E I . S E V I E R

Gliomas in Families: Chromosomal Analysis by Comparative Genomic Hybridization

Atul Patel, Donald J. van Meyel, Gayatry Mohapatra, Andrew Bollen, Margaret Wrensch, J. Gregory Cairncross, and Burt G. Feuerstein

A B S T R A C T : Gliomas that aggregate in otherwise unremarkable famil ies m a y have a heritable genetic basis. To determine the spectrum o f genetic alterations in glioma-susceptible families, we examined tumor DNA from famil ial cases for regions o f chromosomal gain or loss using comparative genomic hybridization (CGH). We compared chromosomal alterations within and among glioma famil ies to those f o u n d in sporadic gliomas. A specif ic chromosomal abnormali ty common to the tumors o f mult iple unrelated probands with glioma or a specif ic chromosomal abnormali ty common to mult iple af fected pers'ons in a single glioma-pro;m fami l y would support the hypothesis o f an inherited predisposit ion to glioma and at the same t ime identify specif ic regions o f the genome harboring putat ive glioma suscepti- bility genes. Tumor DNA from 11 pat ients from seven famil ies with two or more individuals with glioma was analyzed, including three members o f a remarkable fami l y having I0 affected individuals. We ]ound no chromosomal abuormali ty common to all tumors (if all prebends nor did we f ind family-spe- cific abnormalit ies in two o f three glioma-prone kindreds. There were frequent copy number aberrations (CNAs) on chromosomes 7, 10. 19, and the sex chromosomes; other CNAs included +3q(13.3-29), - 4 q , +5q. -9q34 , ~- 12, --13q(21--933), - 15. - 16p, + 17qter, - 18, - 21 , and -22 . Ampli f icat ions occurred at + + 7p(1 I. 1 --~12), + ~ 7q(21.2---)33), Jr + 12q(13.2-~14), and ~- 4 12p(I 1-412}. Although there: were several novel CNAs [ - I 6 p , and + + 1213(I 1-p12)], none could readily explain the inheritance o f these tumors. ~ Elsevier Science Inc., 1998

I N T R O D U C T I O N

Most cancers result from the accumtdat ion of genetic mutat ions at critical loci causing act ivat ion of oncugenes and inact ivat ion of tumor suppressor genes [1], and many occur in a familial form [2]. Genetic predisposi t iun to gl ioma occurs in the (:ontext of the Li-Fraumeni syndrome, ueurof ibromatosis type 1, and tuberous sclerosis, mult i - sy s t em disorders tar which causat ive genes have been localized or c loned [3-6]. Gliomas also occur in families not known to have these syndromes [7], and presumably

From the Cam:er Genetic.s Program. UCSF Cancer Center, the Department of Laborato~ Medicine (A. P., G. M.. B. G. F.}. t3rain Tumor Research (,'enter. the Department of Neurological SurgeD" {A. P.. 13. G. F.), U;zit of Neuropathology. the'. Department of Pathology (A. 13.), and the Department of Epidemiology and Bio- statistics (M. W.). University of Califigrnia, San Francisco, U.S.A., and the Department of Microbiology and hnmunology (D. J. v. M.. J. G. C.). the Department of Oncology (D. ]. v. M., J. G. C.), and the; Dtwartmcnt of Clinical :Veuroh;gical Sciences (]. (;. C.}, Univer- sity of Western Ontario and the London Regional Cancer Cetztrc. London. Ontario. Canada.

Address reprint requests to: Burr (;. Feuerstein. Cancer (;enet- ics Program, LICSF Cancer (h.'nter. University of California, San Francisco, CA 94143-0808.

Received April 3, 1997: accepted June 19, 1997.

C a n c e r Genet Cy't ogenet 100 :77 -83 (1998) .(:" E l sev ie r ,ql:it!n(:e Inc., 1998 655 : \ w m u e of the A m e r i c a s . N e w York. NY 1001{1

these familial aggregations also have a heri table genetic basis, al though this has not been proven. Studies of fami- lies with glioma as a strategy for identifying brain tumor suscept ib i l i ty genes mav have broader relevance; familial cancers often arise as a result of the same genetic alter- at ions that affect their sporadic (non-familial) counterpar ts [8]. To determine the spect rum of genetic al terat ions in gl ioma-suscept ible families, we examined tumor DNA from familial cases of regions of chromosomal gain or loss, using comparat ive genomic hybr id iza t ion (CGH). CGH survevs the entire genome for copy number aberrat ions (CNAs), s imul taneous ly mapping regions of tumor DNA that have been ampl i f ied or deleted [9, 10]. Since the p rebend may be the only affected member alive when a familial t endency is recognized, s tudies of uncomnmn, fatal, inheri ted cancers often are compromised by tissne availabil i ty. CGH is well sui ted to thmilial s tudies because archiwd tumor tissue van be used [11] and pat ient-matched DNA from normal tissues is not required. Using an innova- tive strategy to identify tumor suppressor genes, a suscepti- bi l i ty locus h)r cancer-prone families with Peutz-Jeghers syndrome was tbund recently by C(;H analysis of tumors of affected members [12]. CGH revealed a common site of DNA loss on chromosome 19p, and subsequent analysis of a po lymorph ic marker in this region showed unambiguous

{) 1 ( i5-4{i08/98/$19.00 }'II S0165-4608{97)00275-6

78 A. Patel et al.

genetic l inkage with the Peutz-Jeghers phenotype, indicat- ing the h)cation of a single causat ive gene of high pene- trance [12]. In addi t ion. CG! 1 analysis of familial renal (:ell car(:inomas has revealed CNAs common among the tumors of members of the same family, indicating a heritable mec'hanism resulting in ( 'hromosome-spe.cific genomic instabil i ty and/or selection tot specific lnutaticms [131. To assess whether gliomas in families harbor distinct (:ytoge- netic abnormali t ies at the sites of glioma susceptibi l i ty genes, we comt)ared chromosomal alterations within and hetwee.,~ glioma families and also compared alterations seen in glioma kindreds to those found in sporadic gliomas.

MATERIALS AND METHODS

Tumor Spec imens

Tumor DNA from 11 pat ients froln seven families with two or more ind iv idua ls with glioma was ana lyzed - - t h r ee affected membe, rs of one kindred, two affected n)emhers of two kindreds , and four unrelated proban(ts. 'Fable 1 Stlln- marizes cl inical and pathologic:al data on these pat ients and Figure 1 depicts the famih, pedigrees. The tumor spec:imens were reviewed, classified, and grade, d bv the 1993 World Health Organizat ion criteria [14]. In addi t ion, a quart i le est imate (<25%, 25-50%, 50-75%, or >75%) of the alnount of tumor present in each sample was made. All tumors were formalin-fixed and paraffin embectded; 30 sec'tions eac:h lOtxln ill thickness were used per specimen for DNA isolation. Normal DNA was isolate(l from leuko- cyte, s of nornml male and female donors.

DNA Isolation

Tissues were treated in xyhme at 55°C to relnove wax, washed in ethanol, and then washed in t)hosphate-buff- e.red saline. Cellular melnbranes were lysed in a buffer containirtg 1% Tr i t on -X- lO0 .5 mM Mg(]lz, O.32 M sucrose.

anct 10 mM Tris (pH 8.0); next, nuclear membranes were lysed in a buffc, r of 20 mM EIYI'A, 100 mM NaC-;1, and 10 mM Tris (ptl 8.0). Samph.'s were inc:ubate.ct for 3 days at 50°C adding proteinase K (500 p.g/nil) and SDS (to 0.5%) e, verv 8-12 hours, extracted twice with phenol:chloro- forin:isoamyl alcohol (25:24:1), ethanol precipi tated, and resuspended in lO mM Tris (plI 8.0). 1 mM EI')TA.

DNA Probe l,abeling

Probes were hlbelect by nick translat ing total genera l ( DNA (1 txg) from normal or tumor tissue. Normal DNA was labeh;d with f luorescein-11-dUTP and tumor DNA with Texas Red-5-dUTP (DuPont: Wilmington, DE) in sep- arate 50 gl rcac:tions, following the manufacturer ' s proto- (:c~l. Probe fragment size (optimal range, 600-2000 base l)airs) was de termined hy agarose eh. 'clrophoresis, l,abels were "reversed" by incorporat ing Texas Red inl(.) normal DNA and tluores(:ein into tumor I)NA to w.,rifv CNAs and identify false posit ive results 115].

Comparative Genomic Hybridization

Slides with metaphase, chromosome spreads from a nor- real, healthy mah: (toner were prepared as ternt)lates for hvhrictization, using s tandard procedures. The methods for s l ide t)retreatment, hybridizat ion, washing, and fluo- resc:ence image a(_:quisition were performed as previously descr ibed 1151. hnages of t ight or mort; (:()pies of eac'h a t l t O S O l l l e a n d four copies of each sex chrolllOSOllle were acquired pe.r hybridizat ion, lehmrescence intensi ty ratios were (:ah-:uhded as a function of chromosome le.nglh and ncnTnalized ~,ppropriah'.ly It) form ratio profiles for e.ac:h (:hrolnosome. At dllV point along a (:hrolnosonm, a ratio gr(;ater than 1,0 indicated relat ively greater b inding by the fiuoro, s(:ein-labeled pr()l)e., whi le a ratio of less than 1.0 indicated relat ively greater binding of the probe labeled with Texas Re, d. ' the fluores(:ence ratios from all copies of

Table 1 Clinicx)pathologic characteris t ics and cytogenetic abnormal i t ies (h:tecte.d t) 3, comparative. genera l ( hybr id iza t ion in familial gl iomas

Tumor in Genetic gains and losses l'atimll no. Family' Age (yrs)/gende.r Pathology specimen de(co:ted by C(;tl

1 A 13/M AA 25-5(} Y 2 A 6/NI A .... 25 No dctectabh.' abnormality 3 A 7 3 / M ( ;M 511 - 7 % 10, 13(t(21-"13) , 21

4 B 5 2 / M ( ;M . :25 • - : -7p(11.1---~12),- 19

5 B 25/1" A t ) :-75 l p . ~ 3 q ( 1 3 . 3 - 2 9 ) ,

4 ( i , - 1 2 . - 1 5 . . 1 ! 1 { 1, X 6 (; 7 7 / M (.;M 5(1-75 }-7p(11 .21) . ~ . t -7p{11 .1-+12) .

• 7 q [ 2 1 . 2 - 3 3 ) . (. |q34. 10. H i p . - 1 9 . - 2 2

7 (.] 4 4 / M ( ;M 5 0 - 7 5 + 7 . 1(1.. 1 2 p . ' t 12p(11-~.12}. • ~ t 1 2 q ( 1 3 . 2 -+14 ) , - 13q[21--<.13}

8 D 68/M (,M 50-75 ~7, 10, 18

9 E 39/I" O 511-75 I p.- 19( t

10 F' 33/1" A 50 , 5 . - 7 . ~ I'L- Hi lL- X 11 (] 3 a / M ( )A -- 25 t 1 7 q t e r , Y

Abbrt~rations: A. aslrl)Q,'ttmla, AA. anaplasli{: aslr(It:vhm'ia: (). oligo¢lendrt~glhmta: At). anapla~,lit: oligi~dendrt~gli¢ml;l; ()A. mixed oligt)astrocytorna: (;M. glii)l)lastonla mullifin'me.

' ( : r . ss referent:e wilh Figure 1.

CGH in Familial Gliomas 79

each chromosome were compiled and averaged, and the mean and standard deviations were plotted and evaluated for divergence from 1.0. This information was combined with visual inspection of color images to determine areas of relative gain or loss in genomic material, defined as copy number aberrations (CNAs). A ratio of greater than 1.2 or less than 0.8 was defined as significant if present in

both fl)rward and reverse CGH profiles. 3"his is based on results from 20 hybridizations that compared the intensity of normal DNA labeled with Texas Red with the intensity of normal DNA labeled with fluorescein. We set the cut- offs at 2.67 standard deviations. The average standard deviation of CGII ratio over the genome was 0.075. Because of tissue constraints, some san]pies were less than

A

Figure 1 Pedigrees of seven glioma fiimili(:s. L. unaffe(:ted male, C:, unaffected female, I , aftee:ted xllale; @, affected f(:mah;: ~. unknown sex. Numbers represent tho number of individuals in the; indicated class. Tumors e)f 11 affected individuals studied by CGH tire,' noted #1 through 011 below the,' corresponding patient symbol. Patho- logic diagnosis of all affected individuals (filled symbols) also noled, with lhe following abbreviations: A, astro(:y- toma: AA, anaplastic astro(:ytoma; (), oligoeiendroglioma; AO, anaplastic oligodentiroglioma: OA. mixed oligoastrocytoma: GM, glioblaste)ma mtfltifi)rme: G, glioma of unknown pathology.

,° T G G G G M

# 3 ( G M )

B

GM

#5 AO

AA #2 A

C D

#7 GM GM

GM #1 A A

A #8 GM

E

#9 O

F G

#10 AA #11 A OA

A

80 A. Patel et al.

50% tumor. C(;H ratio deviations were only considered significant if the, standard deviation of the ratio did not overlap 1.0, and if the ratio was confirmed in a reverse hybridization. Due to the blocking effects of Cot-1 DNA, ratio (:hanges at hetero(:hromatic regions were unreliable and therefore disregarded.

RESUI,TS

Table I summarizes the CNAs in 11 familial glioma speci- mens. Tabh; 2 presents the compiled CGIt data by chromo- sonie and indi(:ates the frequency of (;NAs found previously ii1 72 unsele('led primary glioblastomas (GBM) [161. CNAs were frequent in. the familial cases on (:hromosomes 7, 10, 19, and on the sex chromosomes. Chromosome 7 deniers- strafed various patterns of gain in six n f l l sainples, includ- ing gain of the entire chrornosoine, gain of the q arm (rely, and interstitial gains or amplifications at both 7p(11.1-->21) and 7q(21.2--)33). Chromosonm 10 was lost in four of 11 samples, always in gliohlastonms and always in its entirety. [,ikewise, entire sex chromosoines we, re lost in four of 11 samples (two cases with loss of the X (:hromn- some and two with loss of the Y). Chromosonw, 19 aherra- tions occnrred in four of 11 samples, including gaiu of the entire ( 'hromosolne in one sample and loss in three others; (:hromosonm 19 was lost entirely in ()lie case and the q- arm only was It)st in two others. N(itably, the h)ss of 19q o(:l:nrred in the two oligodendroglial ttnnors (samples #5 and #9); also COlnmon to these oligodendroglioma samples was deletion of lp, a (]NA not observed in the astrol:yti(: specimens. ()ther CNAs include(t: +3q(13.3--)29). - 4 ( t , t5q, -9q34, ~12, -13q(21--+33),-15, 16p,--17qter.- 18, -21 , and - 22.

Amplifications were obse, rved ill several patient sam- ples: these inl:luded + +Tp(l l . l - - ,12)(samples #4 and #6); ~-~-7(t(21.2-.~33)(samtAc, #6); +-+12q(13.2-+14)(saml~le #7)

and a_ +-12p(l l~12)(sample #7).

DISCUSSION

Since glinmas sometimes oc(:ur in mnlt iple members of otherwise unremarkahh; families, it is reasonable to hypothesize that certain individuals are at risk for glimna [)y virtue of inherited mutat ions in one or more glioma susceptibili ty genes. The finding bv ('GH of a specific chromosomal abnormali ty cninmon tn the tumors of mul- tiple unrelated probands with glioma or the finding of a spe(:ifi(: chrornosonml abnorlnali tv coninlon to multiple affecte, d persons in a single glioma-iirone family woul(t support tile hyp(ithesis of an inherited predisposit ion to glioma and at the same tilne identify specific regions of the genoine harhoring putative glioma susceptihility genes. In this study, we found no CNA comin(m to all tumors (if (ill probands nor did we find family-specifi(: ahnorinalit ies ill two of three glioma-prnne kindreds (fam- ilies A and B in Table 1 and Figure 1). Two glioblastomas in family C shared gains of chromos(nne, 7 and losses of (:hrnm(isome 10. These (:hanges are (:ommonlv ollserved in glioblastomas (Tahle 2) and likely represent a(:quired

somatic mutat ions asso(-iated wilh progression ill each CASH. Multiple CNAs were detected in gliomas that aggre- gote ill families and now, for the first tilne, their cytoge- netic tn'ofiles (:an he (:oinpared to abnormalities known to occur in unselected gliomas.

Several areas of the gen()nie are, altered commonly in astrocytomas, including losse, s of genetic information on (:hr(nnosomes 9p, 10. 13q. 17p, 19q, 22, and the sex chro- m(isolnes and gains of gene, tic information or gene amplifi- cations (m chronmsnnm 7 [171. In the, 11 falnilial cases, we found 40 CNAs, most affecting regions also perturbed in mise, le, cted cases. (]oinplete loss of chrornosnme 10 was seen in four gliohlastomas. Loss of one or more tumor sup- pressor genes on (:hrolnOSOlne 10 is helieved |o represent a critical step in Ill(,, progre, ssion of intermediate grade glio- mas to high grade GBM: deleti(m nmpping has lot:areal three areas of interest: a telonw, ri(: region on 10 t) and two hn:i (m 10( t [1,% 1{t]. hilerstilia] loss on chromosome 13 (13q21-q33) was observed in two GBM in this study and previously in unselected (;BM 120]; this region does not overlap the retinoblasloma snsceplibil i ty gene (13q14), and likely points to the locati(m of another tumor suppres- sor gene on 13q. CNAs involving chronmsome 19 o(:(:urred in two GBM and two oligodendrogliomas, l.osses on (:hro- mosome 19 are common in oligodendroglionlas [211 and have heen reported in some high-gra(h.' astrocytic gliolnas [22, 231; tile region 19q13.2-q13.3 may harbor an inlpor- rant tumor suppressor gene [24]. The oligodendrogliomas in the present study had s imultaneous losses of l(,)q and lp. t:(mfirrning previous reports 125, 26]. The X andY (:hro- niosnines were ]nst in four tunmrs; iiOlle were glioblast(nnas. It has heen argued that rnonosonw ()f the sex c h r o n i ( i s o n l e s

niav retlect heterogeneity within non-neot)lastic tissue [27, 281 and that h)ss (if Y in somali(: (:ells may tie a fre- quent finding in aging males [281; hi)wever, in ihis study, we fottnd sex chr(inlosome loss (llll',' in younger patients (ages 13. 25, 33, and 38). 13eleti(m of 16p wt;re detected in two ln lnors ; chron los t ln le 16 aberrations have })een de- s(:ribed in ependynmnlas in ( 'hildren [291 hnt are rare ill (:ases (if adult glinnuis. The signifi(:ance of 16tl loss in these falnilial glinmas is unknown, hut (mr results sltggest this aberration may be present in fanfilial tnln(irs more comm(mly than in sporadi(: glicnnas.

The gene en(:o(ling e,t)idermal growth fa('t()r recel)tor (E(;FR) and the MET oncogen(,, Oil c:hronmsonle 7 are am- plifie(i in sporadic gli(nnas as are the GI,I, SAS, and MDM2

o " ~ S on(:ogenes on chronmsonm 12 130-36]. Two ~lm )la,'tomas in this study had aml)lifi(:alions at 7ti11.1-p12, the sit(; of EGFR (7t)11-p13); one had a se(:ond amplification al 7q21.2-(t33 en(:()mf)assing tile ME'I" locus (7(131). Although EGFR and ME'r are like, Iv amplified ill the glioblastomas with these tw() amplifi(:ali(nas, other growth-controllilag geims may be located in these large, amplified regions of (:hronlosonle 7 material. An extra copy (if chromosonw 12 was ohserx:ed in one ltlin()r, and ill anolher there were Ix.,'() distinct sites of amplifi(:ation, L ±12(t13.2~(114 and -:- '12p11-+p12: the ~-+12q13.2~q14 anlpli(.:(nl (;n(:(nn- passes GIA, SAS, and MI)M2, but the + ~-12p11--+p12 ah- mlrmalitv has not been noted previously all(l is a likely sit(; for an (is '`'tlt un(:hara(:terize, d Oli(:()g, e n e .

CGH in Familial Gliomas 81

¢.t3

©

" 5 3

r - '

g t--,

r "

¢.-,

5 ~d3

o ~

° ~

° ~

° ~

° ~

O

c-.,

c..;

¢ ' 4

, , ,m

X

P-,

g

r--t

r- , ~D

1.3

¢--t

.,2,1

¢ . ,

G -

t-. .

¢-,

t ~

g

,"1

,_.q

i

I

I t

+

7 7

I

+ +

I

7

+

I

. -~ .-',~ ,~'~ i ~ L~,-. ~ ~ . ~ I - ~

Some chromosomal abnormalities commonly encoun- tered in sporadic gliomas were not detected in these famil- ial cases. For example, we found no abnormalities of 9p, the site of the tumor suppressor genes encoding p16 INK4A and p15 ~NK4~, nor of 17p, the site of the p53 gene. Muta- tions in these genes, as well as deletions, loss of heterozy- gosity, and structural rearrangements of 9p and 17p, occur frequently in sporadic gliomas [37, 38] and CGH reports to date appear to underestimate occurrence at 17p [16, 39, 40]. The findings in this study contrast with earlier work in which we observed frequent loss of heterozygosity (I,OH) at 9p21 and 17p13 in familial gliomas [41]; these genetic alterations were using polymorphic microsatellite markers. Six samples in the present study [samples #4, #5, #7, #8, #9, and #11) were included in our LOIt analysis and all six displayed I,OH for at least one marke, r on 9p. Several possibilities may account for the discrepancy be- tween I,OH and CGH results. First, CGH will not detect deletions less than 5-10 Mb in length. Second, CGH using DNA derived from the paraffin-embedded tissue utilized in this study is less sensitive than CGH performed using DNA from frozen samples. Finally, LOH at 9p and 17p due to mitotic recombination [421 will not alter the gene copy number and, as a consequence, will escape detection by CGH.

In conclusion, we have identified novel CNAs in these 11 familial gliomas [ - 1 6 p and + + 12p(11-p12)]. but none that readily explain the inheritance of these tumors. The lesions we identified are largely consistent with those ob- served previously by CGH or karyotypic analyses of sporadic cases. Gliomas exhibit heterogeneous clinical, pathologic, and nmlecular characteristics [31, 43], but cases with dif- ferent histology can still share (:lasses of genetic abnormal- ities such as p53 mutation [16], and gliomas that appear similar microscopically segregate into subsets with dis- tinct molecular profiles [44, 45]. We have proceeded on the assumption that. within a family, the same he, ritable defect could give rise to histologically dissimilar glial tu- mors, given that each may arise only folh)wing subsequent acquired nmtations peculiar Io that histological subtype. However, the large number of frequent genetic aberrations that occur in tamilial tumors of different histologies com- plicates the task of identifying cytogenetic abnormalities of etiologic significance in tumor tissue. To better exploit CGt/ as a prelude to targeted linkage analysis in glioma families, a strategy proven successful in P(,'utz-Jeghers syn- drome [12J. future analyses should focus on families (like family A in this study) in which the inheritance pattern is striking (i.e., many affected members, early age of onset, uniform pathology), on families in which glioma has oc- curred in association with other cancers, and on tumors where there are few aberrations (i.e., low grad(; tumors). If tissue availability permits, such studies should employ complementary technical approaches (i.e., LOlt analysis, fluorescence in situ hybridization) to minimize the possi- bility of overlooking genetic alterations of interest and maximize the types of mutations that (:an be screened.

This work was funded in part by NIH grants CA13525, CA52(189, CA61147, and CA64898. the Quarterly Research Committo.e.

82 A. Patel e,t al.

UCSF. and tile DepartInent of Laboratory M(;dicine. UCSI". We also thank Joe Gray ff)r put t ing our respect ive groups in contact for this stucty.

REFERENCES

1. Bishop JM (1991): Mole(:ular thelnes in on(:c~gene, sis. Cell 64:235-248.

2. l.i FP (1990]: Familial (:ancer svn(lromes anti chlsters. Curr Probl Cancer 14:73-114.

3. Malkin D. l,i lq' . Strong I,C. l. 'raumeni ll"l. Nelson CE. Kim DH. Kassel l, (;ryka MA. Bischoff FZ. Tainsky MA. l.'rien(l SH (199(I): (;ernl line p53 muta t ions in a familial synd rome of breast can(:er, sarcomas, alld other lleopl~ssms. St:ien(:e 250:1233-1238.

4. Xu G. () 'Ccmnell P. Viskoc:hil D. ( ;awthon R. Robertson M. Culv(;r M. Dunn D. S tewms I, Gesteland R. White R. Weiss R (1990): The neurof ibromatos is type 1 gene en(:odes a pr()tein related to GAP. ( 'ell 62:5,99--008.

5. Murrell l, ' l 'rofatter l. Rutter M. Cutone S, St()tler C. Ratter J. Long K. Turner A. l)e, aven L. Buckler A. Mc:Cormick MK (1995): A 5(}0-kilobase region conta ining tile tuber(ms sr:h:ro- sis locus [TSC1) in a 1.7-megabase YAC and (:(ismid contig. (;enolni(:s 25:59-65.

6. Wit:neck(,' R, Konig A. l)eClue JE (1995): htentifit:aticm of tuberin, the tuberous sc:lerosis-2 produ(:t. Tuber in tlt)SSe, sses spe,(:ific Rap l ( ;AP activity. J Biol Chem 270:16409-10414.

7. lkizler Y. van Mewd DI. Ramsay I)A. Ab(lallah GI,, Allaster RM. Ma(xtonal(I DR. Cavenee WK. Cairn(ross J(; [19,'42): (;lit)- mas in families. (]an J Nel,rol Sc:i 1(-t:4(,t2-497.

8. Eng (], l~oncler BA (1993): The role of gene, tnutat ions in the genesis of familial c:anc:ers. FASEB J 7:910-919.

9. Kall ioniemi A. Kallicmiemi (_)P. Sudar I). Rutovitz D, (_;rav JW, Waldman 1", Pinkel D {1,q92]: Comparat ive genomi(: hvtlr idization for nmle(:ular (:ytogeneti(: analysis of solid tumors. S(:iellce 258:818-821.

10. Kalli,miemi O1', Kallioniemi A. Piper ]. lsola ], Waldumn FM. Gray JW, Pinkel D (19941: Optinlizing (:onq)arative genomic hybridization for analysis of I)NA sequen(:e copy ilumbe, r (:hange.s in solid tun)ors. Genes (Sm)mosom Cancer 10:231-243.

11. Isola l, l.)eVries S. Chu 1,. Ghazvini S, Walchnan F [19{14]: Analysis of (:hange.s of DNA sequence (:opy number by i:om- parative genomic hybr idizat ion in archival t)araffin-em.t)ed - tied tumor samples . Am J l 'atho[ 145:13(11-1308.

12. l lemminki A, Toml inson I, Markie D. Jarvinen I I. S is tonen P. Bjorkvist A-M. Knuuti la S, Salovaara R, Bodme.r W. Shibata D, de. la Chappe.lle A, Aaltone.n I,A (19.97): Localization of a suscept ibi l i ty lot:us for Peutz-Jeghers syndronm to 19p using c:tnnparative genolnic: hyllr idization and targeted linkage, analysis. Nat (;enet 15:87-90.

13. Bentz M. Bergel'heinl I, ISR. Li C. ]oos S. Wert)er (;iX. Baudis M. (_;narra l. Merino MI. Zbar B. l , inehan WM. Lichter P (19901: Chromosonle i lnbalances in papil lary renal cell (arc:i- noma an(l first (:ytogenetic data of familial (:ases analvzed bv conlparat ive gentmlic hybridizat ion. Cytogenet Celi (,(;net 75:17-21.

14. Kleihues P. Burger PC. Schei thauer BW (19931: llistoh~git:al typing of tun)ours of tile central nervous system. In: World t leal th O ro(o tll'lZatmn,' International I tistological Classifit:ati(m of Tumotlrs . 2rid Ed. Springer-Verlag, Berlila.

15. Mohapatra G, Kim DII, Fmmrstein 13G (1995): Dete(:tion c,f mul t ip le gains and losses of genetic material in ten glioma cell l ines t) 5' COlnparative gel)omit: hybridizat ion. Gene.s ( ]hromosom Cant:er 13:86-93.

1(';. Mohaptra G. Bollen AW, Kim l)tt. l ,amborn K, Moore D, l. 'euerstein B(; (in press]: Genetic analvsis of gl ioblastoma mult i fornm provides ev idence for subgrclut~s wi th in the grade. ( ,enes ( ;hrolnosmll Callcer.

17. Coll ins VP. Janles CD (19(.):3): ( ;ene ancl ( :hromosomal alter- at ions assoc:iated with the developme, nt of human gliomas. FASEB J 7:926-930.

18. Karlhom AE. James CD, Boethius J, C, avenee WK. Collins VP, Nordenskjold M. Larsson C (19.q3): l,oss t~f heterozygosi ty in malignant gl iomas involves at least three dist inct regions (ill chrclmosome 10. Hum Gelmt 92:169-174.

19. Rasheed BK, Fuller (;N, Fr iedman All. Bigner Dl3. Bigner Slt (19(`)2): 1,oss of heterozygosi ty for 10q loci in human gliomas. (]elles Chrolm)SOm (;ance, r 5:75-82.

20. Kim Dtl. Mohal)atra (;, l~,olhm A, \Vahhnar~ I"M, l, 'euerstein B(; (19,(t5): Chromosomal ahnornmli t ies in glioblastoma mul- t iforme tun)ors and gliolna cell l ines detected by r:ornparative gencimic hybx'idization, lnt ] Cancer 60:812-819.

21. Ransom DT, Rithmd SR, Kimlnel 1)W. Moertel CA, Dahl RI, St:heithauer BW, Kelly Pl. lmlkins RB (1(`192): Cytogem;tic: and h)ss of heterozygosi ty s tudies in ependylllOlllas, pilocytic: astro(:vtonlas, and oligocte.l)drr~gliomas. Genes Chrom(Isom (;am:er 5:348-356.

22. Ritlan(l SR. (;anju V. ]enkins RB {1995): Region-spe,(:ific h)ss of hetert)zygosity on chromos(ram 1,9 is related h) the mor- phoh)gic: type of human glioma. (;eta0,s (Sm)mosom (~anc:er 12: 277-282.

23. yon l)eimling A, Bender B. lahnke R. Waha A. Kraus I. Albrecht S. Welle, nreuther R. I"assbe, nder 1.. Nagel J. Menon AG. Louis I.)N. Lenartz D. S(:hramn) I. Vv'iesller ()l) (1994): l,oc:i assoc:i- ated with malignant l)rogression in astm(:vtomas: a candidate (m chrcm)osome 19q. Canc:er Res 54:13(37-1401.

24. Yong \VH. Chou D. Ueki K. Harsh GR. yon l)ein)ling A. (.;usella ll". Mohrtmweist :r t I\V. Louis DN (19951: Chr(mm- some 19q (lelelions in hunlall glicm)as overlap teh)nmri(: to 111(`1S219 and may target a 425 kll region (:m~tromeric: to I319S112. J Ne.uropathol Exl) Nmu'ol 54:622-626.

25. Kraus JA, Kcmpnmnn l, Kaskel P, Maintz 11. Branclner S. St:hi')rain J, Louis DN, Wiestler (.)D. v . n Deimling A {1(`195): Shared a l M i c losses on c:hromosomes 1 I) alld 19( l suggest a t:Ollllnoll origin of (~lig()(lendroglionm and oligoastro(:ytoma, l Ne, ur(q~athol Exp Neurol 54:(`11-(,)5.

26. l~,eit'enberger J, Reifenberger (;. l,iu I,. James CD. Wechsh:r W. ( ;ollins VP (1 (,194): Mole(:ular gene, tic atudysis of ol igodendro- glial tumors shm:-,s prefl:rential allelic dele, ticms on 19q and 111. Am J Pathol 145:1175-119(I.

27. l le(:ht BK. Tur(: Card B, (;hatel M. Pa(tuis P. (;ioal)ni l. Attias R. ( ;audrav 1'. Hecht 1" (1995): Cytogenetit:s of malignant glio- mas. I1. The sex chr(mll)son)es with reli;rem:e It) X isodisolny all(] the role of numeri(:al X/Y challges. ( ' ,ancer GO, llel Cyt()go- rmt 84:,¢-1-14.

28. Yalna(ta K, Kasanla M. Kon(h)T . Shin(tara N. Yoshi(ika M (1,q94): Chromosome s tudies ill 7(I brain tumors with spe(:ial attel)tit)ll It) sex ¢:hroTIlosome loss and s ingle autosolna l tri- sorer . Can(:(,'/ (_]tmet Cyt(~genet 73:46-52.

29. Neumalm E. Kalousek I_)K, Nol'nlan MG. Steinbok P. Cochrane DD, (;clddard K (19931: Cytogeneti(: analysis, of 109 llediatril: (:enlral nerv(ms system ttmlors. (]an(:er (;enel (Jyl(Jgermt 71:40-49.

30. R(:ifenberger G. Liu I,. lc:himura K, Sc:hmidt EE. C(dlins VI' (19t.13): Ampli f ica t ion and overexpressicm of the MI)M2 gene in a st,bset of hun)an nlalignant gl iomas wil lmut p53 rnuta- firms. Callcer Rt~S 53:273(i-2739.

31. Collil)s VP (1993}: Ampl i f ied g(me, s in human gliomas. Semin (;anc:cr Biol 4:27-32.

32. Muh:ris M. Ahm;ida A, l)utri l laux AM. Pruc:hon E. Vega F. l)elatlre IY. Poisson M. MalfcLv B. Dutrillaux B (19941: ()nt:o- g e n t alnl'difit:atioll ill hllrnall gliomas: il lllOlOctllar cytoge- neti(: analysis. On(:ogcme !t:2717-2722.

33. Reitiu)tnwger (;, Reifenberger I, lchilnura K. Meltzer PS. Collins VP (1994): Amplif icat ion of mult iple genes from (:hromosolnal region 12q13-14 in human malignant glioinas: prelinfinary

C(;H in Fami l i a l G l i o m a s 83

mapping of the amplicons shows preferential involvement of CDK4, SAS, and MI)M2. Cancer Res 54:4299-4303.

:34. Fischer U, Muller I-tW. Sattler JP, Feiden K. Zang KD, Mecse E {1995): Amplific:ation of the MET gene in glioma. Genes Chromosom Cancer 12:6:3-65.

:35. Wullich B, Muller t/W, l:ischer L), Zang KD, Meese E (1993}: Amplified met gene linked to double minutes in human glio- blastoma. Eur J Cancer 29A: 1991-1995.

36. Wullich B, Sattler HP, Fist:her U. Meese E (1994): Two inde- pench,nt amplification events on chromosome 7 in glioma: amplification of the epidermal growth factor receptor gene and amplification of the oncngene MET. Ant |cancer Res 14: 5 7 7 - 5 7 9 .

37. Schmidt EE, h:himura K, Reifcnberge, r G, Collins VP (1994): CDKN2 (pl6/MTS1) gene cteletion or CDK4 amplification occurs in the majority of glioblastomas. Cancer Res 54:6321- 6324.

:38. He J, Allen JR, Collins VP. Allalunis Turner MJ, Godbout R, I)ay RS, James CD (1994): CDK4 amplification is an alterna- live mechanism to p16 gene hontozygous d(:lelion in glioma cell lines. Cancer Res 54:5804-5807.

39. Weber RG, Sommer C, Albert FK, Keissling M, Cromer T (1996}: Clinically distinct subgroups of glioblastoma multi-

forme stuctied by comparative genomic hybridization. Lab Invest 74:108-116.

40. Schlegel J, Scherthan H, Arens N, Stumm G, Keissling M (1,(}96): Detection of cmnplex genetic alterations in human glioblastolna mult i ibrme using comparative genomic hybrid- ization. J Neuropath Exp Neuro] 55:81-87.

41. Wailing CL wm Meyel DJ, Ramsay DA, Macdonaht DR, Cairn- cross ]G (1995): l,oss of heterozygosity analysis of chromo- somes 9, 10 and 17 in gliomas in families. Can J Neurol Sci 22:17-21.

42. lames CD, Carlbom E, Nnrdenskiold M, Collins VP, Cavenee WK (1989): Mitotic recombination of chromosome 17 in astrocytomas. Proc Nail Acad Sci USA 86:2858-2862.

43. Kleihues P, Soylemezoglu t,', Schatible B. Scheithauer BW, Burger PC (1995): Histopathology, classification, and grading of gliomas. Glia 15:211-221.

44. IJeki K. Ono Y, Henson JW, Efirct JT, yon De|ruling A, Louis DN (1996): CDKN2/p16 or RB alterations occur in the major- ity of glioblastomas and are inw',rsely correlated. Cancer Res 5{i:150- 15:3.

45. von l)eimling A, w,n Ammon K, Schocnfeht D, Wiestler OD. Seizinger BR, I,otnis DN (1`(193): Subse, ts of glioblastonm mul- tiforme defined by molecular genetic anah'sis. Brain Pathol 3:19-26.