC is positive in glial tumor areas and negative in mesenchymal tumor areas. B: DNA was extracted from glial (Glial) and mesenchymal (Mes) tumor areas separately. The scraped borders were displayed with dotted lines (before) and solid lines (after DNA extraction).
      DNA was extracted as described previously.
      • Huang J.
      • Pang J.
      • Watanabe T.
      • Ng H.K.
      • Ohgaki H.
      Whole genome amplification for array comparative genomic hybridization using DNA extracted from formalin-fixed, paraffin-embedded histological sections.
      Briefly, tumor samples scraped from histologic slides were deparaffinized in xylene and rehydrated in ethanol. After overnight incubation in 1 mol/L sodium thiocyanate solution, samples were suspended in DNA extraction buffer composed of ATL buffer and proteinase K (DNeasy Mini kit; Qiagen, Valencia, CA) and were incubated for 60 hours. The samples were then incubated with RNase for 10 minutes, and ATL buffer was added. After incubation with a mixture of 450 μL of ATL buffer and 450 μL of 100% ethanol for 5 minutes, the samples were loaded onto DNeasy Mini spin columns (Qiagen). After washing with buffer AW1, purified genomic DNA was eluted with 21 μL of nuclease-free H2O. DNA concentrations were determined by spectrophotometer (NanoDrop Technologies, Wilmington, DE). Absorption was measured at 230, 260, and 280 nm, and the DNA quality was evaluated by A260/A230 and A260/A280 ratios.

      Whole Genome Amplification

      WGA was performed using the REPLI-g FFPE kit (Qiagen) as the reference.
      • Huang J.
      • Pang J.
      • Watanabe T.
      • Ng H.K.
      • Ohgaki H.
      Whole genome amplification for array comparative genomic hybridization using DNA extracted from formalin-fixed, paraffin-embedded histological sections.
      Briefly, purified genomic DNA in a total volume of 10 μL was heated to 95°C for 5 minutes for denaturation. After cooling the samples on ice for 5 minutes, 8 μL of FFPE buffer, 1 μL of ligation enzyme, and 1 μL of FFPE enzyme were added, and the samples were incubated at 24°C for 30 minutes, followed by heat inactivation at 95°C for 5 minutes. Thirty microliters of the reaction mix (29 μL of reaction buffer and 1 μL of Midi Phi29 DNA polymerase) was added to the denatured DNA to a total volume of 50 μL. The mix was incubated at 30°C for 1 hour. After amplification, the Phi29 enzyme was inactivated by heating at 95°C for 10 minutes.

      Array CGH

      Samples after WGA were purified using the NucleoTraP CR kit (Macherey-Nagel, Düren, Germany) before DNA labeling, as described previously.
      • Huang J.
      • Pang J.
      • Watanabe T.
      • Ng H.K.
      • Ohgaki H.
      Whole genome amplification for array comparative genomic hybridization using DNA extracted from formalin-fixed, paraffin-embedded histological sections.
      • Nobusawa S.
      • Lachuer J.
      • Wierinckx A.
      • Kim Y.H.
      • Huang J.
      • Legras C.
      • Kleihues P.
      • Ohgaki H.
      Intratumoral patterns of genomic imbalance in glioblastomas.
      Genome-wide chromosomal imbalance was assessed using a CGH oligonucleotide microarray (105K; Agilent Technologies, Santa Clara, CA; 15.0 Kb average probe resolution) according to the manufacturer's instructions. Briefly, the sample (1 μg) and the sex-matched reference DNA were chemically labeled with ULS-Cy5 and ULS-Cy3, respectively, at 85°C for 30 minutes using an oligonucleotide array CGH labeling kit for FFPE samples (Agilent Technologies). The labeled samples were purified with the genomic DNA purification module (Agilent Technologies), combined, mixed with human Cot-1 DNA, denatured at 95°C using an oligonucleotide array CGH hybridization kit (Agilent Technologies), and applied to microarrays. After hybridization at 65°C for 40 hours, microarrays were washed in oligonucleotide array CGH wash buffer 1 at room temperature for 5 minutes and in wash buffer 2 at 37°C for 1 minute. After drying, the microarrays were scanned by a DNA microarray scanner (G2565BA; Agilent Technologies), and data (log2) were extracted from the raw microarray image files using Feature Extraction software version 9 (Agilent Technologies). Data were analyzed by DNA Analytics software version 3.5 (Agilent Technologies) using default filter settings. The aberration detection method 2 algorithm with fuzzy zero correction was used to define aberrant intervals.

      Amplification of the STOML3, FREM2, and LHFP Genes

      Amplification of the STOML3, FREM2, and LHFP genes was assessed by quantitative PCR using the CF sequence as a reference in pairs of glial and mesenchymal tumor areas from 64 gliosarcomas. Primer sequences were as follows: 5′-TCACCAGAGACTCCGTAACT-3′ (sense) and 5′-AGAAATGTTGCTTGATGGAC-3′ (antisense) for STOML3 (PCR product, 109 bp), 5′-TCCAACCTCCTGGATTATAC-3′ (sense) and 5′-GACAAGCTGTACTGGTAAGGA-3′ (antisense) for FREM2 (PCR product, 110 bp), and 5′-CCTGTGCATGATGAGAGTC-3′ (sense) and 5′-GTCACTATGGTGCAGATCCT-3′ (antisense) for LHFP (PCR product, 110 bp), and 5′-GGCACCATTAAAGAAAATATCATC TT-3′ (sense) and 5′-GTTGGCATGCTTTGATGACGCTTC-3′ (antisense) for the CF (PCR product, 79 bp). PCR reactions were performed in a total volume of 20 μL, with 10 μL of iQ SYBR green (Bio-Rad Laboratories, Hercules, CA), 6.4 μL of primer sets (1.25 μmol/L of each primer), and 20 ng of DNA, with cycling parameters as reported previously.
      • Nigro J.M.
      • Takahashi M.A.
      • Ginzinger D.G.
      • Law M.
      • Passe S.
      • Jenkins R.B.
      • Aldape K.
      Detection of 1p and 19q loss in oligodendroglioma by quantitative microsatellite analysis, a real-time quantitative polymerase chain reaction assay.
      PCR was performed in triplicate on a 96-well optical plate using an iCycler iQ5 detection system (Bio-Rad Laboratories). The copy number calculation was performed using the comparative CT method, as described previously.
      • Nigro J.M.
      • Takahashi M.A.
      • Ginzinger D.G.
      • Law M.
      • Passe S.
      • Jenkins R.B.
      • Aldape K.
      Detection of 1p and 19q loss in oligodendroglioma by quantitative microsatellite analysis, a real-time quantitative polymerase chain reaction assay.
      • Biernat W.
      • Huang H.
      • Yokoo H.
      • Kleihues P.
      • Ohgaki H.
      Predominant expression of mutant EGFR (EGFRvIII) is rare in primary glioblastomas.
      To calculate the average δCT [δCT (normal)], DNA was isolated from 14 FFPE normal tissues. The gene copy numbers of the samples are calculated by the following formula: δCT = [CT (target) – CT (reference)] and δδCT = [δCt(tumor) – δCT (normal)]. The relative gene copy numbers are calculated by the expression 2 × 2–δδCt. Using this method, a δδCt ratio, 2 × 2–δδCt more than 2.88 was considered to indicate amplification.

      IHC Analysis

      Immunohistochemical (IHC) analysis was performed for 64 gliosarcomas and 10 primary (de novo) glioblastomas using the UltraVision Quanto detection system (Thermo Fisher Scientific Inc., Fremont, CA) according to the manufacturer's protocol. STOML3 antibody (1:250 dilution) and FREM2 antibody (1:100) were purchased from Atlas Antibodies (Stockholm, Sweden). Immunoreactivity was evaluated separately in mesenchymal and glial tumor areas of gliosarcomas and was scored as follows: – indicates <10%; +, 10% to 50%; ++, 51% to 90%; and +++, >90% of neoplastic cells with immunoreactivity. LHFP antibody for IHC on FFPE sections was not available.

      Statistical Analyses

      The Student's t-test was performed to analyze differences in mean ages between patients with tumors with or without genetic alterations. Fisher's exact test was used to assess group differences in the analysis of qualitative features in IHC analysis data. Pearson's correlation was used for analysis of the correlation between two variables. The Mann-Whitney nonparametric test was used for numerical variables. The Kaplan-Meier curve with the log-rank test was used for survival analysis. P < 0.05 was considered statistically significant.

      Results

      Array CGH

      The overall pattern of chromosomal imbalance in array CGH analysis was largely similar in glial and mesenchymal tumor areas (Figure 2). The imbalance commonly detected in >10-Mb chromosomal regions included gain at 7p22.3-q36.3 (8 of 13; 62%) and 20p13-q13.33 (3 of 13; 23%) and loss at 8q24.3 (3 of 13; 23%), 10p15.3-q26.3 (7 of 13; 54%), 11p15.5-q25 (4 of 13; 31%), and 13q11-q34 (6 of 13; 46%). Loss of several loci containing well-characterized tumor-associated genes was observed at 9p21.3-p21.2 (CDKN2A, CDKN2B) (3 of 13; 23%), 10p15.3-q26.3 (PTEN) (7 of 13; 54%), and 13q14.2 (RB1) (1 of 13; 8%).
      Figure thumbnail gr2
      Figure 2Genome-wide chromosomal imbalance (log2 ratio >1.0) in glial (A) and mesenchymal (B) tumor areas in gliosarcomas (n = 13). Note that the overall pattern of chromosomal imbalance is largely similar in glial and mesenchymal tumor areas.
      We then focused on gain (log2 ratio >1.0) in mesenchymal but not in glial tumor areas. Direct comparison between glial and mesenchymal components of the same tumors showed the highest gain at 13q13.3-q14.1 (log2 ratio >3.0) containing the STOML3, FREM2, COG6, C13orf23, LOC387921, LHFP, and UFM1 loci uniquely in the mesenchymal tumor area of a single gliosarcoma (Figure 3). Amplification of the STOML3, FREM2, and LHFP genes at 13q13.3 was confirmed by quantitative PCR in mesenchymal but not in glial tumor areas of this gliosarcoma.
      Figure thumbnail gr3
      Figure 3Array CGH showing gain at 13q13.3-q14.1 in a mesenchymal tumor area (red line) but not in a glial tumor area (blue line) of a gliosarcoma.

      Amplification of STOML3, FREM2, and LHFP Is Characteristic of Mesenchymal Tumor Areas of Gliosarcomas

      We examined 64 gliosarcomas (57 tumors clearly containing both glial and mesenchymal tumor areas and 7 tumors largely presenting a mesenchymal phenotype) for amplification of the STOML3, FREM2, and LHFP genes. STOML3 amplification was restricted to mesenchymal tumor areas (14 of 64 mesenchymal tumor areas; 22%), and no glial tumor areas showed STOML3 amplification (0 of 57, 0%; P < 0.0001) (Figure 4A). FREM2 amplification was observed in mesenchymal tumor areas only in 10 of 64 cases (16%) but in none of the 57 glial tumor areas analyzed (P = 0.002). Similarly, LHFP amplification was seen in mesenchymal tumor areas only in 7 of 64 cases (11%), whereas no glial tumor areas revealed LHFP amplification (0 of 57, 0%; P = 0.014). Furthermore, direct comparison of 2–δδCt (mesenchymal area)/2–δδCt (glial area) values for cases showing amplification or nonamplification in mesenchymal areas showed significant differences for STOML3 (mean, 3.58/1.02 for amplification/nonamplification; P < 0.0001), FREM2 (mean, 1.86/1.01 for amplification/nonamplification; P < 0.0001), and LHFP (mean, 1.47/0.97 for amplification/nonamplification; P = 0.007). There was no significant difference in age at diagnosis in patients with gliosarcoma with or without amplification of these genes. Amplification of the STOML3, FREM2, or LHFP genes was not detected in any of 10 ordinary glioblastomas (not gliosarcoma) analyzed.
      Figure thumbnail gr4
      Figure 4STOML3 IHC in gliosarcomas. A: Note that STOML3 immunoreactivity is positive in mesenchymal tumor areas (Mes) and negative in glial tumor areas (Glial). B: Higher magnification of a mesenchymal tumor area of gliosarcoma showing nuclear and cytoplasmic staining for STOML3. Original magnification, ×400. C: Higher magnification of a glial tumor area showing that STOML3 immunoreactivity is largely negative. Original magnification, ×400.

      STOML3 and FREM2 IHC

      Cytoplasmic and nuclear STOML3 expression was detected by IHC in >50% of neoplastic cells in 25 of 64 gliosarcomas (39%) analyzed. STOML3 expression was more extensive in mesenchymal than in glial tumor areas except for 2 cases (Figure 4 and Table 1). STOML3 expression was significantly correlated with STOML3 gene amplification in gliosarcomas (0.34; 95% CI, 0.11 to 0.55; P = 0.006; Figure 5).
      Table 1Amplification and Overexpression of STOML3, FREM2, and LHFP in Glial and Mesenchymal Tumor Areas of Gliosarcomas
      Case no.STOML3FREM2LHFP
      Amplification (Glial/Mes)IHC (Glial/Mes)Amplification (Glial/Mes)IHC (Glial/Mes)Amplification (Glial/Mes)
      1−/−−/−−/−−/−−/−
      2−/−−/−−/−−/−−/−
      3−/amp−/++−/amp−/−−/amp
      4−/amp−/−−/amp−/−−/−
      5−/−−/−−/−−/+−/−
      6−/−−/−−/−−/−−/−
      7−/−−/−−/−+/−−/−
      8−/−−/−−/−−/+−/−
      9−/−−/−−/−−/−−/−
      10−/amp+/++−/−−/+−/−
      11−/−+++/+−/−−/−−/−
      12−/amp−/+++−/amp−/+++−/−
      13−/amp++/+++−/−−/+++−/−
      14−/−−/−−/−−/−−/−
      15−/−−/+++−/−−/+−/−
      16−/−+/+−/−−/−−/−
      17−/−+/+++−/amp++/++−/amp
      18−/−−/−−/−+/−−/−
      19−/−−/−−/−++/−−/−
      20−/−−/+++−/amp+/++−/−
      21−/−+/+−/−+/++−/−
      22ND/−ND/+ND/−ND/−ND
      23−/−−/−−/−−/−−/−
      24−/−++/++−/−−/−−/−
      25−/−−/−−/−−/++−/−
      26−/−−/+++−/amp+/++−/amp
      27−/−−/−−/−−/+−/−
      28−/−−/−−/−−/−−/−
      29−/−−/−−/−−/−−/−
      30−/−−/+−/−−/−−/−
      31−/−−/−−/−−/−−/−
      32−/−−/+−/−−/−−/−
      33−/−−/+++−/−+/+++−/−
      34−/amp++/++−/amp−/+−/−
      35−/amp−/++−/−−/++−/−
      36−/−+/+−/−+/+−/−
      37−/−+/++−/−+/−−/−
      38−/−−/−−/−−/−−/−
      39−/−+/++−/−−/+−/−
      40ND/ampND/++ND/−ND/++ND
      41ND/−ND/+ND/−ND/−ND
      42−/−+/+++−/−−/−−/−
      43−/−+/−−/−+/+−/−
      44−/−−/−−/−−/++−/−
      45−/−−/−−/−−/−−/−
      46−/−−/−−/−−/−−/−
      47ND/ampND/++ND/−ND/+++ND/amp
      48−/−−/−−/−−/−−/amp
      49−/amp+/++−/−−/+++−/−
      50−/−−/−−/−−/+−/−
      51−/−++/++−/−−/−−/−
      52ND/−ND/+ND/−ND/−ND
      53−/−−/−−/−−/−−/−
      54−/−−/+++−/−−/+−/−
      55−/amp−/+++−/−+/++−/−
      56−/amp+/++−/−+/++−/−
      57−/−−/−−/−+/+−/−
      58−/−−/−−/−−/−−/−
      59ND/−ND/−ND/−ND/−ND
      60−/−−/++−/−−/+−/−
      61−/−−/−−/amp−/−−/−
      62−/amp++/+++−/amp+/+++−/amp
      63−/−−/−−/−+/+−/−
      64ND/ampND/++ND/ampND/−ND/amp
      The fraction of STOML3− or FREM2−positive cells in IHC was evaluated as follows: − indicates less than 10%; +, 10% to 50%; ++, 51% to 90%; and +++, greater than 90% of positive cells.
      amp, amplification; Glial, glial tumor area; Mes, mesenchymal tumor area; ND, not determined.
      Figure thumbnail gr5
      Figure 5Correlation between amplification and overexpression of the STOML3 gene in mesenchymal tumor areas of gliosarcomas. The boxes represent interquartile range, and whiskers represent maximal points that are not outliers, which are shown as the white circle. The bold line represents the median. Overexpression was assessed by IHC and was scored as follows: – indicates <10%; +, 10% to 50%; ++, 51% to 90%; and +++, >90% of neoplastic cells with immunoreactivity.
      Cytoplasmic FREM2 expression was detected by IHC analysis in more than 50% of neoplastic cells in 16 of 64 gliosarcomas (25%) analyzed. FREM2 expression was more extensive in mesenchymal than in glial tumor areas except in 4 cases (Table 1). There was a non–statistically significant correlation between FREM2 amplification and expression.
      We also performed STOML3 and FREM2 IHC in 10 ordinary glioblastomas. Only one glioblastoma showed immunoreactivity to STOML3 (70% of neoplastic cells) and one to FREM2 (20% of neoplastic cells).

      STOML3, FREM2, or LHFP Overexpression/Amplification and Patient Survival

      Overexpression of the STOML3, FREM2, or LHFP genes was not prognostic of survival in patients with gliosarcoma (data not shown). We compared survival of 138 patients with glioblastoma with and without STOML3 expression in The Cancer Genome Atlas database (http://www.cbioportal.org/public-portal, last accessed August 9, 2011).
      Cancer Genome Atlas Research Network
      Comprehensive genomic characterization defines human glioblastoma genes and core pathways.
      Mean survival in patients with glioblastoma with STOML3 overexpression was 5.9 months (95% CI, 3.35 to NA months), which was significantly shorter than that in patients without STOML3 overexpression (14.3 months; 95% CI, 12.6 to 18.0 months; P = 0.015).

      Discussion

      EMT is a critical step during embryogenesis and is also considered to play a significant role in cancer invasion and metastasis.
      • Greenburg G.
      • Hay E.D.
      Epithelia suspended in collagen gels can lose polarity and express characteristics of migrating mesenchymal cells.
      • Stoker M.
      • Perryman M.
      An epithelial scatter factor released by embryo fibroblasts.
      The migratory and invasive phenotype and stem cell characteristics of cancer cells may result from EMT.
      • Kurrey N.K.
      • Jalgaonkar S.P.
      • Joglekar A.V.
      • Ghanate A.D.
      • Chaskar P.D.
      • Doiphode R.Y.
      • Bapat S.A.
      Snail and slug mediate radioresistance and chemoresistance by antagonizing p53-mediated apoptosis and acquiring a stem-like phenotype in ovarian cancer cells.
      Several signaling pathways and molecules seem to be involved in EMT, including transforming growth factor β and RTK/Ras signaling, autocrine factors and Wnt, Notch, Hedgehog, and nuclear factor κB–dependent pathways.
      • Kalluri R.
      • Weinberg R.A.
      The basics of epithelial-mesenchymal transition.
      • Ouyang G.
      • Wang Z.
      • Fang X.
      • Liu J.
      • Yang C.J.
      Molecular signaling of the epithelial to mesenchymal transition in generating and maintaining cancer stem cells.
      Loss of E-cadherin is a critical step for EMT, leading to the breakdown of cell-cell adhesion and acquisition of invasive growth properties in cancer cells.
      • Peinado H.
      • Portillo F.
      • Cano A.
      Transcriptional regulation of cadherins during development and carcinogenesis.
      Down-regulation of E-cadherin may be caused by several transcription factors, such as Twist, dEF1/Zfh1 family (Zeb1, SIP1), and Snail family (SNAI1/Snail, SNAI2/Slug),
      • Bolos V.
      • Peinado H.
      • Perez-Moreno M.A.
      • Fraga M.F.
      • Esteller M.
      • Cano A.
      The transcription factor Slug represses E-cadherin expression and induces epithelial to mesenchymal transitions: a comparison with Snail and E47 repressors.
      • Thiery J.P.
      • Sleeman J.P.
      Complex networks orchestrate epithelial-mesenchymal transitions.
      • Martin T.A.
      • Goyal A.
      • Watkins G.
      • Jiang W.G.
      Expression of the transcription factors snail, slug, and twist and their clinical significance in human breast cancer.
      • Karreth F.
      • Tuveson D.A.
      Twist induces an epithelial-mesenchymal transition to facilitate tumor metastasis.
      or gene mutations.
      • Peinado H.
      • Portillo F.
      • Cano A.
      Transcriptional regulation of cadherins during development and carcinogenesis.
      Little is known as to whether similar mechanisms operate in glioma. It has been shown that Twist1 promotes invasion through mesenchymal changes in glioblastoma,
      • Mikheeva S.A.
      • Mikheev A.M.
      • Petit A.
      • Beyer R.
      • Oxford R.G.
      • Khorasani L.
      • Maxwell J.P.
      • Glackin C.A.
      • Wakimoto H.
      • Gonzalez-Herrero I.
      • Sanchez-Garcia I.
      • Silber J.R.
      • Horner P.J.
      • Rostomily R.C.
      TWIST1 promotes invasion through mesenchymal change in human glioblastoma.
      and SNAI2/Slug promotes growth and invasion in glioma,
      • Yang H.W.
      • Menon L.G.
      • Black P.M.
      • Carroll R.S.
      • Johnson M.D.
      SNAI2/Slug promotes growth and invasion in human gliomas.
      suggesting that glial to mesenchymal transitions may play a role in invasion in gliomas.
      The present study is the first array CGH analysis to separately assess genome-wide chromosomal imbalance in glial and mesenchymal tumor areas in the same gliosarcomas. In line with the results of previous conventional CGH and genetic analyses in several selected genes, we showed that genome-wide chromosomal imbalance was similar between glial and mesenchymal tumor areas of gliosarcomas.
      • Biernat W.
      • Aguzzi A.
      • Sure U.
      • Grant J.W.
      • Kleihues P.
      • Hegi M.E.
      Identical mutations of the p53 tumor suppressor gene in the gliomatous and the sarcomatous components of gliosarcomas suggest a common origin from glial cells.
      • Reis R.M.
      • Konu-Lebleblicioglu D.
      • Lopes J.M.
      • Kleihues P.
      • Ohgaki H.
      Genetic profile of gliosarcomas.
      • Actor B.
      • Cobbers J.M.
      • Buschges R.
      • Wolter M.
      • Knobbe C.B.
      • Reifenberger G.
      • Weber R.G.
      Comprehensive analysis of genomic alterations in gliosarcoma and its two tissue components.
      • Boerman R.H.
      • Anderl K.
      • Herath J.
      • Borell T.
      • Johnson N.
      • Schaeffer-Klein J.
      • Kirchhof A.
      • Raap A.K.
      • Scheithauer B.W.
      • Jenkins R.B.
      The glial and mesenchymal elements of gliosarcomas share similar genetic alterations.
      In the present study, we focused on genes at 13q13.3-q14.1, where significant gain was restricted to mesenchymal tumor areas only (not glial tumor areas) in 1 of 13 gliosarcomas analyzed. Several known genes are situated at this locus, including STOML3, FREM2, COG6, C13orf23, LOC387921, LHFP, and UFM1. We selected STOML3, FREM2, and LHFP genes for further analyses based on the level of gain and their known functions.
      We found amplification of the STOML3, FREM2, and LHFP genes in 22%, 16%, and 11%, respectively, of 64 gliosarcomas analyzed. Amplification of these genes was restricted to mesenchymal tumor areas only and was absent in glial tumor areas in any of the gliosarcomas analyzed. Mesenchymal tumor area–specific amplification of the STOML3 and FREM2 genes was associated with overexpression as confirmed by IHC analysis. These results suggest that mesenchymal differentiation in a small fraction of gliosarcomas may be due to additional genetic alterations.
      STOML3 encodes a protein related to stomatin that regulates the activity of ion channels.
      • Tan P.L.
      • Barr T.
      • Inglis P.N.
      • Mitsuma N.
      • Huang S.M.
      • Garcia-Gonzalez M.A.
      • Bradley B.A.
      • Coforio S.
      • Albrecht P.J.
      • Watnick T.
      • Germino G.G.
      • Beales P.L.
      • Caterina M.J.
      • Leroux M.R.
      • Rice F.L.
      • Katsanis N.
      Loss of Bardet Biedl syndrome proteins causes defects in peripheral sensory innervation and function.
      It is essential for touch sensation of mouse skin.
      • Wetzel C.
      • Hu J.
      • Riethmacher D.
      • Benckendorff A.
      • Harder L.
      • Eilers A.
      • Moshourab R.
      • Kozlenkov A.
      • Labuz D.
      • Caspani O.
      • Erdmann B.
      • Machelska H.
      • Heppenstall P.A.
      • Lewin G.R.
      A stomatin-domain protein essential for touch sensation in the mouse.
      • Kadurin I.
      • Huber S.
      • Grunder S.
      A single conserved proline residue determines the membrane topology of stomatin.
      The STOML3 protein is expressed in receptor neurons of the olfactory epithelium, suggesting that it has some function in olfactory neurons.
      • Goldstein B.J.
      • Kulaga H.M.
      • Reed R.R.
      Cloning and characterization of SLP3: a novel member of the stomatin family expressed by olfactory receptor neurons.
      LHFP is a frequent translocation partner with HMGIC (at 12q15) in lipomas.
      • Petit M.M.
      • Schoenmakers E.F.
      • Huysmans C.
      • Geurts J.M.
      • Mandahl N.
      • Van de Ven W.J.
      LHFP, a novel translocation partner gene of HMGIC in a lipoma, is a member of a new family of LHFP-like genes.
      • Nilsson M.
      • Mertens F.
      • Hoglund M.
      • Mandahl N.
      • Panagopoulos I.
      Truncation and fusion of HMGA2 in lipomas with rearrangements of 5q32–>q33 and 12q14–>q15.
      The Fras1/Frem gene family, including the FREM2 gene, encodes extracellular matrix proteins that are localized in epithelial basement membranes in eyelids, limbs, kidneys, lungs, gastrointestinal tract, and the central nervous system.
      • Chiotaki R.
      • Petrou P.
      • Giakoumaki E.
      • Pavlakis E.
      • Sitaru C.
      • Chalepakis G.
      Spatiotemporal distribution of Fras1/Frem proteins during mouse embryonic development.
      • Pavlakis E.
      • Chalepakis G.
      pH-dependent antigen unmasking in paraformaldehyde-fixed tissue cryosections.
      The Fras1/Frem gene family plays critical roles in epithelial-mesenchymal interaction during embryonic development.
      • Petrou P.
      • Makrygiannis A.K.
      • Chalepakis G.
      The Fras1/Frem family of extracellular matrix proteins: structure, function, and association with Fraser syndrome and the mouse bleb phenotype.
      Timmer et al
      • Timmer J.R.
      • Mak T.W.
      • Manova K.
      • Anderson K.V.
      • Niswander L.
      Tissue morphogenesis and vascular stability require the Frem2 protein, product of the mouse myelencephalic blebs gene.
      showed that loss of FREM2 function results in defects in developmental events associated with morphogenetic rearrangements of the vasculature and of tissues arising from all germ layers in mice. FREM2 mutations are known to cause epidermal adhesion defects
      • Petrou P.
      • Makrygiannis A.K.
      • Chalepakis G.
      The Fras1/Frem family of extracellular matrix proteins: structure, function, and association with Fraser syndrome and the mouse bleb phenotype.
      and are associated with Fraser syndrome, which is not linked to Fras1.
      • Petrou P.
      • Makrygiannis A.K.
      • Chalepakis G.
      The Fras1/Frem family of extracellular matrix proteins: structure, function, and association with Fraser syndrome and the mouse bleb phenotype.
      • Jadeja S.
      • Smyth I.
      • Pitera J.E.
      • Taylor M.S.
      • van H.M.
      • Bentley E.
      • McGregor L.
      • Hopkins J.
      • Chalepakis G.
      • Philip N.
      • Perez A.A.
      • Watt F.M.
      • Darling S.M.
      • Jackson I.
      • Woolf A.S.
      • Scambler P.J.
      Identification of a new gene mutated in Fraser syndrome and mouse myelencephalic blebs.
      In The Cancer Genome Atlas data (http://tcga-portal.nci.nih.gov/tcga-portal/AnomalySearch.jsp, last accessed August 9, 2011),
      Cancer Genome Atlas Research Network
      Comprehensive genomic characterization defines human glioblastoma genes and core pathways.
      FREM2 gain was found in 1 of 372 (0.3%) glioblastomas; LHFP gain and overexpression were found in 1 of 372 (0.3%) and 109 of 424 (26%) glioblastomas, respectively; and STOML3 gain and overexpression were found in 1 of 372 (0.3%) and 4 of 138 (3%) glioblastomas, respectively, and STOML3 overexpression was associated with poor prognosis in patients with glioblastoma. Thus, overexpression and amplification of the STOML3, FREM2, and LHFP genes seem to be rare in ordinary glioblastomas.
      In summary, amplifications of the STOML3, FREM2, and LHFP genes are additional genetic alterations associated with the mesenchymal phenotype in a small fraction of gliosarcomas. It remains to be shown whether these genes or other genes co-amplified at 13q13.3-q14.1 are directly associated with glial-mesenchymal transition in glioblastomas. The relatively low frequency of chromosomal copy number abnormalities specific to mesenchymal tumor areas also suggests that other mechanisms, such as epigenetic regulation, may be operative in mesenchymal differentiation in gliosarcomas.

      Acknowledgment

      We thank Christine Carreira for technical assistance.

      References

      1. Louis D.N. Ohgaki H. Wiestler O.D. Cavenee W.K. WHO Classification of Tumours of the Central Nervous System. IARC, Lyon, France2007
        • Biernat W.
        • Aguzzi A.
        • Sure U.
        • Grant J.W.
        • Kleihues P.
        • Hegi M.E.
        Identical mutations of the p53 tumor suppressor gene in the gliomatous and the sarcomatous components of gliosarcomas suggest a common origin from glial cells.
        J Neuropathol Exp Neurol. 1995; 54: 651-656
        • Reis R.M.
        • Konu-Lebleblicioglu D.
        • Lopes J.M.
        • Kleihues P.
        • Ohgaki H.
        Genetic profile of gliosarcomas.
        Am J Pathol. 2000; 156: 425-432
        • Actor B.
        • Cobbers J.M.
        • Buschges R.
        • Wolter M.
        • Knobbe C.B.
        • Reifenberger G.
        • Weber R.G.
        Comprehensive analysis of genomic alterations in gliosarcoma and its two tissue components.
        Genes Chromosomes Cancer. 2002; 34: 416-427
        • Boerman R.H.
        • Anderl K.
        • Herath J.
        • Borell T.
        • Johnson N.
        • Schaeffer-Klein J.
        • Kirchhof A.
        • Raap A.K.
        • Scheithauer B.W.
        • Jenkins R.B.
        The glial and mesenchymal elements of gliosarcomas share similar genetic alterations.
        J Neuropathol Exp Neurol. 1996; 55: 973-981
        • Mani S.A.
        • Guo W.
        • Liao M.J.
        • Eaton E.N.
        • Ayyanan A.
        • Zhou A.Y.
        • Brooks M.
        • Reinhard F.
        • Zhang C.C.
        • Shipitsin M.
        • Campbell L.L.
        • Polyak K.
        • Brisken C.
        • Yang J.
        • Weinberg R.A.
        The epithelial-mesenchymal transition generates cells with properties of stem cells.
        Cell. 2008; 133: 704-715
        • Thiery J.P.
        Epithelial-mesenchymal transitions in tumour progression.
        Nat Rev Cancer. 2002; 2: 442-454
        • Huang J.
        • Pang J.
        • Watanabe T.
        • Ng H.K.
        • Ohgaki H.
        Whole genome amplification for array comparative genomic hybridization using DNA extracted from formalin-fixed, paraffin-embedded histological sections.
        J Mol Diagn. 2009; 11: 109-116
        • Nobusawa S.
        • Lachuer J.
        • Wierinckx A.
        • Kim Y.H.
        • Huang J.
        • Legras C.
        • Kleihues P.
        • Ohgaki H.
        Intratumoral patterns of genomic imbalance in glioblastomas.
        Brain Pathol. 2010; 20: 936-944
        • Nigro J.M.
        • Takahashi M.A.
        • Ginzinger D.G.
        • Law M.
        • Passe S.
        • Jenkins R.B.
        • Aldape K.
        Detection of 1p and 19q loss in oligodendroglioma by quantitative microsatellite analysis, a real-time quantitative polymerase chain reaction assay.
        Am J Pathol. 2001; 158: 1253-1262
        • Biernat W.
        • Huang H.
        • Yokoo H.
        • Kleihues P.
        • Ohgaki H.
        Predominant expression of mutant EGFR (EGFRvIII) is rare in primary glioblastomas.
        Brain Pathol. 2004; 14: 131-136
        • Cancer Genome Atlas Research Network
        Comprehensive genomic characterization defines human glioblastoma genes and core pathways.
        Nature. 2008; 455: 1061-1068
        • Greenburg G.
        • Hay E.D.
        Epithelia suspended in collagen gels can lose polarity and express characteristics of migrating mesenchymal cells.
        J Cell Biol. 1982; 95: 333-339
        • Stoker M.
        • Perryman M.
        An epithelial scatter factor released by embryo fibroblasts.
        J Cell Sci. 1985; 77: 209-223
        • Kurrey N.K.
        • Jalgaonkar S.P.
        • Joglekar A.V.
        • Ghanate A.D.
        • Chaskar P.D.
        • Doiphode R.Y.
        • Bapat S.A.
        Snail and slug mediate radioresistance and chemoresistance by antagonizing p53-mediated apoptosis and acquiring a stem-like phenotype in ovarian cancer cells.
        Stem Cells. 2009; 27: 2059-2068
        • Kalluri R.
        • Weinberg R.A.
        The basics of epithelial-mesenchymal transition.
        J Clin Invest. 2009; 119: 1420-1428
        • Ouyang G.
        • Wang Z.
        • Fang X.
        • Liu J.
        • Yang C.J.
        Molecular signaling of the epithelial to mesenchymal transition in generating and maintaining cancer stem cells.
        Cell Mol Life Sci. 2010; 67: 2605-2618
        • Peinado H.
        • Portillo F.
        • Cano A.
        Transcriptional regulation of cadherins during development and carcinogenesis.
        Int J Dev Biol. 2004; 48: 365-375
        • Bolos V.
        • Peinado H.
        • Perez-Moreno M.A.
        • Fraga M.F.
        • Esteller M.
        • Cano A.
        The transcription factor Slug represses E-cadherin expression and induces epithelial to mesenchymal transitions: a comparison with Snail and E47 repressors.
        J Cell Sci. 2003; 116: 499-511
        • Thiery J.P.
        • Sleeman J.P.
        Complex networks orchestrate epithelial-mesenchymal transitions.
        Nat Rev Mol Cell Biol. 2006; 7: 131-142
        • Martin T.A.
        • Goyal A.
        • Watkins G.
        • Jiang W.G.
        Expression of the transcription factors snail, slug, and twist and their clinical significance in human breast cancer.
        Ann Surg Oncol. 2005; 12: 488-496
        • Karreth F.
        • Tuveson D.A.
        Twist induces an epithelial-mesenchymal transition to facilitate tumor metastasis.
        Cancer Biol Ther. 2004; 3: 1058-1059
        • Mikheeva S.A.
        • Mikheev A.M.
        • Petit A.
        • Beyer R.
        • Oxford R.G.
        • Khorasani L.
        • Maxwell J.P.
        • Glackin C.A.
        • Wakimoto H.
        • Gonzalez-Herrero I.
        • Sanchez-Garcia I.
        • Silber J.R.
        • Horner P.J.
        • Rostomily R.C.
        TWIST1 promotes invasion through mesenchymal change in human glioblastoma.
        Mol Cancer. 2010; 9: 194
        • Yang H.W.
        • Menon L.G.
        • Black P.M.
        • Carroll R.S.
        • Johnson M.D.
        SNAI2/Slug promotes growth and invasion in human gliomas.
        BMC Cancer. 2010; 10: 301
        • Tan P.L.
        • Barr T.
        • Inglis P.N.
        • Mitsuma N.
        • Huang S.M.
        • Garcia-Gonzalez M.A.
        • Bradley B.A.
        • Coforio S.
        • Albrecht P.J.
        • Watnick T.
        • Germino G.G.
        • Beales P.L.
        • Caterina M.J.
        • Leroux M.R.
        • Rice F.L.
        • Katsanis N.
        Loss of Bardet Biedl syndrome proteins causes defects in peripheral sensory innervation and function.
        Proc Natl Acad Sci U S A. 2007; 104: 17524-17529
        • Wetzel C.
        • Hu J.
        • Riethmacher D.
        • Benckendorff A.
        • Harder L.
        • Eilers A.
        • Moshourab R.
        • Kozlenkov A.
        • Labuz D.
        • Caspani O.
        • Erdmann B.
        • Machelska H.
        • Heppenstall P.A.
        • Lewin G.R.
        A stomatin-domain protein essential for touch sensation in the mouse.
        Nature. 2007; 445: 206-209
        • Kadurin I.
        • Huber S.
        • Grunder S.
        A single conserved proline residue determines the membrane topology of stomatin.
        Biochem J. 2009; 418: 587-594
        • Goldstein B.J.
        • Kulaga H.M.
        • Reed R.R.
        Cloning and characterization of SLP3: a novel member of the stomatin family expressed by olfactory receptor neurons.
        J Assoc Res Otolaryngol. 2003; 4: 74-82
        • Petit M.M.
        • Schoenmakers E.F.
        • Huysmans C.
        • Geurts J.M.
        • Mandahl N.
        • Van de Ven W.J.
        LHFP, a novel translocation partner gene of HMGIC in a lipoma, is a member of a new family of LHFP-like genes.
        Genomics. 1999; 57: 438-441
        • Nilsson M.
        • Mertens F.
        • Hoglund M.
        • Mandahl N.
        • Panagopoulos I.
        Truncation and fusion of HMGA2 in lipomas with rearrangements of 5q32–>q33 and 12q14–>q15.
        Cytogenet Genome Res. 2006; 112: 60-66
        • Chiotaki R.
        • Petrou P.
        • Giakoumaki E.
        • Pavlakis E.
        • Sitaru C.
        • Chalepakis G.
        Spatiotemporal distribution of Fras1/Frem proteins during mouse embryonic development.
        Gene Expr Patterns. 2007; 7: 381-388
        • Pavlakis E.
        • Chalepakis G.
        pH-dependent antigen unmasking in paraformaldehyde-fixed tissue cryosections.
        Appl Immunohistochem Mol Morphol. 2008; 16: 503-506
        • Petrou P.
        • Makrygiannis A.K.
        • Chalepakis G.
        The Fras1/Frem family of extracellular matrix proteins: structure, function, and association with Fraser syndrome and the mouse bleb phenotype.
        Connect Tissue Res. 2008; 49: 277-282
        • Timmer J.R.
        • Mak T.W.
        • Manova K.
        • Anderson K.V.
        • Niswander L.
        Tissue morphogenesis and vascular stability require the Frem2 protein, product of the mouse myelencephalic blebs gene.
        Proc Natl Acad Sci U S A. 2005; 102: 11746-11750
        • Jadeja S.
        • Smyth I.
        • Pitera J.E.
        • Taylor M.S.
        • van H.M.
        • Bentley E.
        • McGregor L.
        • Hopkins J.
        • Chalepakis G.
        • Philip N.
        • Perez A.A.
        • Watt F.M.
        • Darling S.M.
        • Jackson I.
        • Woolf A.S.
        • Scambler P.J.
        Identification of a new gene mutated in Fraser syndrome and mouse myelencephalic blebs.
        Nat Genet. 2005; 37: 520-525