Differential Involvement of Protein 4.1 Family Members DAL-1 and NF2 in Intracranial and Intraspinal Ependymomas Pratima K. Singh, M.D., David H. Gutmann, M.D., Ph.D., Christine E. Fuller, M.D., Irene F. Newsham, Ph.D., Arie Perry, M.D. Departments of Pathology (PKS, CEF, AP) and Neurology (DHG), Washington University School of Medicine, St. Louis, Missouri; and Department of Anatomy (IFN), Henry Ford Hospital, Detroit, Michigan KEY WORDS: Brain tumor, DAL-1, Ependymoma, Ependymomas are malignant CNS neoplasms with Glioma, NF2, Tumor suppressor gene. highly variable biologic behavior, including a gen- Mod Pathol 2002;15(5):526–531 erally better prognosis for intraspinal tumors. Inac- tivation of the NF2 gene on 22q12 and loss of its Ependymomas are glial tumors thought to arise protein product, merlin, have been well docu- from primitive ependymal or subependymal cells in mented in subsets of meningiomas and ependymo- the vicinity of the ventricles and remnants of the mas. DAL-1, a related tumor suppressor and protein central spinal canal (1, 2). Although they constitute 4.1 family member on 18p11.3, has also been re- only 5% of all neuroepithelial tumors, they are the cently implicated in meningioma pathogenesis, third most common brain tumor in the pediatric though its role in ependymoma remains unknown. age group. In children, the vast majority of ependy- Therefore, we evaluated 27 ependymomas (12 intra- momas are intracranial and are associated with fre- cranial and 15 spinal) using fluorescence in situ quent recurrences, whereas in adults more then hybridization (FISH) and immunohistochemistry half are intraspinal, and recurrence is rarer (3). The (IHC) to determine NF2/merlin and DAL-1/DAL-1 World Health Organization (WHO) divides ependy- status at the DNA and protein levels. Demonstrable mal tumors into conventional (WHO Grade II) and NF2 and DAL-1 gene deletions were each detected in anaplastic (WHO Grade III) categories, with a WHO 6 (22%) ependymomas. All 5 merlin losses by IHC Grade I designation reserved for two distinctively ؍ occurred in spinal ependymomas (P .047), indolent subtypes, subependymoma and myxopap- whereas 5 (71%) DAL-1–negative cases were intra- illary ependymoma (2). Unfortunately, histologic ؍ cranial (P .185). The former result is consistent classification has thus far proven to be an unreli- with prior observations that NF2 mutations are gen- able predictor of clinical behavior, with primary erally limited to spinal ependymomas. In contrast tumor location and extent of resection remaining to meningiomas, simultaneous merlin and DAL-1 the most meaningful prognostic determinants. losses were not encountered. Our findings suggest In comparison to diffuse astrocytomas and oligo- that (1) NF2 and DAL-1 losses are involved in the dendrogliomas, relatively little is known about the pathogenesis of spinal and intracranial ependy- molecular pathogenesis of ependymomas. The moma subsets, respectively and (2) given the num- most frequent abnormality reported to date is de- ber of cases with no demonstrable losses, other cel- letion of chromosome 22 (4–13). Gains of chromo- lular perturbations must also be critical for tumori- some 7 and losses involving chromosomes 6, 9, 11, genesis. and 13q have also been identified (10, 11, 14–16). Given the increased frequency of intramedullary ependymomas in neurofibromatosis 2 (NF2) pa- Copyright © 2002 by The United States and Canadian Academy of tients, it is interesting that chromosome 22 dele- Pathology, Inc. VOL. 15, NO. 5, P. 526, 2002 Printed in the U.S.A. tions and NF2 (22q12) gene mutations have been Date of acceptance: January 8, 2002. encountered primarily in spinal ependymomas (13, Supported in part by grants from the National Institutes of Health (NS/ CA41520 to DHG) and National Cancer Institute (CA777300 to IFN). 17–20). These findings were presented in part at the Annual ASCP/CAP Fall DAL-1 (Differentially expressed in Adenocarci- Meeting in Philadelphia, PA, October 20-23, 2001. Address reprint requests to: Arie Perry, M.D., Division of Neuropathology, noma of the Lung), a related tumor suppressor gene Box 8118, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110-1093; e-mail: [email protected]; fax: on chromosome 18p11.3, has recently been identi- 314-362-4096. fied (21, 22). As its name implies, this gene was 526 originally identified from studies on non–small cell saline citrate (SSC) and allowed to air dry. A carcinomas of the lung but has since been impli- fluorescein-labeled P1-derived probe targeting the cated in the tumorigenesis of breast cancer and DAL-1 gene (21) and a rhodamine-labeled cosmid- glioblastomas as well (22). Like NF2, DAL-1 is a derived NF2 probe cocktail (donated by Dr. Mia member of the Protein 4.1 superfamily, and alter- MacCollin, Massachusetts General Hospital, Bos- ations of both genes have been implicated in the ton, MA) were paired for each hybridization. The pathogenesis of meningiomas (23, 24). Given the probes were diluted in DenHyb hybridization homology between these two genes, we set out to buffer (Insitus Biotechnologies, Albuquerque, NM) better define the pathogenic roles of NF2 and at a concentration of 1:50. The hybridization mix DAL-1 in ependymoma. Our results suggest prefer- (10 ␮L per slide) was applied to the sections, fol- ential loss of merlin in spinal ependymomas and lowed by simultaneous denaturing of probe and DAL-1 in intracranial ependymomas. target at 90° C for 13 minutes. Overnight hybridiza- tion at 37° C took place in a humidified chamber. Posthybridization washes in 50% formamide/1ϫ MATERIALS AND METHODS SSC (5 minutes) and 2ϫ SSC (5 minutes) were per- Patient/Tumor Cohort formed at room temperature, and the slides were again allowed to air dry. DAPI (0.5 ␮L/mL; Insitus Surgical specimens from ependymomas resected Laboratories) was used as a nuclear counterstain, between 1990 and 2000 were retrieved from the files and the sections were viewed under an Olympus of the Lauren V. Ackerman Laboratory of Surgical BX60 fluorescent microscope with appropriate fil- Pathology at the Washington University Medical ters (Olympus, Melville, NY). Center, St. Louis, Missouri. All available slides were Sections showing sufficient hybridization effi- reviewed, and a representative paraffin block was ciency (Ͼ90% nuclei with signals) were evaluated, selected per case for further study. Five-micron- with 100–200 intact nonoverlapping nuclei scored thick sections were cut and mounted on poly-L- for the number of fluorescent signals. Cutoffs for lysine–coated slides for fluorescence in situ hybrid- abnormalities or deletions were based on counts ization (FISH) and immunohistochemistry (IHC). from nonneoplastic control specimens (temporal Clinical records were reviewed for evidence of neu- lobectomy specimens for seizure control) for each rofibromatosis Type 2 (NF2), and National Insti- probe. Interpretation of deletion required Ͼ50% of tutes of Health diagnostic guidelines were applied nuclei containing one NF2 or DAL-1 signal (mean ϩ (25, 26). 3 standard deviations in controls). Because nuclei with more than two signals were rarely seen in Immunohistochemistry nonneoplastic controls, polysomies (gains) were ar- Ͼ Slides were stained using automated IHC with a bitrarily defined as 5% nuclei containing three or DAKO autostainer (DAKO, Carpenteria, CA) as de- more signals. scribed elsewhere (24). Affinity-purified rabbit polyclonal antibodies against merlin (WA30) and Statistical Methods DAL-1 (3A1) were each applied at 1:500 dilutions, Associations between deletion frequencies (NF2 and antigen retrieval was achieved using 0.4% pep- or DAL-1) and tumor location (intracranial or in- sin in 0.01 N HCl for 30 minutes at 37° C. Positive traspinal) were evaluated based on results of the controls for merlin and DAL-1 included biopsy Fisher’s exact test. Reported P values of Ͻ.05 were specimens of sural nerve. Study cases were consid- considered statistically significant. ered positive if Ͼ1% of neoplastic cells displayed cytoplasmic staining. Positivity seen only at tissue edges was considered artifactual, and therefore, a RESULTS given case was considered negative if no staining was seen centrally within the tumor section. The study cohort consisted of 27 cases. Patient and tumor characteristics, as well as FISH and IHC results, are summarized in Table 1. There were 15 Fluorescence In Situ Hybridization spinal ependymomas, including one in the cervi- Dual-color FISH experiments were performed as comedullary junction. The intracranial tumors in- previously reported (27). After deparaffinization, cluded nine posterior fossa and three supratento- the sections were subjected to target retrieval by rial ependymomas. The median age for the entire steam cooking in citrate buffer for 20 minutes, fol- group was 27 years (range, 1–72), and none of the lowed by a 20-minute cool-down period and a patients had clinical evidence of NF2. However, 5-minute wash (distilled water). This was followed spinal ependymomas presented at a median age of by pepsin (4 mg/mL) digestion at 37° C for 30 min- 40 years, in comparison to intracranial examples, utes. The slides were then washed in 2ϫ standard which presented at a median age of 9 years. The 15 DAL-1 and NF2 in Ependymomas (P.K. Singh et al.) 527 TABLE 1. Summary of Patient–Tumor Cohort and Fluorescence In Situ Hybridization Immunohistochemical Results Case No. Age (y)/Sex Diagnosis Site NF2 Merlin DAL-1 DAL-1 1 15/F E Spinal Polysomy ϩ Polysomy ϩ 2 44/F E Spinal Deleted ϩ Polysomy ϩ 3 16/F MPE Spinal Polysomy ϩ Polysomy ϩ 4 41/F E Spinal Normal Ϫ Polysomy ϩ 5 38/F E CMJ Deleted Ϫ Normal ϩ 6 65/M E Spinal Deleted Ϫ Polysomy ϩ 7 33/F MPE Spinal Polysomy ϩ Deleted Ϫ 8 56/M E Spinal Deleted ϩ Deleted Ϫ 9 25/F MPE Spinal Normal ϩ Polysomy ϩ 10 13/F E Spinal Polysomy Ϫ Polysomy ϩ 11 53/M E Spinal Normal ϩ Polysomy ϩ 12 28/M E Spinal Polysomy ϩ Polysomy ϩ 13 40/M MPE Spinal NI ϩ NI ϩ 14 69/F E Spinal Polysomy ϩ Normal ϩ 15 44/F MPE Spinal Deleted Ϫ Normal ϩ 16 1/M E Post.