Advances in Brief

SPARC: A Potential Diagnostic Marker of Invasive Meningiomas1

Sandra A. Rempel,2 Shugang Ge, andJorge A. GutierrezHenry Ford Midwest Neuro-Oncology Center and the Departments ofNeurosurgery [S. A. R., S. G.] and Pathology [J. A. G.], Henry FordHealth Sciences Center, Detroit, Michigan 48202

AbstractSPARC, a secreted, extracellular matrix-associated

protein implicated in the modulation of cell adhesion andmigration, was evaluated as a marker for invasive meningi-omas. Although the majority of meningiomas are clinicallyand morphologically benign, ;10% progress into atypicaland malignant tumors, according to the standard criteria.However, a subset of meningiomas presents as histomorpho-logically benign tumors (WHO grade I), but they are clini-cally invasive. It has been suggested that these tumorsshould be classified as malignant, and that the patients mayrequire adjuvant therapy and closer follow up. Unfortu-nately, a significant number of these tumors may not berecognized because the surgical specimen used to assess thegrade of a tumor lacks the infiltrative interface with thebrain, which is currently necessary to determine its invasivecharacter. Therefore, a marker of heightened invasivenesswould greatly facilitate the identification of this subset ofpatients. In this study, the immunohistochemical expressionof SPARC in benign, noninvasive primary meningiomas wascompared with its expression in invasive, aggressive, pri-mary and recurrent meningiomas. SPARC was not ex-pressed in the 9 benign, noninvasive tumors, but was highlyexpressed in the 20 invasive tumors, regardless of the grade.The findings suggest that SPARC is a potential diagnosticmarker of invasive meningiomas and is capable of distin-guishing the histomorphologically benign noninvasive fromthe histomorphologically benign but invasive meningiomas,in the absence of the infiltrative interface.

IntroductionMeningiomas are proposed to arise from the arachnoidal

cap cells of the meninges (1) and may progress from benign tomalignant tumors (2, 3). They are the second most commonbrain tumor comprising;15% of intracranial and 25% of the

intraspinal primary adult central nervous system tumors (4). Therevised WHO system grades these tumors as benign (grade I),atypical (grade II), and anaplastic/malignant (grade III), basedon their histomorphological criteria (5). One criterion used isinvasion, the presence of which is usually associated with themalignant grade. However, there is a subset of meningiomasthat is histomorphologically benign but infiltrates the brain,suggesting that their histomorphologically benign features belietheir aggressive nature. Patients harboring these morphologi-cally benign but invasive tumors may therefore require closerfollow-up and adjuvant chemo- and/or radiation therapy. How-ever, a significant number of these patients may not be recog-nized and consequently not treated because the specimen used toassess the grade of tumor lacks the infiltrative interface that ispresently necessary to determine its invasive character. There-fore, a marker of heightened invasiveness other than the actualdemonstration of the invasive edge would greatly facilitate theidentification of this subset of patients. SPARC3 may be such amarker.

SPARC(also known as BM-40 and osteonectin; Ref. 6) isa developmentally regulated gene that is expressed in a numberof cell types (7). It is secreted into the ECM, where it maymodulate cell adhesion through a receptor-mediated event (6)that induces changes in cytoplasmic components associated withfocal adhesions (8), or it may interact directly with ECM com-ponents such as vitronectin to modulate cell adhesion (9). Thus,SPARC expression is associated with the modulation of celladhesion and migration.

We have recently identified and characterizedSPARCas agene that is overexpressed in the highly infiltrative humanastrocytic tumors (grades II–IV; Ref. 10). Using anin vivomodel, we have demonstrated that increased SPARC expressionis observed not only in the tumor mass itself but also in theindividual infiltrating astrocytic tumor cells (10). To further testthe hypothesis that increased SPARC expression correlates withtumor cell invasion, we examined its expression in meningiomasthat progress from benign to malignant grade. In addition, wehypothesized that increased SPARC expression would be able todifferentiate the histomorphologically benign, noninvasive men-ingiomas from the histomorphologically benign, but invasivemeningiomas. To test these hypotheses, we compared SPARCexpression in benign, noninvasive, nonrecurrent meningiomasand in invasive, aggressive, recurrent meningiomas.

Materials and MethodsTissue Samples. Institutional Review Board-approved

informed consent was obtained from all patients. Immediatelyupon surgical removal, tumor samples were fixed in 10% buff-ered formalin and embedded in paraffin for pathological diag-nosis. Tissue sections (5mm) were stained with H&E for his-

Received 8/11/98; revised 11/25/98; accepted 11/25/98.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisementin accordance with 18 U.S.C. Section 1734 solely toindicate this fact.1 This work was supported in part by a 1996–1997 Elsa U. PardeeFoundation Grant.2 To whom requests for reprints should be addressed, at Department ofNeurosurgery, Room 3096, Education and Research Building, HenryFord Health Sciences Center, 2799 West Grand Boulevard, Detroit, MI48202. Phone: (313) 916-8689; Fax: (313) 916-9855; E-mail: nssan@neuro.hfh.edu.

3 The abbreviations used are: SPARC, secreted protein acidic and rich incysteine; ECM, extracellular matrix.

237Vol. 5, 237–241, February 1999 Clinical Cancer Research

Research. on November 26, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

topathological evaluation. Two populations of tumors wereanalyzed. One set was composed of nine tumors from patientswithout recurrence that were histologically benign and wereconfirmed noninvasive by histomorphological examination ofthe tumor/normal tissue interface. The second set was composedof 20 primary and/or secondary invasive, aggressive tumors ofall grades derived from 9 patients with recurrences. Invasivenesswas assessed by examining the tumor sections for tumor/normaltissue interface and from descriptions within the surgical report.

Tumor Grading. Meningioma specimens were gradedas benign, atypical, or malignant/anaplastic and histologicallyclassified as meningothelial, transitional, fibroblastic, psammo-

matous, or papillary according to the WHO system (5). Theoccurrence (primary or recurrence number) of the tumor wasdocumented.

Immunohistochemistry. Immunohistochemical analysiswas performed as reported previously (10). Briefly, formalin-fixed, paraffin-embedded 5-mm tissue sections were subjectedto routine deparaffinization and rehydration. The subsequentsteps were performed at room temperature unless otherwisespecified. Sections were incubated for 10 min in 3% hydrogenperoxide, immersed in 10 mM sodium citrate buffer (pH 6.0) andboiled for 10 min, and then cooled for 20 min. Sections wererinsed in PBS solution and incubated with 10% horse serum in

Table 1 Grade, histomorphological type, invasive features, and SPARC expression levels and pattern in meningiomas

Patient no.a Specimen no.b Surgery date Histologyd

InvasiveeSPARCpatternf

SPARCintensitygGradec S/T On section By report

Nonrecurrent030 036-P 04/93 B M No 2033 039-P 07/93 B F No 2038 044-P 10/93 B M No 2044 050-P 01/94 B P No 2081 103-P 10/94 B M No 2083 105-P 11/94 B M No 2086 108-P 01/95 B T No 2094 116-P 03/95 B P No 2121 144-P 06/96 B T/P No 2

Recurrent003 004-R1 10/81 B M NP Dura D 41

072-R2 08/86 B M Dura Dura D 41

004 079-P 01/74 B/A M NP Brain S 51005-R1 11/81 A M Brain Brain S 4.51009-R2 11/83 A M/Pa Brain Brain S 51

007 008-R3 09/83 A/M T Brain Brain D 21 to 41

016 077-P 10/86 A/M M NP Brain, bone S 31 to 41021-R1 12/87 M M Dura, bone Bone, dura, SM D/S 4.51

020 080-P 07/92 A F Brain, dura ? D/S 41025-R1 01/93 A M Dura Brain, dura D/S 51081-R2 06/93 B/M Metastatic Bone SM, bone D/S 51

067 082-P 02/94 B T Brain Brain D 51083-R1 12/95 B/A T Brain Brain D 21

068 084-P 10/95 A/B F NP NP D/S 41085-R1 12/95 B/A M Dura Dura D 41

069 086-P 10/93 B M NP NP D 41087-R1 03/94 A/B M Dura Dura D/S 41088-R2 04/94 B M NP NP D 41

071 090-P 06/91 A/M T Brain Brain D 21 to 41091-R2 11/92 M/A M Brain, dura Brain D/S 11 to 41

a Each patient was assigned a study number. Note: Patient #020 had two brain tumors and a metastatic tumor to the spine.b Each specimen was assigned a study number: P, primary; R, recurrence followed by recurrence number.c As noted either on S, the section used for immunohistochemistry, or T, the entire tumor specimen; B, benign; A, atypical; M, malignant. Only

one grade was indicated if the grade was the same for both categories.d M, meningotheliomatous; F, fibroblastic; T, transitional; P, psammomatous; Pa, papillary.e As noted either histomorphologically on the sections or as noted in surgical reports. No, no invasion was observed at the tumor/normal tissue

interface or tissue invaded is indicated: brain, dura, bone, skeletal muscle (SM). NP, interface was not present and invasion could not be assessed.f D, diffuse, cytoplasmic staining pattern; S, speckled, intense, cytoplasmic, perinuclear staining pattern.g 2, negative;1, positive, on a scale of 0.5 to 5.

238 SPARC: Potential Meningioma Invasion Marker

Research. on November 26, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

PBS for 60 min. The sections were then incubated overnight at4°C with a 1:4800 dilution (0.74mg/ml) of primary anti-SPARCantibody (Haematological Technologies Inc., Essex Junction,VT) in PBS. After three washes in PBS buffer, the sections wereincubated for 30 min with biotinylated secondary antibody(1:200 dilution in PBS), washed, and incubated for 45 min withthe avidin-biotin complex according to the manufacturer’s in-structions (Vectastain ABC kit; Vector Laboratories, Burl-ingame, CA). Finally, the sections were washed, reacted withdiaminobenzidine in 0.1M Tris buffer (pH 7.6) with 0.03%hydrogen peroxide, followed by rinsing in tap water, counter-staining, and mounting. Negative controls were performed omit-ting the primary anti-SPARC antibody. Sections were blindlyreviewed and scored by a neuropathologist. Staining intensitywas graded as negative (2), weak (1), moderate (21to 31) orstrong (41 to 51) for 0% staining,,20% of cells staining,20–50% of cells staining, and.50% of cells staining, respec-tively. In addition, tumor cells were characterized as havingdiffuse cytoplasmic staining and/or intense, perinuclear cyto-plasmic staining patterns.

ResultsThe summary of occurrence, grade, histology, invasive-

ness, and SPARC intensity and staining patterns is presented inTable 1. All major histological subtypes and grades were rep-resented.

The different patterns of SPARC expression found in themeningiomas are illustrated in Fig. 1. The histologically benigntumors were either SPARC negative (Fig. 1A) or SPARC pos-itive (Fig. 1B). Atypical (Fig. 1C) and malignant/anaplasticmeningiomas (Fig. 1D) were also SPARC positive.

To determine whether a correlation existed betweenSPARC expression and tumor invasiveness and aggressiveness,SPARC expression was examined in histomorphologically be-

nign, noninvasive tumors selected from patients having nonre-current disease and compared with that of multiple tumors froma set of patients having aggressive, recurrent disease with doc-umented invasion. SPARC expression was found to be negativein all of the nine benign, noninvasive tumors (Table 1). Incontrast, SPARC expression was increased in all of the invasivetumors of the recurrent group, regardless of histological grade(Fig. 1, B, C, andD; Table 1).

SPARC-positive tumor cells and interstitium were ob-served at the invading margin of meningiomas that displayed adiffuse infiltrative pattern (11) where tumor cells and adjacentbrain cells intermingle, as observed with astrocytic tumors (Fig.2; Ref. 10), or in meningiomas that penetrate the brain in anodular fashion (Ref. 11; Fig. 3).

Increased SPARC expression was observed in tumor cellsinvading not only into the brain parenchyma (Figs. 1–3; Table 1)but also into other tissues such as dura (Fig. 4), skeletal muscle,and bone (Table 1).

We noted both extracellular matrix-associated (interstitial)and intracellular SPARC staining. Two patterns of intracellularstaining were observed: one a diffuse, cytoplasmic pattern; andthe other a strong, punctate perinuclear pattern.

DiscussionIn this study, we observed that increased SPARC expres-

sion is associated with the invasive phenotype in meningiomas,regardless of tumor histological grade. These results suggest thatSPARC is a candidate marker to identify those potentially oractually invasive tumors that are otherwise histomorphologi-cally benign and for which there is no tumor/brain invasiveinterface available for pathological assessment.

SPARC is a secreted, ECM-associated glycoprotein that isimplicated in the modulation of cell adhesion, migration, andangiogenesis during development (6). Increased expression has

Fig. 1 Immunohistochemicallocalization of SPARC in men-ingiomas of all grades. SPARCprotein was analyzed in benign,noninvasive meningioma (A),benign, invasive meningioma(B), atypical, invasive meningi-oma (C), and malignant, inva-sive meningioma (D) using a1:4800 dilution of anti-SPARC(osteonectin) antibody and de-tection with a diaminobenzi-dine reaction as described in“Materials and Methods.”SPARC protein was present inall of the invasive tumors (B–D), regardless of grade, but notin the noninvasive tumor (A).When present, SPARC was lo-calized to the cytoplasm of themeningeal tumor cells and theinterstitium. SPARC expressionwas also found in tumor endo-thelial cells (arrow). 340.

239Clinical Cancer Research

Research. on November 26, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

been associated with other cancer types (12–14), suggesting thatthe inappropriate expression may contribute to cancer progres-sion. We have demonstrated increased SPARC expression intumor cells and neovessel endothelial cells of astrocytic tumorsof all grades (10). In addition, we have demonstrated SPARCexpression in human glioblastoma cells invading rat brain in anin vivo model (10). We therefore hypothesize that the inappro-priate expression of SPARC may play two roles in brain tumorpathobiology, one in angiogenesis and one in tumor invasion.The present demonstration of the lack of SPARC expression in all

of the benign, noninvasive meningiomas and the high expression inthe tumor cells of aggressive, invasive meningiomas would supportthis hypothesis. That some of these aggressive tumors were histo-morphologically benign suggests that SPARC may be a specificmarker of invasion rather than a nonspecific marker of increasedaggressiveness related to tumor progression.

For those few tumors having high SPARC expression forwhich we could not document invasiveness in the initial resec-tion (Table 1, specimens 084, 086, and 088), the aggressivenessof these tumors and the demonstrable invasiveness of the recur-rent tumors (Table 1, specimens 085 and 087) support thehypothesis that SPARC may be a useful marker when the

Fig. 2 Diffuse infiltrative invasion at the tumor/brain interface.SPARC protein was immunohistochemically localized as described inFig. 1. SPARC-positive tumor cells (left) invading adjacent normal braintissue (right) are shown. Note the intense, perinuclear speckled patternof SPARC expression (C). Arrows,magnified region.A, 310; B, 320;C, 340.

Fig. 3 Nodular invasion at the tumor/brain interface. SPARC proteinwas immunohistochemically localized as described in Fig. 1. HighSPARC expression is observed at the invading tumor edge. SPARC-positive cells are also present throughout the invading nodular mass.A,310; B, 320; C, 340.

240 SPARC: Potential Meningioma Invasion Marker

Research. on November 26, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

interface is not available for evaluation. The usefulness of thisprotein as a marker is further supported by the observation thatincreased SPARC expression was observed on sections of thetumor that were of lower grade than that assigned to the wholetumor (Table 1, specimens 079, 008, 077, 081, 085, and 090).These observations suggest that the more histologically aggres-sive region of the tumor may not be required for the diagnosisof invasiveness. Furthermore, for patient 020, SPARC washighly expressed in the primary, recurrent, and metastatic tu-mors, an observation consistent with expression associated withthe invasive phenotype.

Increased SPARC expression was observed in tumor cellsinvading not only into the brain parenchyma but also into othertissues such as dura, skeletal muscle, and bone. Although inva-sion into tissues other than brain is not necessarily considered anindicator of malignancy, increased SPARC expression in tumorsinvading these tissues is consistent with a correlation of expres-sion with the invasive phenotype.

We noted ECM-associated staining, consistent with thereports that SPARC is a secreted protein. Intracellularly, both adiffuse, cytoplasmic pattern and a strong, punctate perinuclearpattern were observed. The significance of these patterns is notpresently understood, but the perinuclear staining observed inthe tumors has been observed in normal human fetal astrocytesin culture (data not shown) and in calvaria bone cells in culture

(15) and may reflect increases in protein synthesis, processing,secretion, and thus increased function.

On the basis of these encouraging preliminary data, immu-nohistochemical and correlative analyses of a larger number ofspecimens, including the rarer subtypes, are presently under wayto statistically confirm that SPARC expression correlates withthe invasive phenotype and is capable of predicting invasivebehavior in otherwise histologically benign tumors, therebyproviding a much needed diagnostic marker.

References1. Kepes, J. J. Meningiomas. Biology, Pathology, and Differential Di-agnosis.In: Stephen S. Sternberg (ed.), Masson Monographs in Diag-nostic Pathology, p. 1–2. New York: Masson Publishing USA, Inc.,1982.

2. Weber, R. G., Bostrom, J., Wolter, M., Baudis, M., Collins, V. P.,Reifenberger, G., and Lichter, P. Analysis of genomic alterations inbenign, atypical, and anaplastic meningiomas: toward a genetic modelof meningioma progression. Proc. Natl. Acad. Sci. USA,94: 14719–14724, 1997.

3. Rempel, S. A. Molecular biology of central nervous system tumors.Curr. Opin. Oncol.,10: 179–185, 1998.

4. Russell, D. S., and Rubenstein, L. J. Pathology of the Tumors of theNervous System, Ed. 5, pp. 1–1012. London: Edward Arnold, 1989.

5. Kleihues, P., Burger, P. C., and Scheithauer, B. W. The new WHOclassification of brain tumors. Brain Pathol.,3: 255–268, 1993.

6. Sage, E. H. Terms of attachment: SPARC and tumorigenicity. Nat.Med., 3: 144–146, 1997.

7. Mundlos, S., Schwahn, B., Reichert, T., and Zabel, B. Distribution ofosteonectin mRNA and protein during human embryonic and fetaldevelopment. J. Histochem. Cytochem.,40: 283–291, 1992.

8. Murphy-Ullrich, J. E., Lane, T. F., Pallero, M. A., and Sage, E. H.SPARC mediates focal adhesion disassembly in endothelial cellsthrough a follistatin-like region and the Ca11-binding EF-hand. J. Cell.Biochem.,57: 341–350, 1995.

9. Rosenblatt, S., Bassuk, J. A., Alpers, C. E., Sage, E. H., Timpl, R.,and Preissner, K. T. Differential modulation of cell adhesion by inter-action between adhesive and counter-adhesive proteins: characterizationof the binding of vitronectin to osteonectin (BM40, SPARC). Biochem.J., 324: 311–319, 1997.

10. Rempel, S. A., Golembieski, W. A., Ge, S., Lemke, N., Elisevich,K., Mikkelsen, T., and Gutierrez, J. A. SPARC. A Signal of AstrocyticNeoplastic Transformation and Reactive Response in Human Primaryand Xenograft Gliomas. J. Neuropathol. Exp.. Neurol.,57: 1112–1121,1998.11. Nakasu, S., Hirano, A., Llena, J. F., Shimura, T., and Handa,J. Interface between the meningioma and the brain. Surg. Neurol.,32:206–212, 1989.12. Porter, P. L., Sage, E. H., Lane, T. F., Funk, S. E., and Gown, A. M.Distribution of SPARC in normal and neoplastic tissue. J. Histochem.Cytochem.,43: 791–800, 1995.13. Bellacene, A., and Castronovo, V. Increased expression of os-teonectin and osteopontin, two bone matrix proteins, in human breastcancer. Am. J. Pathol.,146: 95–100, 1995.14. Porte, H., Chastre, E., Prevot, S., Nordlinger, B., Empereur, S.,Basset, P., Chambon, P., and Gespach, C. Neoplastic progression ofhuman colorectal cancer is associated with overexpression of thestromelysin-3and BM-40/SPARCgenes. Int. J. Cancer,64: 70–75,1995.15. Nefussi, J. R., Brami, G., Modrowski, D., Oboeuf, M., and Forest,N. Sequential expression of bone matrix proteins during rat calvariaosteoblast differentiation and bone nodule formationin vitro. J. Histo-chem. Cytochem.,45: 493–503, 1997.

Fig. 4 Tumor invasion into the adjacent dura. SPARC protein wasimmunohistochemically localized as described in Fig. 1. Strong SPARCsignal is present in sheets of invading tumor.Arrow, magnified region.A, 310; B, 320.

241Clinical Cancer Research

Research. on November 26, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

1999;5:237-241. Clin Cancer Res   Sandra A. Rempel, Shugang Ge and Jorge A. Gutiérrez  MeningiomasSPARC: A Potential Diagnostic Marker of Invasive

  Updated version

  http://clincancerres.aacrjournals.org/content/5/2/237

Access the most recent version of this article at:

   

   

  Cited articles

  http://clincancerres.aacrjournals.org/content/5/2/237.full#ref-list-1

This article cites 12 articles, 1 of which you can access for free at:

  Citing articles

  http://clincancerres.aacrjournals.org/content/5/2/237.full#related-urls

This article has been cited by 18 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/5/2/237To request permission to re-use all or part of this article, use this link

Research. on November 26, 2020. © 1999 American Association for Cancerclincancerres.aacrjournals.org Downloaded from