Biological and Molecular Analysis of a Low-Grade Recurrence of a Glioblastoma Mult i forme 1

I N T R O D U C T I O N Genomic instability is inherent in cancer cells (1, 2) including gliomas (3). As tumor cells divide, the resulting daughter cells contain the genetic alterations present in the parent cell, but they may also continue to evolve additional karyotypic and/or molecular changes. Local expansion of these divergent clones results in a regionally heterogeneous tumor. Thus, specific genetic abnormalities present in some tumor regions will not occur in others. This can result in local areas of altered phenotypic characteristics such as therapy resistance, increased malignancy, growth rate, microvascular proliferation, and invasiveness (4-13). We recently analyzed 38 separate regions of a low-grade oligoastrocytoma and demonstrated the presence of regional heterogeneity using both cytogenefic and flow cytometric analyses (14). These data demonstrated the presence of regions with enhanced genetic instability that generated several distinct clonal abnormalities occupying different regions of the tumor. Although one clone appeared to be more widely disseminated than the other three clonal populations, each was adjacent to an area in which >50% of the cells consisted of nonclonal aberrations. Difficulties in the diagnosis of glial tumors due to their regional heterogeneity is recognized, and has spawned the search for molecular markers of progression that may be detected in tumor specimens which erroneously reveal low histopathological malignancy. A number of potential markers have been identified such as various karyotypic abnormalities (1520), aberrant expression of the EGFR 3 (10, 21-30), loss of expression of the tumor suppressor gene DCC (31), increased expression of proliferation markers (32-35), aberrant expression of p53 and mouse double minute 2 (36-39), and aberrant expression of growth factors (40-45). In addition to impacting tumor diagnosis, the use of prognostic markers to study regional heterogeneity has further implications for clinical tumor management. Surgical removal of regions with local malignant transformation may be more effective in promoting regression and improving survival than a less directed resection. Reports of patients whose recurrent tumor is of a lower grade than the primary tumor are largely anecdotal, and an in-depth analysis of such tumor pairs has not been reported. We now report the molecular, cytogenetic, flow cytometric, and histological analysis of tumor samples from a primary and recurrent tumor from the same patient in which the recurrent tumor was of a lower histological grade than the primary tumor. M A T E R I A L S A N D M E T H O D S Patient Data and Tumor Histopathology. The patient was a 30-year-old female undergoing craniotomy for resection of a primary malignant brain tumor in the fight frontal lobe. A subtotal gross resection was performed, and the tumor was identified as a GBM (grade 4 astrocytoma) using the RingertzReceived 7/17/95; revised 9/8/95; accepted 9/15/95. This work was supported by NIH Grants CA50931 and CA25956. 2 To whom requests for reprints should be addressed, at Neuro-Oncology Research, Barrow Neurological Institute of St. Joseph's Hospital and Medical Center, 350 West Thomas Road, Phoenix, AZ 85013. Phone: (602) 406-3647; Fax: (602) 406-7172; E-mail: [email protected]. 3 The abbreviations used are: EGFR, epidermal growth factor receptor; DCC, deleted in colon cancer; GBM, glioblastoma multiforme; FISH, fluorescence in situ hybridization; FCM, flow cytometry; DI, DNA index; RPTP13, receptor protein tyrosine phosphatase 13; RT, reverse transcription. Research. on April 20, 2017. © 1996 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 188 Analysis of a Low-Grade Recurrence of a GBM Table 1 Primers used for PCR Analyses Primer Fragment Gene Primer sequence location a size b (bp) Reference DCC 5'-TTCCGCCATGGTFTTTAAATCA-3 ' Exon O 233 84 5'-AGCCTCATTTTCAGCCACACA-3' Exon P 5 '-AGGTCTTGAAGGCTGTCC-3 ' 1467-1484 200 25 5'-CTTCTCCACTGGGTGTAAGAGG-3' 1645-1666 5 '-TGTACCATCGATGTCTACAT-3 ' 3001-3020 832 a 25 5'-CTATCCTCCGTGGTCATGCT-3' 3813-3832 577/507 5'-GGGCTCTGGAGGAAAAGAAA-3' 254-273 1151 d A.C. Scheck e 5'-CAGCAAAAACCCTGTGATTT-3' 1385-1404 331 5'-AGTGTGCAAGTGCTTGCCTAT-3' 5716-5736 534 85 5'-TGCAGAATAGTCACTCTGCTG-3' 6250-6270 5'-TGCCTACTTCCCAACTGAG-3' 2295-2313 509 S.A. Norman f 5'-GCAGCATGAGTAGTGGAC-3' 2787-2804 5'-TGCCTACTTCCCAACTGAG-3' 2295-2313 2945 d S.A. Norman f 5'-ATTGCTCCGACATCATCTG-3' 5222-5240 375 5 '-CCACTGAACTTCTGATTCGC-3 ' 172-191 215 48 5'-GCGTGCTAGCTGGATGTCTT-3' 367-386 EGFR (TM) ~ EGFR (Cyt.) EGFR (Ext.) RPTP~3 (intact) RPTP~ (secreted) RPTP~ (truncated) Histone 3.3 a Nucleotide number or exon location. b Size expected from amplification of eDNA. c TM, transmembrane domain; Ext., extracellular domain; Cyt., cytoplasmic domain. a Lower number is the size of the truncated message. RT-PCR will not demonstrate the 2945-bp fragment due to its length. The cytoplasmic domain of the gene-encoding EGFR has two truncated forms that can be demonstrated by this primer pair; however, only the 507-bp fragment was detected in tumors LH and LHR. e Primers were designed to different exons using the Oligo 4.1 primer design software. ;Primers were designed using the sequence given in Kreuger and Saito (85). Burger (46) and revised WHO (47) systems for astrocytoma. The patient was treated with 6060 cGy to the tumor and a chemotherapeutic regimen that included 10-ethyl-10deazaaminopterin (edatrexate, 10-EDAM) and carboplatin. A second treatment regimen with focused ultrasound was also administered; however, the tumor recurred 5 months after the original surgery. At this time, a second large resection was performed and only diffusely infiltrative low-grade astrocytoma (grade 2) was found. The large tissue samples were dissected into 1-cm regions, and each region was further divided, with one sample going into formalin for routine paraffin-embedding, a second sample being frozen and stored in liquid nitrogen for molecular analysis, and the third sample used for cytogenetics. For cytogenetic and molecular analyses, the primary tumor sample was obtained as a single sample and coded LH. The recurrent tissue was received as five separate samples and coded LHR-A, LHR-B, LHR-C, LHR-D, and LHR-E in compliance with our Internal Review Board. Isolation of Total Cellular RNA and RT-PCR Analysis. Total cellular RNA was isolated from frozen tumor tissue using RNAzol B (Cinna/Biotecx Laboratories, Inc. Houston, TX) and conditions specified by the manufacturer. RT followed by the PCR (RT-PCR) was used to determine gene expression in a semiquantitative manner. Total RNA (1 Ixg/20 Ixl RT reaction) was reverse transcribed using the GeneAmp RNA PCR kit and random hexamer primers (Perkin-Elmer Corp., Norwalk, CT) with conditions specified by the manufacturer. Six-txl aliquots of this reaction were used for PCR with primers described in Table 1. Since these primers span introns, they will amplify the appropriate size fragments from cDNA only. The use of one RT reaction for multiple PCR reactions allowed us to use the "housekeeping" gene histone 3.3 as a " loading" control (31, 48). Samples were analyzed by electrophoresis through a 1.8% agarose gel, visualized using ethidium bromide, and a Polaroid photograph was taken of the resulting gel. Quantitation was done by densitometric analysis of the photographic negative (49). Negative controls without RNA were performed for each experiment. Tumor Dissociation for Cytogenetics and Cell Culture. The tumor samples were dissociated mechanically into singlecell suspensions by mincing the tissue to a fine pulp with scalpels, then aspirating several times through an 18-gauge needle as previously reported (15, 50). The single-cell suspension was plated in 25-cm 2 flasks at a concentration of 3-5 × 104 cells/flask to obtain short-term chromosome preparations for the analysis of first division cells (15). The remaining tissue fragments and single cells were plated to establish monolayer cultures. Conditioned medium was applied to all monolayer cultures. Conditioned medium was a 1:1 ratio (v/v) of standard growth medium (Waymouth 87/3 medium; Grand Island Biological Company, Grand Island, NY) containing 25% FCS (Hyclone, Logan, UT), and "spent" medium removed from a vigorously growing glioma cell line AO2V 4 (51). When cell cultures became semiconfluent or demonstrated active growth, the medium was changed completely. No antibiotics were used. All primary and passage 1 cell cultures not used in cytogenetic studies or drug resistance assays were frozen and maintained in liquid nitrogen. The primary untreated freshly resected tumor sample coded LH was a large mass of tissue. For the purpose of cytogenetic analysis, multiple small pieces were removed from different parts of this sample; these were combined and handled as a single specimen. The recurrent tumor consisted of five individual samples removed from different regions of the brain which were dissociated and cultured separately. Only the LHR-E sample yielded viable cells for cytogenetic analysis and monolayer Research. on April 20, 2017. © 1996 American Association for Cancer clincancerres.aacrjournals.org Downloaded from Clinical Cancer Research 189 Fig. 1 A, primary tumor LH: GBM with marked nuclear pleomorphism, vascular proliferation, and focal necrosis; B, recurrent tumor LHR: low-grade astrocytoma composed of well-differentiated fibrillary astrocytes with low cellularity, little pleomorphism, and no necrosis or vascular proliferation. cultures. Each of the other LHR regions had limited cellular outgrowth from one or more of the tissue explants. The cells migrating from these explants were very abnormal, containing multiple or gigantic nuclei. After several weeks in culture, large vacuoles formed and degeneration ensued (52). This is not an unusual feature in recurrent tumors. One explanation for the frequent death of recurrent tumor cells is that the bulk of such samples consists of necrotic tissue resulting from radiation and/or chemotherapy treatment. Cytogenetic analyses of first-division cells (primary cells) were performed on short-term cultures (15) within 72 h after plating. A minimum of two flasks were prepared for each sample, and the karyotypic deviation had to occur in both flasks before being scored. The cytogenetic analysis of the tumors in this report follows the International System for Human Cytogenetic Nomenclature (53, 54). Rearranged chromosomes were considered clonal markers if present in two or more karyotypes, and the gain or loss of a whole chromosome(s) had to occur in five or more karyotypes to be defined as a clonal population. Nonclonal cells were cells in which: (a) a rearranged chromosome occurred only once; (b) a specific loss and/or gain of chromosome(s) occurred in less than five metaphases; and/or (c) a metaphase contained a chromosome exchange, telomeric fusions, and single-stranded chromatin strands as previously reported (14, 50). Research. on April 20, 2017. © 1996 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 190 Analysis of a Low-Grade Recurrence of a GBM A biopsy of galea was obtained from the patient at the time of primary resection prior to therapy. The constitutional karyotype of the patient was determined by counting the chromosome number of 25 metaphases and karyotyping 5 cells. All metaphases were analyzed by standard Gor Q-banding techniques. FISH. Slides containing 4-5-1zm-thick paraffin-embedded sections were air dried and baked at 65°C overnight. The slides were incubated for 1 h in a 1:1 alcohol:chloroform mixture at room temperature, then treated with RNase for 1 h at 37°C, rinsed four times with 2x SSC (150 mM NaC1, 15 mM Na3C6HsO7) at room temperature, and dehydrated using cold, graded alcohols. The slides were air dried, denatured in 1% formamide in PBS at 70°C, dehydrated in cold, graded alcohols, and air dried. The slides were hybridized with chromosome 7-specific o~ satellite DNA probe in hybridization mix (Oncor, Inc., Gaithersburg, MD), sealed with a glass coverslip, and incubated at 37°C for 4 to 16 h. The sealant was removed, and the slides were incubated in 65% formamide at 43°C for high stringency (single probe) or in 50% formamide at 37°C for 2 rain for less stringent conditions. The slides were washed twice with 2X SSC (pH 7.0) at 37°C for 4 min and once with 1X phosphate-buffered detergent (Oncor, Inc.). Upon removal they were treated with blocking agent, then treated with fluoresceinlabeled avidin. The paraffin-embedded preparation was stained with propidium iodide/antifade, and the slides were then scanned using the appropriate barrier and exciter filters on a Zeiss Axioskop microscope. The signal ratio:nucleus in paraffin-embedded sections is very difficult to ascertain because of the number of incomplete cells in the preparation and cells lying on top of one another. Therefore, representative fields observed in the FISH analysis of the LH and LHR-E samples are provided without quantification. FCM. FCM was performed on paraffin-embedded material. After 3-1xm sections were removed for routine diagnostic purposes, three to five 50-1xm sections were cut for FCM. The tissue was processed using our modifications of the method of Hedley (34, 55). The 50-1~m sections were dewaxed in xylene and rehydrated through graded alcohols overnight in an Autotechnicon (Technicon Corp., Tarrytown, NY). After mincing, the specimens were incubated at 37°C for 2 h in a solution of 5 g/liter pepsin (Sigma Chemical, St. Louis, MO) in normal saline, with the pH adjusted to 1.8. After mechanical dissociation by forceful passage through a 22-gauge needle, the cell suspensions were filtered through 150and 30-1xm mesh (Small Parts, Inc., Miami, FL). Following centrifugation at 400 X g for 5 min at room temperature, the supernatant was discarded, and the pellet was resuspended in a staining solution consisting of 100 g/liter propidium iodide (Sigma Chemical), 1 g/liter ribonuclease (Sigma Chemical), and 1 ml/liter NP40 detergent (Sigma Chemical) in PBS. The propidium iodide-stained cell suspensions were analyzed on an Epics Profile flow cytometer (Coulter Electronics, Hialeah, FL) using 15 mW 488 nm light from a 25-mW argon laser. The DNA histograms were analyzed using Multicycle (Phoenix Flow Systems, San Diego, CA). DI (ratio of the DNA content of aneuploid and diploid populations) between 1.90 and 2.10 were considered near-tetraploid. Controls were not used; the first peak was defined as diploid by convention. Ki-67 Immunohistochemistry. Ki-67 immunostaining was performed using modifications of the original methods of Gerdes et al. (56). Five-t~m sections were cut from the frozen tissue block, fixed for 10 rain in cold acetone, and air dried for at least 2 h. The slides were incubated for 20 rain at room temperature with a 1:50 dilution of Ki-67 antibody (DAKO, Inc., Santa Barbara, CA) and developed using an alkaline phosphatase-antialkaline phosphatase kit with fast red chromogen. Intrinsic alkaline phosphatase activity was blocked with levamisole (Vector Laboratories, Burlingame, CA). The labeled sections were lightly counterstained with hematoxylin. The percentage of Ki-67-positive cells (labeling index) was determined from the area with the highest number of labeled cells. One thousand nuclei in contiguous microscopic fields were counted at X400. Cells pertaining to blood vessels (endothelial cells and pericytes) were not counted. Reactive astrocytes and residual non-neuronal parenchymal cells could not be distinguished from tumor cells on the frozen section slides. A nucleus was considered positive if it demonstrated either a diffuse or punctate (nucleoar) distribution of reaction product.