Cancer Genetics 204 (2011) 541e549

High resolution genomic profiling and classicalcytogenetics in a group of benignand atypical meningiomasHeidrun Holland a,1, Kristin Mocker b,1, Peter Ahnert a,c,g, Holger Kirsten a,c,f,g,Helene Hantmann a, Ronald Koschny d, Manfred Bauer e, Ralf Schober e,Markus Scholz c,g, J€urgen Meixensberger b, Wolfgang Krupp b,*aTranslational Centre for Regenerative Medicine (TRM), University of Leipzig, Leipzig, Germany; bDepartment ofNeurosurgery, University of Leipzig, Leipzig, Germany; c Institute for Medical Informatics, Statistics and Epidemiology,University of Leipzig, Leipzig, Germany; dDepartment of Internal Medicine, University of Heidelberg, Heidelberg, Germany;eDivision of Neuropathology, University of Leipzig, Leipzig, Germany; fFraunhofer Institute for Cell Therapy and Immunology,Leipzig, Germany; g LIFE Center (Leipzig Interdisciplinary Research Cluster of Genetic Factors, Phenotypes andEnvironment), University of Leipzig, Leipzig, Germany

Received Dece

12, 2011; accepted

* Corresponding

E-mail address:1 These authors c

2210-7762/$ - see

doi:10.1016/j.cance

Meningiomas are classified as benign, atypical, or anaplastic. The majority are sporadic, solitary,

and benign tumors with favorable prognoses. However, the prognosis for patients with anaplastic

meningiomas remains less favorable. High resolution genomic profiling has the capacity to provide

more detailed information. Therefore, we analyzed genomic aberrations of benign and atypical

meningiomas using single nucleotide polymorphism (SNP) array, combined with G-banding by

trypsin using Giemsa stain (GTG banding), spectral karyotyping, and locus-specific fluorescence

in situ hybridization (FISH). We confirmed frequently detected chromosomal aberrations in menin-

giomas and identified novel genetic events. Applying SNP array, we identified constitutional

de novo loss or gain within chromosome 22 in three patients, possibly representing inherited

causal events for meningioma formation. We show evidence for somatic segmental uniparental

disomy in regions 4p16.1, 7q31.2, 8p23.2, and 9p22.1 not previously described for primary menin-

gioma. GTG-banding and spectral karyotyping detected a novel balanced reciprocal translocation

t(4;10)(q12;q26) in one benign meningioma. A paracentric inversion within 1p36, previously

described as novel, was detected as a recurrent chromosomal aberration in benign and atypical

meningiomas. Analyses of tumors and matched normal tissues with a combination of SNP arrays

and complementary techniques will help to further elucidate potentially causal genetic events for

tumorigenesis of meningioma.

Keywords Meningioma, cytogenetics, single nucleotide polymorphism array, segmental

uniparental disomies, paracentric inversion

ª 2011 Elsevier Inc. All rights reserved.

Accounting for 24e30%of all primary intracranial neoplasms inadults, meningiomas are histologically graded according to theWorld Health Organization (WHO) as benign (WHO grade I),atypical (WHO grade II), or anaplastic (WHO grade III)

mber 7, 2010; received in revised form October

October 17, 2011.

author.

wkrupp@medizin.uni-leipzig.de

ontributed equally to this study.

front matter ª 2011 Elsevier Inc. All rights reserved.

rgen.2011.10.007

meningiomas (1). Mainly, meningiomas occur as sporadic,solitary, slow-growing, and benign tumors (2,3).

Atypical and anaplastic meningiomas represent 8e22% ofmeningiomas. They show a more aggressive biologicalbehavior, a less favorable prognosis (4,5), and high recur-rence rates between 38e78% (6,7). Although the grade ofmalignancy and the completeness of tumor resection areimportant predictors for tumor recurrence, several geneticaberrations are associated with amore aggressive phenotype(8,9). Meningiomas are among the best-described tumors interms of tumor genetics. However, our understanding of

542 H. Holland et al.

genetic aberrations associated with aggressive behavior ofmeningioma is still limited.

One of the most prominent chromosomal aberrationsdetected by comparative genomic hybridization (CGH) inmeningiomas is partial or complete loss of chromosome 22,found in 24e51% of published cases (10). Loss of hetero-zygosity (LOH) of chromosome 22 is observed in 47e72% ofmeningiomas. In one third of these cases showing LOH ofchromosome 22, somatic mutation of the NF2 gene, locatedon 22q12.2, is described as an early event in tumorigenesis(11). The NF2 gene appears to be less often affected withinthe meningothelial subtype in comparison with other menin-gioma subtypes (12). The same holds for skull base locali-zation of meningiomas in comparison with other localizations(13). Alternative NF2-independent pathways of tumor initia-tion of meningioma still remain unclear despite the identifi-cation of several other candidate genes on chromosome 22,such as BAM22, LARGE, SMARCB1, and MN1 (14).

The second most common genetic abnormality in menin-gioma is partial or complete loss of 1p, which simultaneouslyrepresents the most frequent progression-associated chro-mosomal aberration in meningiomas (15). Concentrating onchromosomal loci 1p13, 1p32, and 1p36, several genes, suchas TP73,RASSF1A,CASP9, JUN, and TNFRF25, have beenfocused on as candidates for tumor initiation and/or progres-sion (14,16e18). Further frequently detected abnormalities inmeningiomas, such as deletions of 6q, 10, 14q, and 18q andchromosomal gains on 1q, 9q, 12q, 15q, 17q, and 20, havebeen described to be associated with tumor progression anda higher grade of malignancy (11,14,19). Comprehensivegenomic characterization of meningiomas may improve theunderstanding of meningioma formation and progression.

As a possible new mechanism of tumorigenesis,segmental uniparental disomy (UPD) was our focus. Somaticrecombination or nondisjunction in mitosis may lead to loss ofone allele, and the remaining allele is then reduplicated.Another possibility is chromosomal breakage followed byreduplication to compensate for the loss of a segment (20).These events result in LOH without copy number change.Possible consequences are inactivation of tumor suppressorgenes or activation of oncogenes.

Our recent analyses of a case of atypical meningioma anda case of multiple meningioma showed segmental UPDregions as potential recurrent genetic events in meningiomas(21,22).

The aim of the current study was to assess occurrence ofpreviously described segmental UPD regions in a group of 10sporadic benign and atypical meningiomas. Furthermore, wesought data on other potentially recurrent segmental UPDevents, and we wanted to perform a comprehensive analysisfor novel chromosomal aberrations. For this purpose, wecombined cytogenetic techniques with molecular cytogeneticmethods. GTG-banding, spectral karyotyping (SKY), andlocus-specific FISH were complemented by analyses of copynumber and heterozygosity using Affymetrix Genome-WideHuman SNP Array 6.0. GTG-banding produces reproduciblebanding patterns on metaphase chromosomes and allowsdetection of aberrations on the single chromosome level(resolution 5e10 Mb). SKY is a multicolor FISH techniquethat allows for detection of cryptic balanced or unbalancedtranslocations and complex rearrangements in the karyo-type. With a resolution as low as 2e3 Mb, SKY allows

assignment of additional chromosomal material to theirchromosomes of origin. Two major limitations of SKY incontrast to GTG-banding are that duplications and deletionsare visible only with additional procedures, and intra-chromosomal aberrations are not detectable.

Applying this combination of techniques, we confirmedfrequently detected chromosomal aberrations and identifiednovel genetic events.

Materials and methods

Patients

Tumor resection was performed in 10 patients with primarysporadic meningiomas at the Department of Neurosurgery,University Hospital of Leipzig. Group 1 comprised fiveconsecutive cases with meningioma of WHO grade I (n Z 3male, n Z 2 female, mean age at time of surgery 54.8 y).Group 2 consisted of five consecutive cases with meningiomaof WHO grade II (nZ 1male, nZ 4 female, mean age at timeof surgery 64.0 y). Time intervals between admissions ofpatients in group 2 were longer than those in group 1 becauseof the lower incidence of atypical meningiomas. Informedconsent was obtained from all patients.

Using magnetic resonance imaging (MRI), tumor lesionswere found in the following intracranial regions: olfactorygroove (n Z 4), convexity (n Z 3), lateral sphenoid (n Z 1),tuberculum sellae (n Z 1), and falx (n Z 1). Detailed infor-mation on age, sex, tumor location, tumor grade, histopath-ological subtype, proliferation index, and surgical procedure(Simpson grade) is given in Table 1.

All patients lacked a family history of brain tumors andunderwent neither chemotherapy nor radiotherapy prior tosurgery, except patient 8, who received local adjuvant radio-therapy of a rectal carcinoma five years before meningiomadiagnosis. The proliferation index was calculated aspercentage of cells positive for proliferation marker Ki67evaluating 10 high power fields (�400 magnification). Themean MIB (Ki67) proliferation index was 3.4% for WHO gradeI meningiomas and 7.8% for WHO grade II meningiomas.

Isolation and culturing of primary tumor cells

For each tumor cell isolation, fresh non-necrotic surgicalspecimens were washed in phosphate-buffered saline (PBS)and mechanically disaggregated into small pieces that wereevenly distributed in a 25-cm2 cell culture flask (#83.1810Sarstedt, N€umbrecht, Germany) coated with AmnioMaxmedium (Invitrogen, Carlsbad, CA) and incubated at 37�Cand 5% CO2. Tumor attachment was monitored daily. Aftertumor cell outgrowth, tumor pieces were removed and cellswere covered with AmnioMax. Cells were sub-cultivated ata confluency of 90%.

Chromosome preparation, GTG, and FISH studies

Chromosome preparations were performed on primary tumorcell cultures using standard cytogenetic techniques (colcemidtreatment, hypotonic treatment, and methanol/acetic acidfixation). GTG-banding, SKY (according to manufacturer’s

Table 1 Clinical, molecular, and histopathological patient data

Sample

No.

Age,

y

Location

(main part) Site of origin

Simpson

gradeaWHO

grade Histology

MIB

(Ki-67) NF2 mutationb

1 45 fronto-temporal

right

lateral spenoid II I Meningothelial <1% -

2 53 fronto-basal right tuberculum

sellae

II I Meningothelial 4% -

3 50 frontal olfactory groove II I Meningothelial 4% -

4 57 frontal right convexity I I Microcystic 4% g.76466C>T

5 69 frontal olfactory groove II I Meningothelial 4% -

6 42 occipital right falx II II Atypical 8% g.43498_43499insA

7 58 frontal olfactory groove II II Atypical 8% g.43498_43499insA

g.74810C>T

(Zp.Gln407Stop)

8 65 temporal left convexity I II Chordoid 8% -

9 76 fronto-basal olfactory groove II II Atypical 8% -

10 79 temporal left convexity I II Atypical 7% g.43498_43499insA

a Simpson grade I: complete excision of meningioma, including infiltrated dura and bone.

Simpson grade II: complete excision of meningioma, with supposed reliable coagulation of dural attachment.b Positions of NF2 mutations are according to the NCBI (National Center for Biotechnology Information) reference sequence: NG_009057.1.

Shown are only variants not observed in dbSNP v132.

Cytogenetics of benign and atypical meningiomas 543

instructions; Applied Spectral Imaging, Carlsbad, CA), andlocus-specific FISH (according to manufacturer’s instructions;Abbott/Vysis, North Chicago, IL) were applied for chromo-some analysis using chromosome spreads. For each tumor,we analyzed 25 metaphase cells of the primary cell cultureusing GTG-banding and 15metaphase cells using SKY. FISHanalyses were performed on metaphase/interphase cells withprobes LSI P58/LSI Telomere 1p; 1p36; spectrum orange/green, and LSI TUPLE1/LSI ARSA; 22q11.2/22q13; spectrumorange/green (Abbott/Vysis).

Molecular karyotyping using SNP arrays

High resolution genome wide copy number variation analysisand assessment of heterozygosity were performed using theGenome-Wide Human SNP Array 6.0 (Affymetrix, SantaClara, CA) in resected tumor tissues. The array containsprobes for genotyping and copy number detection for nearlyone million SNPs. In addition, it contains close to one millionadditional probes for detecting copy number in non-polymorphic regions. Thus, the array allows genome widedetection of chromosomal copy number as well as estimationof heterozygosity at high resolution.

Genomic DNA from primary meningioma cells wasextracted using the DNeasy Tissue Kit (QIAGEN GmbH,Hilden, Germany). The integrity of genomic DNA waschecked by agarose gel electrophoresis. Array processingwas performed as a service by AROS Applied BiotechnologyAS (Aarhus, Denmark). Quality control involved severalaspects: Inspection of sample histograms (data not shown)did not reveal any hints at technical problems. In principalcomponent analysis (data not shown) of array intensity data,clustering with “scan date” was not observed, suggestingreproducible laboratory procedures. Male samples clusteredseparately from female samples, as expected. For allpatients, blood and tumor samples from the same individualclustered near each other with similar distances between

blood and tumor. We observed no clustering with tumor grade(WHO grade I or II).

Genotypes were called using the birdseed version 1algorithm (23) implemented in the Affymetrix GenotypingConsole software version 4.0, with standard settings. Forblood samples, the overall call rate was greater than 99.2%.For tumor samples, the overall call rate was greater than98.7%. Copy number analyses and detection of regions withLOH were performed in Partek Genomics Suite version 6.5(release number 6.10.1020, Partek, St. Louis, MO) accordingto given workflows with standard settings.

Copy number aberrations were detected with genomicsegmentation in Partek Genomics Suite. Copy numberregions were reported if a minimum size of 1Mb and anaverage smooth signal of at least 2.4 for gain regions ora maximum smooth signal of 1.6 for loss regions wereobserved. LOH was determined in Partek Genomics Suite forpaired samples to identify somatic changes in tumors versusblood (germline) and unpaired relative to a reference derivedfrom 102 germline samples of the local population to identifyoverall regions of LOH, including constitutive LOH.

Analysis of enrichment of LOH was performed using thesoftware R 2.12.0. The observed percentage of LOH within1p31.1, 6q14.1, 10q21.1, and 14q23.3 within a certain tissuetype was compared with the percentage of LOH within100,000 randomly drawn genomic regions of similar size ofthe same tissue type. On this basis, the odds ratio of thepercentage of LOH in WHO grade II versus WHO grade Iwas determined.

NF2 gene sequencing

All 17NF2 exons were amplified as described previously (24).Polymerase chain reaction (PCR) products were purifiedusing a PCR purification kit (Seqlab, G€ottingen, Germany).Sequencing was performed by Seqlab. Sequences wereanalyzed using FinchTV 1.4 (Geospiza, Seattle, WA) and

544 H. Holland et al.

compared with human genome build hg19 and dbSNP build132. Sequencing the PCR product in the reverse directionvalidated detected variants not reported in dbSNP.

Results

Results of GT- banding, SKY, and locus-specificFISH

In analyzing 25 metaphase cells by GTG-banding and 15metaphase cells by SKY for each tumor, we found 56structural chromosomal aberrations most frequently localizedon chromosomes 1, 2, 3, 4, 6, 7, and 22 in 10 meningiomas(Table 2).

Using these techniques, we identified a novel balancedreciprocal translocation t(4;10)(q12;q26) in 4 of 40 analyzedmetaphase cells in a benign meningioma (sample 3,Figure 1A). SNP array analysis also showed no chromosomalimbalance at the chromosomal breakpoint cluster regions4q12 and 10q26 in this sample (Figure 1B).

FISH analysis with locus-specific probes LSI P58/LSITelomere 1p (primary tumor cells and blood) and LSITUPLE1/ARSA (22q11.2/22q13) were performed. With theapplication of probe LSI P58/LSI Telomere 1p, the previouslydescribed paracentric inversion within chromosomal region1p36 (22) was detected as a recurrent chromosomal aber-ration in 7 of 10 meningiomas (4/5 WHO grade I, 3/5 WHOgrade II). Using the locus-specific FISH probe LSI TUPLE1/ARSA, we confirmed known chromosomal aberrations, suchas segmental deletions on 22q (4 samples: 2 WHO grade I, 2WHO grade II), and/or monosomy 22 (6 samples: 2 WHOgrade I, 4 WHO grade II).

Numerical chromosomal changes were detected inmeningiomas using GTG-banding and SKY, analyzing 25 and15 metaphase cells, respectively. Of 173 identified numericalchromosomal aberrations, the most frequently found weremonosomies of chromosomes 22 (34/400 metaphase cells,8 samples: 5 WHO grade I, 3 WHO grade II), 18 (12/400,

Table 2 Structural aberrations based on the results of GTG-bandi

indicates chromosomal aberrations detected in more than one meta

Sample

No. Gender

WHO

grade Chromomal aberrations

1 F I del(1)(p36.3p36.1)[2/40], d

del(7)(q11q21), del(18)(p1

2 M I del(1)(p36)[3/40], del(14)(q

3 M I t(1;7)(?;?), (t(2;?)(q37;?), t(4

del(7)(p13p15) [2/40], de

4 F I del(1)(p36)[5/40], del(1)(p2

5 M I t(5;9)(?;?)

6 F II del(2)(p16p12), del(3)(p?), d

7 M II del(2)(q32), del(2)(p?), del(3

8 F II del(1)(p36p31), del(4)(q?)[2

9 F II dup(20)(q13)

10 F II tas(13;19)(pter;pter), t(15;19

Note: Bold typing indicates chromosomal aberrations detected in more t

8 samples: 4WHOgrade I, 4WHOgrade II), and 19 (12/400, 7samples: 4 WHO grade I, 3 WHO grade II).

Results of SNP array analyses and NF2sequencing

Using Genome-Wide Human SNP Array 6.0, we detectednumerous chromosomal aberrations. We confirmed previ-ously described chromosomal imbalances, such as losses of1p, 2p/q, 3p, 6, 7p/q, 9p, 14, 19p, 14, 19p, and 22q, and gainof 22q. For the first time in primary benign meningiomas,we detected losses of 2p16.2-p11.1, 7p21.2-p15.3, and7q11.21-q21.11 (Table 3).

Further, one chromosomal imbalance not previouslydescribed for intracranial primary atypical meningiomas wasobserved by SNP array analysis: gain of chromosomalmaterial 19p13.12-p12 in the only chordoid subtype of thisgroup of meningiomas (Figure 1C). This gain was adjacent toa loss known to be recurrent in meningiomas (Table 3).Interestingly, losses or gains of 22q were detected not only inprimary tumor cells but were also found in the germline DNAof the same patients (1/10 and 2/10 analyzed patients,respectively, Figure 1D). To our knowledge, these aberra-tions have not been previously reported to be constitutional.

The Cancer Gene Census (25) lists a number of genes andloci known to be involved in various neoplasms. Ourcomparison of the Cancer Gene Census with our SNP arraydata identified 121 concordant genes within genomic regionsaffected by gains or deletions (data not shown). Of theseaffected genes, 56 are described for solid tumors and 10 havebeen published in connection with brain tumors (Table 3).

SNP array analysis uniquely allows genome wide analysisof heterozygosity. For the studied group of benign andatypical meningiomas, somatic LOH without a copy numberchange was observed in 3 tumor samples on chromosomalregions 4p16.1, 7q31.2, 8p23.2, and 9p22.1, suggestingsegmental UPD. LOH without copy number change wasfrequently found within all samples of both tumor grades aswell as in corresponding germline DNA samples from thesame individuals (Table 4a). Segmental UPD in regions

ng and SKY, (25 and 15 metaphases, respectively). Bold typing

phase spread of the respective tumor sample

el(1)(p21p12), del(1)(p?), del(2)(p?), del(2)(p21p13),

1.2), del(22)(q13)

32)

;7)(?;?), t(4;10)(q12;q26)[4/40], del(5)(q?),

l(16)(q24)

1), del(2)(p?), del(3)(p26), del(3)(p21p13)

up(6)(p?), del(10)(p12), del(12)(p12), dup(12)(q24.3q23)

)(p25p23)[2/40], del(7)(p15p13)

/40], del(6)(p22p12)[2/40], t(7;8)(?;?), t(19;20)(?;?); del(22)(q?)

)(q11;q12)

han one metaphase spread of the respective tumor sample.

Figure 1 Comparison of GTG-banding, SKY, andSNP array copy number analyses ofmeningioma. (A) Ideogramswith GTG-banding

(top), and SKY (bottom) showing the t(4;10)(q12;q26). (B) SNP array copy number analysis showed no significant imbalance within

breakpoint-cluster region 4q12 or 10q26 (blue rectangles indicate breakpoint-cluster regions). (C) SNP array copy number analysis

(Log2Ratio) revealed de novo gain of chromosomal material 19p13.12-p12 in the only chordoid subtype of this group of meningiomas

(top, indicated by the rectangle); hematoxylin and eosin staining showing characteristic chordoid areas (bottom). (D) SNP array copy

number analysis (Log2Ratio) revealed de novo loss or gain within chromosomal regions 22q11.22-q11.23 in normal control DNA and

primary tumor cells of three patients (green rectangles indicate regions with chromosomal gain, red rectangles indicate the region with

chromosomal loss).

Cytogenetics of benign and atypical meningiomas 545

Table 3 Overview of significant somatic aberrations detected by SNP array

Sample

No.

WHO

grade Cytoband

Copy

number

change

Physical

position

Length

(Mbp) Genesa

Confirmation

Start

(Mb)

End

(Mb) GTG FISH Litc

1 I 1p36.33-p22.3 Loss 0.8 86.7 85.9 PAX7, SFPQ, THRAP3,

MYCL1, CDKN2C,b JUN

Yes Yes Yes

1p22.1-p21.3 Loss 93.1 95.3 2.2 No No Yes

1p13.2-p12 Loss 113.3 118.3 5.0 TRIM33, NRAS Yes No Yes

2p22.3-p16.3 Loss 32.3 47.9 15.6 EML4, MSH2, MSH6 Yes No Yes

2p16.2-p11.1 Loss 54.1 91.1 37.0 Yes No No

2q24.3-q37.3 Loss 169.7 242.4 72.7 CHN1, CREB1,IDH1,b FEV,

PAX3, ACSL3, CMKOR1

No No Yes

7p21.2-p15.3 Loss 15.2 19.7 4.5 No No No

7p14.1-p11.2 Loss 40.2 54.7 14.5 No No Yes

7q11.21-

q21.11

Loss 63.8 78.6 14.8 Yes No No

7q22.3-q31.1 Loss 106.4 111.1 4.7 No No Yes

4 I 1p36.33-p22.2 Loss 0.7 91.4 90.7 PAX7, SFPQ, THRAP3,

MYCL1, CDKN2C,b JUN

Yes Yes Yes

1p21.3-p13.3 Loss 94.5 107.4 12.9 Yes No Yes

3p26.3-q11.2 Loss 0.0 95.0 95.0 SRGAP3,b VHL, PPARG,

RAF1,b MLH1, CTNNB1,

SETD2, MITF

Yes No Yes

7 II 3p26.3-q11.2 Loss 0.0 95.0 95.0 SRGAP3,b VHL, PPARG,

RAF1,b MLH1, CTNNB1,

SETD2, MITF

Yes No Yes

7p22.3-q11.21 Loss 0.1 62.8 62.7 Yes No Yes

8 II 1p36.33-p13.1 Loss 0.8 116.1 115.3 PAX7, SFPQ,THRAP3,

MYCL1, CDKN2C,b JUN,

TRIM33, NRAS

Yes Yes Yes

6p25.3-q27 Loss 0.1 170.8 170.7 HMGA1,TFEB, Yes No Yes

0.0 0.0 0.0 ROS1,b GOPC,b MYB

9p22.3-p22.2 Loss 14.3 18.2 3.9 No No Yes

9p21.3 Loss 20.1 23.4 3.3 CDKN2A -p16(INK4a),

CDKN2A- p14ARF

No No Yes

9p21.3-p21.1 Loss 25.0 29.7 4.7 No No Yes

9p21.1 Loss 30.0 31.5 1.5 No No Yes

14q11.1-

q32.33

Loss 18.1 105.3 87.2 HEI10,KTN1, RAD51L1,

TSHR, GOLGA5, AKT1

No No Yes

19p13.3-

p13.12

Loss 0.0 15.3 15.3 STK11, SMARCA4 No No Yes

19p13.12-p12 Gain 15.3 21.0 5.7 MECT1 No No Yes

22q11.1-

q11.21

Loss 14.4 18.2 3.8 No No Yes

22q11.21-

q12.1

Gain 19.4 24.6 5.2 BCR,b SMARCB1b No No Yes

22q12.1-

q13.33

Loss 24.6 49.6 25.0 MN1,b EWSR1, NF2,b

ZNF278, PDGFB, EP300

No No Yes

10 II 22q11.1-

q13.33

Loss 14.9 49.6 34.7 BCR,b SMARCB1,b MN1,b

EWSR1, NF2,b ZNF278,

PDGFB, EP300

No No Yes

Note: Regarding confirmation, concordance with results from GTG-banding and FISH are shown, as well as previous reporting in the

literature (Lit).a Data were obtained from the Wellcome Trust Sanger Institute Cancer Genome Project Web site (25).b Genes, described as aberrant in brain tumors.c Frequent chromosomal sequences by (array) CGH in meningiomas (38).

546 H. Holland et al.

1p31.1, 6q14.1, 10q21.1, and 14q23.3, was described inmeningioma in previous studies (21,22). Within theseregions, we found a higher percentage of copy number

neutral LOH (Table 4b). Enrichment analysis showed that thedetected increased percentage of LOH in WHO grade Icases and further increase in WHO grade II cases were

Table 4 Extent of segmental UPD regions suggested by copy neutral loss of heterozygosity (cnLOH)

WHO grade I WHO grade II

Tumor tissue Blood Tumor tissue Blood

a) Whole genome

Average cnLOH size in Mb (%) 324 (11.3%) 310.9 (10.8%) 325.2 (11.3%) 311.9 (10.9%)

Minimum cnLOH size in Mb (%) 307.8 (10.7%) 287.4 (10%) 294.4 (10.3%) 260.2 (9.1%)

Maximum cnLOH size in Mb (%) 347 (12.1%) 334.6 (11.7%) 346.5 (12.1%) 350.3 (12.2%)

b) 6q14.1, 10q21.1, and 14q23.3 only

Average cnLOH size in Mb (%) 3.2 (17.1%) 3.2 (17.2%) 4.1 (21.5%) 4.2 (22.2%)

Minimum cnLOH size in Mb (%) 1.1 (5.9%) 1.1 (5.9%) 2.2 (11.8%) 2.5 (13.3%)

Maximum cnLOH size in Mb (%) 4.8 (25.4%) 4.8 (25.4%) 5 (26.6%) 5 (26.6%)

Note: Shown is the average cumulative size of cnLOH regions (a) throughout the whole genome from tumor tissues and corresponding

germline DNA (blood). The same is shown for (b) selected regions suggested by previous findings (21,22).

Cytogenetics of benign and atypical meningiomas 547

unlikely to be due to chance (P < 0.05) for both tumor tissueand germline DNA.

Mutation analysis of the NF2 gene revealed five mutationsin 4 of 10 meningioma cases (1 WHO grade I and 3 WHOgrade II). Only one single mutation (g.74810C>T, corre-sponding to p.Gln407Stop in sample 7) showed evidencefor functional impairment of NF2 (Table 1). In addition, fiveknown SNPs were found to be polymorphic in our samples(rs13055076, rs2530664, rs140086, rs5763378, andrs79901896). These SNPs were located in intronic regions ofNF2 without evidence for functional implications.

Discussion

Whereas most meningiomas are sporadic, solitary, benigntumors with a good chance for complete surgical resectionand relatively optimistic outcome, the prognosis for patientswith atypical and anaplastic meningiomas remains lessfavorable. Genetics of WHO grade I and II meningiomas hasbeen considered to be among the best studied in humantumors (9,26). Recently, the combination of cytogenetictechniques and genome wide high resolution SNP arrays hasmade it possible to detect a much broader spectrum ofchromosomal aberrations than classical cytogenetic tech-niques alone. For instance, SNP array techniques uniquelyallow for genome wide analysis of heterozygosity. Examplesinclude solid tumors and hematological diseases (27e29).Here we present a study of benign and atypical meningiomasemploying GTG-banding, SKY, FISH, SNP array, and NF2gene sequencing.

The occurrence of copy neutral LOH and UPD werepreviously described for different malignancies (27,29e31),suggesting UPD as a possibly important mechanism intumorigenesis. Previously, in a case of atypical meningiomaand a case of multiple meningiomas, we identified segmentalUPD in regions 1p31.1, 6q14.1, 10q21.1, and 14q23.3(21,22). In this study, copy neutral LOH that suggests UPDwas enriched in three of these chromosomal regions incomparison with that of the genome wide average. This wasobserved for tissue samples from benign and atypicalmeningiomas and matched germline DNA (blood). Theenrichment of LOH was higher in samples of meningiomaWHO grade II compared with WHO grade I (Table 4). Thisenrichment appeared to be due to a higher number of LOHblocks as well as higher minimum and median LOH block

sizes. With our previous findings (21,22) and additionalevidence for constitutional copy neutral LOH on 14q inmeningiomas (24), this enrichment corroborates the notionthat constitutional segmental UPD in these regions may playa still unknown role in meningioma development. Evidence forthe involvement of constitutional segmental UPD also existsfor other tumor entities such as breast, prostate, and head andneck carcinomas (32). In paired analyses of tumor tissue andblood from the same patient, we identified somatic segmentalUPD regions 4p16.1, 7q31.2, 8p23.2, and 9p22.1 as newlydescribed events in primary meningioma cells. It remainsunclear whether these regions play a mechanistic role inmeningioma.

Possible consequences of segmental UPD are inactiva-tion of tumor suppressor genes or activation of oncogenes byvarious genetic and epigenetic mechanisms (20). In our data,the flanking of some UPD regions by regions of chromo-somal loss or gain suggests that repair mechanisms providea likely basis for segmental UPD in meningiomas (data notshown). Further investigations are necessary to elucidate therole of segmental UPD in meningiomas.

In this study, SNP array analysis detected previouslydescribed aberrations. In addition, we detected novel chro-mosomal imbalances, such as losses of 2p and 7p/q inbenign meningiomas, gain of 19p13.12-p12 in an atypicalmeningioma, and one loss and two gains within chromosome22 occurring in blood and tumor tissue simultaneously.

Loss of 22q is described as a common event in menin-gioma in general (12). Therefore, at first sight, it may appearconspicuous that we detected loss of 22q in only two cases ofWHO grade II meningiomas, described as chordoid andatypical, respectively. However, statistically, the observed20% of cases with loss of 22q do not fall outside of what wouldbe expected for the given number of samples. The presenceof large deletions of 22q correlates with histopathologicalsubtypes of meningiomas: Hansson et al (12) found losses of22q in only 39% of menigothelial meningiomas compared with65% and 86% in transitional and fibroblastic subtypes,respectively. In WHO grade I classification, four of five of ourmeningiomas were of the meningothelial subtype. WHOgrade II meningiomas without losses of 22q showed menin-gothelial differentiation in two of three of our cases. This mayfurther explain the relatively low number of observations of theloss of 22q.

On chromosome 22, NF2 is discussed as the major targetin meningioma (11). Mutational analysis of NF2 exons

548 H. Holland et al.

revealed mutations in 4 of 10 cases; 1 in WHO grade I and 3in WHO grade II (Table 1). This finding is in line with mutationfrequencies of 30e60% in sporadic meningiomas found inthe literature (14). Only one of the detected mutationsshowed evidence for potential functional impairment of NF2in the form of a premature stop codon. Interestingly,a correlation of absence of NF2 mutations with meningo-thelial subtype was described by Hannsson et al (12) andEvans et al (33). None of our meningothelial meningiomasharbored NF2 mutations. This may corroborate the notionthat in the meningothelial subtype, tumor mechanisms otherthan those involving NF2 may be of relevance.

Using GTG-banding and SKY, we detected a novelbalanced reciprocal translocation t(4;10)(q12;q26) [4/40metaphase cells] in a benign meningioma (Figure 1A). Thismay be of mechanistic relevance because balanced trans-locations are discussed as possible initiation events in varioustumor entities (34). The translocation t(4;10)(q12;p11), similarto the t(4;10)(q12;q26) detected by us, was described ina case of myeloproliferative syndrome with hypereosinophiliaas a new cytogenetic variant (35). Generally, balancedtranslocations may result in the formation of gene fusionproducts with altered biological function. Cancer geneslocated within the identified breakpoint regions are REST,RASL11B, PDGFRA (4q12) and ADAM12, BNIP3, DMBT1,FGFR2 (10q26) (25,36). Some of these cancer genes havebeen described specifically for brain tumors. For instance,Schrock et al (37) implicated the REST locus as an amplifi-cation site identified in glioma mapped to chromosome 4q12.

Recently, we detected a novel paracentric inversion withinchromosomal band 1p36 in a case of multiple meningiomas(22). This chromosomal aberration was seen only in thegrade II lesion. This led to the question of whether a para-centric inversion within chromosomal region 1p36 mightrepresent a molecular step in the transition from WHO gradeI to WHO grade II meningiomas. In the present groupof meningiomas, we found this paracentric inversion asa recurrent chromosomal aberration in seven independentpatients with meningiomas of WHO grade I or II. Therefore, itmay more likely represent an early chromosomal event inthe tumorigenesis of meningiomas. In the chromosomalregion of the paracentric inversion, the Atlas of Genetics andCytogenetics in Oncology and Haematology (36) lists 17entries: MDS2, PAX7, TNFRSF14, ARID1A, PRDM16,RPL22, SDHB, MIB2, CAMTA1, ERFI1, ENO1, MTHFR,PRDM2, CASP9, TP73, EPB41, and ALPL. Five of thesegenes are described in meningiomas: ENO1, MTHFR, TP73,EPB41, ALPL (14,38). Of these, TP73, EPB41, and ALPLare discussed as meningioma candidate genes (14). TheEPB41 protein, which links cell membrane proteins to thecytoskeleton, showed decreased expression (0e80%) inmeningiomas (39). M€uller et al (40) described a decrease inexpression of ALPL (alkaline phosphatase) of 14% inmeningiomas (WHO grade I), 46% in meningiomas (WHOgrade II), and 29% in meningiomas (WHO grade III). Furtherresearch is needed to clarify the exact role of these cancergenes in meningiomas.

Taken together, advances in acquiring SNP array data onan increasing number of tumors and matched germline DNAwill help to further elucidate the role of UPD in cancer.Further investigations are necessary for better understandingof the genetics of meningiomas.

Acknowledgments

The authors thank Rainer Baran-Schmidt and Karen Freybergfor excellent technical assistance, James Downs for readingthe manuscript, and Stephane Goutagny for the NF2sequencing protocols and primer sequences.

This project was supported by the German FederalMinistry for Education and Research by grants no. PtJ-Bio0315883 to H.H. and H.K. and research grant no. 01KN0702to P.A., and by grant no. 927000-040 from the Faculty ofMedicine, University of Leipzig, to K.M.

H.K., M.S., and P.A. were supported by the LeipzigResearch Center for Civilization Diseases (LIFE, Universit€atLeipzig). LIFE is funded by means of the European Union, bythe European Regional Development Fund (ERFD), theEuropean Social Fund and by means of the Free State ofSaxony within the framework of the excellence initiative.

References

1. Louis DN, Ohgaki H, Wiestler OD, et al. The 2007 WHO clas-

sification of tumours of the central nervous system. Acta Neu-

ropathol 2007;114:97e109.

2. Rockhill J, Mrugala M, Chamberlain MC. Intracranial meningi-

omas: an overview of diagnosis and treatment. Neurosurg Focus

2007;23:E1.

3. Rosenberg LA, Prayson RA, Lee J, et al. Long-term experience

with World Health Organization grade III (malignant) meningi-

omas at a single institution. Int J Radiat Oncol Biol Phys 2009;

74:427e432.

4. Bruna J, Brell M, Ferrer I, et al. Ki-67 proliferative index predicts

clinical outcome in patients with atypical or anaplastic menin-

gioma. Neuropathology 2007;27:114e120.

5. Chamberlain MC, Blumenthal DT. Intracranial meningiomas:

diagnosis and treatment. Expert Rev Neurother 2004;4:

641e648.

6. Jaaskelainen J, Haltia M, Servo A. Atypical and anaplastic

meningiomas: radiology, surgery, radiotherapy, and outcome.

Surg Neurol 1986;25:233e242.7. Perry A, Giannini C, Raghavan R, et al. Aggressive phenotypic

and genotypic features in pediatric and NF2-associated menin-

giomas: a clinicopathologic study of 53 cases. J Neuropathol Exp

Neurol 2001;60:994e1003.

8. Yang SY, Park CK, Park SH, et al. Atypical and anaplastic

meningiomas: prognostic implications of clinicopathological

features. J Neurol Neurosurg Psychiatry 2008;79:574e580.9. Zang KD. Meningioma: a cytogenetic model of a complex benign

human tumor, including data on 394 karyotyped cases. Cyto-

genet Cell Genet 2001;93:207e220.

10. Bayani J, Pandita A, Squire JA. Molecular cytogenetic analysis

in the study of brain tumors: findings and applications. Neuro-

surg Focus 2005;19:E1.

11. Campbell BA, Jhamb A, Maguire JA, et al. Meningiomas in 2009:

controversies and future challenges. Am J Clin Oncol 2009;32:

73e85.

12. Hansson CM, Buckley PG, Grigelioniene G, et al. Comprehen-

sive genetic and epigenetic analysis of sporadic meningioma for

macro-mutations on 22q and micro-mutations within the NF2

locus. BMC Genomics 2007;8:16.

13. Kros J, de GK, van TA, et al. NF2 status of meningiomas is

associated with tumour localization and histology. J Pathol 2001;

194:367e372.

14. Simon M, Bostrom JP, Hartmann C. Molecular genetics of

meningiomas: from basic research to potential clinical applica-

tions. Neurosurgery 2007;60:787e798.

Cytogenetics of benign and atypical meningiomas 549

15. Lopez-Gines C, Cerda-Nicolas M, Gil-Benso R, et al. Associa-

tion of loss of 1p and alterations of chromosome 14 in menin-

gioma progression. Cancer Genet Cytogenet 2004;148:

123e128.

16. Lomas J, Aminoso C, Gonzalez-Gomez P, et al. Methylation

status of TP73 in meningiomas. Cancer Genet Cytogenet 2004;

148:148e151.

17. Martinez-Glez V, Alvarez L, Franco-Hernandez C, et al.

Genomic deletions at 1p and 14q are associated with an

abnormal cDNA microarray gene expression pattern in menin-

giomas but not in schwannomas. Cancer Genet Cytogenet

2010;196:1e6.

18. Nakane Y, Natsume A, Wakabayashi T, et al. Malignant

transformation-related genes in meningiomas: allelic loss on

1p36 and methylation status of p73 and RASSF1A. J Neurosurg

2007;107:398e404.19. Perry A, Gutmann DH, Reifenberger G. Molecular pathogenesis

of meningiomas. J Neurooncol 2004;70:183e202.

20. Tuna M, Knuutila S, Mills GB. Uniparental disomy in cancer.

Trends Mol Med 2009;15:120e128.

21. Krupp W, Holland H, Koschny R, et al. Genome-wide genetic

characterization of an atypical meningioma by single-nucleotide

polymorphism array-based mapping and classical cytogenetics.

Cancer Genet Cytogenet 2008;184:87e93.

22. Mocker K, Holland H, Ahnert P, et al. Multiple meningioma with

different grades of malignancy: case report with genetic analysis

applying single-nucleotide polymorphism array and classical

cytogenetics. Pathol Res Pract 2011;207:67e72.

23. Korn JM, Kuruvilla FG, McCarroll SA, et al. Integrated genotype

calling and association analysis of SNPs, common copy number

polymorphisms and rare CNVs. Nat Genet 2008;40:1253e1260.

24. Goutagny S, Yang HW, Zucman-Rossi J, et al. Genomic

profiling reveals alternative genetic pathways of meningioma

malignant progression dependent on the underlying NF2 status.

Clin Cancer Res 2010;16:4155e4164.

25. The Cancer Genome Project. Available at: http://www.sanger.

ac.uk/genetics/CGP/. Accessed on September 3, 10.

26. Lee Y, Liu J, Patel S, et al. Genomic Landscape of Meningi-

omas. Brain Pathol 2009.

27. Andersen CL, Wiuf C, Kruhoffer M, et al. Frequent occurrence of

uniparental disomy in colorectal cancer. Carcinogenesis 2007;

28:38e48.

28. Gondek LP, Haddad AS, O’Keefe CL, et al. Detection of cryptic

chromosomal lesions including acquired segmental uniparental

disomy in advanced and low-risk myelodysplastic syndromes.

Exp Hematol 2007;35:1728e1738.

29. Mohamedali A, Gaken J, Twine NA, et al. Prevalence and

prognostic significance of allelic imbalance by single-nucleotide

polymorphism analysis in low-risk myelodysplastic syndromes.

Blood 2007;110:3365e3373.

30. Kloth JN, Oosting J, van WT, et al. Combined array-comparative

genomic hybridization and single-nucleotide polymorphism-loss

of heterozygosity analysis reveals complex genetic alterations in

cervical cancer. BMC Genomics 2007;8:53.

31. Yamamoto G, Nannya Y, Kato M, et al. Highly sensitive method

for genomewide detection of allelic composition in nonpaired,

primary tumor specimens by use of affymetrix single-nucleotide-

polymorphism genotyping microarrays. Am J Hum Genet 2007;

81:114e126.

32. Assie G, LaFramboise T, Platzer P, et al. Frequency of germline

genomic homozygosity associated with cancer cases. JAMA

2008;299:1437e1445.

33. Evans JJ, Jeun SS, Lee JH, et al. Molecular alterations in the

neurofibromatosis type 2 gene and its protein rarely occurring in

meningothelial meningiomas. J Neurosurg 2001;94:111e117.

34. Mitelman F, Johansson B, Mertens F. The impact of trans-

locations and gene fusions on cancer causation. Nat Rev

Cancer 2007;7:233e245.

35. Huret JL. t(4;10)(q12;p11). Available at: http://AtlasGenetics

Oncology.org/Anomalies/t0410q12p11ID1440.html. Accessed on

October 5, 10.

36. Atlas of Genetics and Cytogenetics in Oncology and Haematol-

ogy. Available at: http://AtlasGeneticsOncology.org. Accessed

on October 5, 10.

37. Schrock E, Thiel G, Lozanova T, et al. Comparative genomic

hybridization of human malignant gliomas reveals multiple

amplification sites and nonrandom chromosomal gains and los-

ses. Am J Pathol 1994;144:1203e1218.38. Cancer Genome Anatomy Project. Available at: http://cgap.

nci.nih.gov/Chromosomes/AbnCytSearchForm. Accessed on

September 3, 10.

39. Piaskowski S, Rieske P, Szybka M, et al. GADD45A and EPB41

as tumor suppressor genes in meningioma pathogenesis.

Cancer Genet Cytogenet 2005;162:63e67.

40. Muller P, Henn W, Niedermayer I, et al. Deletion of chromosome

1p and loss of expression of alkaline phosphatase indicate

progression of meningiomas. Clin Cancer Res 1999;5:

3569e3577.