Isochromosome 17q demonstrated by interphase fluorescence in situ hybridization in primitive neuroectodermal tumors of the central nervous system

GENES, CHROMOSOMES & CANCER 1495-96 (1995)

lsochromosome 17q Demonstrated by Interphase Fluorescence In Situ Hybridization in Primitive Neuroectodermal Tumors of the Central Nervous System Jaclyn A. Biegel, Lucy B. Rorke, Anna J. Jans, Leslie N. Sutton, and Annette H. Parmiter

Divisions of Human Genetrcs and Molecular Biology (J A B , A H P) and Neurology (A J J ) and Departments of Pathology (L B R ) and Neurosurgery (L N.S ), The Children's Hospital of Philadelphia. and Department of Pediatrics, University of Pennsylvania School of Medicine (J A B ), Philadelphia, Pennsylvania

We previously reported an i( l7q) as a non-random finding in childhood primitive neuroectodermal tumors (PNETs) of the central nervous system. In the present study, we describe a two-color interphase fluorescence in situ hybridization (FISH) assay for detection of chromosome 17 abnormalities in tumors. Thirty-four PNETs were analyzed by FISH with a series of chromosome 17-specific probes which map to 17p13.3-17q25. The results from the FISH assay were then compared to the karyotypes prepared from the tumors. Ten of the 34 cases demonstrated an i( I7q) by FISH and standard cytogenetics. Two PNETs were shown to have an i( I7q) by FISH alone, and three additional tumors had deletions of I7p. Thus, a total of I 5 of 34 (44%) of the PNETs in this series had a deletion of I7p. This study confirms and extends our previous reports that an i( I7q) is the most common cytogenetic abnormality in PNETs. The interphase FISH assay which we employed will have clinical utility for diagnosis of children with malignant brain tumors, and it may be used for identification of tumors with 17p deletions for molecular studies aimed at identifying disease genes. Genes Chrornosom Cancer /4:85-96 (1995). 0 1995 Wiley-Liss, Inc.

INTRODUCTION

Primitive neuroectodermal tumor (PNET), the prototype of which is medulloblastoma, is the most frequent malignant tumor of the central nervous system (CNS) in children. PNETs are thought to arise from a neural stem cell which has the capacity to differentiate along neuronal, glial, ependymal, and/or mesenchymal pathways (Molenaar et al., 1989). Little is known regarding the molecular genetic events which lead to the development of CNS PNETs. Cytogenetic studies of a variety of pediatric and adult malignancies have been the ba- sis for molecular studies which ultimately led to the identification oi disease genes. Toward this end, we have shown that an i( 17q) is the most frequent cytogenetic change in pediatric PNETs (Biegel et al., 1989). T h e i(17q) is not specific for PNETs (Mertens et al., 1994). However, we have shown that an i(17q) is a primary cytogenetic event in PNETs, unlike many of the other hematologic and solid tumors in which it has been observed. T h e cytogenetic analysis of PNETs has therefore be- come a useful adjunct to the histologic diagnosis of malignant pediatric brain tumors. In the past, cytogenetic studies of PNETs have been limited by the fact that it is difficult to obtain high-quality metaphase preparations from the biopsy specimens. Interphase fluorescence in situ hybridization (FISH) therefore offers an additional means of

identifying PNETs with an i(17q) (Arnoldus et al., 1991a, 1992; Matsumura et al., 1992) which will be useful in a clinical diagnostic setting.

T h e non-random occurrence of an i(17q) in PNETs suggests that the loss of sequences on 17p, or the gain of sequences on 17q, plays a role in the development of these tumors. Restriction length polymorphism (RFLP) analysis of Southern blots and polymerase chain reaction-based microsatellite analysis of tumor and normal tissues have confirmed the deletion of sequences from 17p in PNETs (James et al., 1990; Raffel et al., 1990, 1993; Cogen, 1991; Thomas and Raffel, 1991; Bie- gel et al., 1992; Albrecht et al., 1994; McDonald et al., 1994). These cytogenetic and molecular studies implicate the presence of a tumor suppressor gene on 17p which may play a role in tumor initiation or progression. TF53 was considered a candidate gene for PNETs, based on its map location to 17~13.1 and the reported high frequency of mutations in adult malignant gliomas (Chung et al., 1991). However, mutations in TP53, specifically in exons 4 to 8, have been found in a very limited number of PNETs (Saylors et al., 1991; Biegel et

Received February 28,1995; accepted April 24, 1995. Address reprint requests to Jaclyn A. Biegel, Ph.D., Division of

Human Genetics and Molecular Biology, T h e Children's Hospital of Philadelphia, 34th Street and Civic Center Boulevard, Philadel- phia, PA 19104, U S A .

0 I995 Wilcy-Us, Inc.

86 BIEGEL ET AL.

al., 1992; Raffel et al., 1993; Felix et al., 1995). Furthermore, the identification of several tumors with deletions involving the more distal D17S34 locus suggested that a P N E T suppressor gene mapped to the distal region of 17~13.3. Interphase FISH with chromosome 17-specific probes pro- vides an additional means of identifying tumors with 17p deletions (Matsumura e t al., 1992), which will be useful in positional cloning strategies designed to identify a PNET-related gene.

We now report on the largest series of newly diagnosed PNETs subjected to cytogenetic analysis. We have shown that an i( 17q) is the most common abnormality in CNS PNETs, and we demon- strate that a two-color interphase FISH assay with chromosome 17-specific probes is a sensitive and specific means of detecting an i(17q) in these tumors.

MATERIALS AND METHODS

Clinical Samples

Central nervous system PNETs were obtained from children having surgery at T h e Children's Hospital of Philadelphia (24 cases) or from outside hospitals (10 cases). Three of the latter 10 specimens (92-04-110, 93-01-A047, and 94-07-POO1) were obtained through the Cooperative Human Tissue Network. Tumors were classified according to the revised World Health Organization classification for pediatric brain tumors (Rorke et al., 1985). Specimens were collected over a 3-year pe- riod from 1992 to 1994.

T h e clinical information, pathologic diagnosis, and cytogenetic data for the 34 P N E T patients are shown in Table 1. There were 20 male and 14 female patients. T h e median age at diagnosis was 6 years (range, 2 to 16 years). Primary biopsy specimens from newly diagnosed pediatric patients were obtained in 32 of the 34 cases. T h e first biopsy specimen from case 93-32 demonstrated a marker chromosome in 2 cells that did not appear to involve chromosome 17, as previously reported (case 2, Biegel et al., 1989). A second biopsy specimen was obtained for this study from the patient's recurrent tumor, which was metastatic to the abdomen. A second recurrent tumor (case 94-92) was from a 10-year-old girl who had received radiation and chemotherapy before the tumor specimen was obtained.

Preparation of Tumor Specimens

Interphase nuclei and metaphase chromosomes were isolated from primary tumor specimens ac-

cording to established methods (Biegel et al., 1989). Direct preparations and 24-hour pellets were used for FISH studies as well as for preparation of karyotypes. In some cases, karyotypes were prepared from 3-14-day cultures. For standard cytogenetic studies, 20 metaphase cells were examined whenever possible. Karyotypes were described according to the Guidelines for Cancer Cytogenetics, a supplement to the ISCN (1991), with the exception of the designation for the i( 17q). Abnormalities not present in every cell are shown in parentheses. Direct preparations and 24- hour pellets were stored at 4°C until needed for the FISH studies.

Fluorescence In Situ Hybridization

A series of probes which map to chromosome 17 were hybridized with the chromosome 17 a-satellite probe as an internal control. Probes labeled with biotin were used with the a-satellite probe labeled with digoxigenin, or vice versa, so that they could be detected simultaneously. T h e probes map to chromosome 17 in the following order:

(17~11.2) - centromere - D17S33 (17ql l ) - HERZ/ NEU (17q11.2-ql2) - D17S7.5 (17q25). Biotin- or digoxigenin-labeled chromosome 17 a-satellite, D17S34, TP53, D17S122 (CMTl) , and HERZINEU probes were obtained from Oncor, Inc. (Gaithers- burg, MD). Cosmid probes ICRFc105D04178 (D17S33) and ICRFc105E04183 (D17S75) were obtained from the reference library, ICRF (Le- hrach et al., 1990), and were labeled with biotin- 14-dUTP (Boehringer Mannheim, Indianapolis, IN) by nick translation (Lichter et al., 1988). T h e commercially available i( 17q) probe (Oncor, Inc., Gaithersburg, MD) consists of a biotin-labeled D17S122 probe (17~11.2) and a digoxigenin-labeled myeloperoxidase (MPO) probe (17q21.3-qZ3) and was not used with the a-satellite probe.

In situ hybridization was carried out as described by Kuwano et al. (1991), with minor modifications. Briefly, the a-satellite probe was added to the chromosome 17 test probe hybridization mix im- mediately before the probes were placed on slides. Following an overnight hybridization, slides were washed two times in 50% forrnamide/Zx SSC for 10 minutes each at 45"C, twice in l x SSC, and once in 0.1 X SSC for 5 minutes each at 45"C, and once in 0.1 X SSC for 5 minutes at room tempera- ture. Probes were detected simultaneously with FITC-avidin/antidigoxigenin rhodamine (Oncor) and stained with DAPUantifade according to the manufacturers' instructions. Slides were analyzed

D17S34 (17~13.3) - TP53 (17~13.1) - D17S122

FISH ASSAY FOR { ( I 7q) IN PNETs 87

TABLE I. Clinical Features and Cytogenetic Findings From 34 PNETs

Case no. AgelSex Diagnosis Cells Culture

Cytogenetic findings analyzed time (days)

92- I29

92- I35

92- I 39

92- I42

92- I48

92- I49

92- I55

92- I 56

92- I62

93-18

93-23

93-32

93-4 I

93-60

93-64 93-88

93- I50 94-6

94- I5

94- I7

94- I9

I OIM

81M

7lM

I UM

6lF

I4lF

6lF

UM

81M

6lM

I UF

I6lF

4lF

UF

UM 81M

81F I OIM

81M

I OIF

UM

PNET with neuronal differentiation


PNET with neuronal and glial differentiation

and glial differentiation

PNET with glial differentiation


PNET with neuronal





PNET, desmoplastic, with neuronal differentiation

PNET with glial differentiation; metastatic to abdomen

PNET. desmoplastic with neuronal and glial differentiation

PNET

PNET PNET with neuronal

differentiation PNET. NOS' PNET. desmoplastic,

NOS

PNET with neuronal, glial, and chondroid differentiation



Tetraploid.add( I7p)

46,XY,t( I ;6)(p2 I ;q I I -I 3)

46,XY,add( I I)(pl I),i( 17q)l 47,XY,+7,add( I I)(pl I), i( I 7 4

N D ~

Complex

Diploid and tetraploid cells with add(l)(p36).- lo,(- 15). (- I6).(del( I7p)), + I8,add(22) (91 l),t marker

Hypodiploid, -6

I abnormal triploid cell

74-90,XXYY,- 3, t 7, + 7, -8,- I I ,- 14,- 14,- 15, - I6,i( I7q)x2, + 2 markers

82-92,XXY?Y,-2, + 7 . 4 .

i( l7q)x2, + markers -10.-11,-13.

ND

45-46.XX, complex, ?del( I7p)

46,XX, + 6, + 8, + 8, + I3,i(l7q),+ 19

70-73.XX. + 3,-4, + 6. + 74 + 714 + 7), + 9, - 14.+ 19,-20,(-22)

ND ND

Diploid, + I I 93-96,XXYY,- 3,- 3, + 7,

-8,-8,i(I Iq)x2,+ 13, - 174 I7q), + markers

46,XY,del( 3q),de1(6q)l near-tetraploid.del(3q). del(6q). + markers

90.XXX.-X.i( 17q)x2,-21

47,XY. + I9

2 DP'

29 DP,7

63 DP

2 DP

84 DP,4

5 I

13 DP.4

31 3

25 DP,2

7

18

20

14

DP

DP,5

12 3 19 DP.5

67 DP,4

61 DP,2

6 DP

(continued)

88 BIEGEL ET AL.

TABLE I. (Continued) ~~ ~~

Cells Culture Case no. AgelSex Diagnosis Cytogenetic findings analyzed time (days)

94-40 94-59 94-69

94-70

94-72

94-82

94-92

94-101

94- I06

94- I08

92-041 I0

93-0 I -A047 94-07-PO0 I

3/M 2IM

I OIM

4lF

6lF

61M

I OIF

6lM

8/F

I6lM

4/M

5/F 3/M

PNET PNET PNET with glial

differentiation PNET with neuronal

and glial differentiation

PNET with neuronal, glial, and photoreceptor (retinal S) differentiation

PNET with extensive neuronal differentiation and minimal glial differentiation

Recurrent PNET with minimal glial differentiation

neuronal, glial, and retinal S differentiation

neuronal, glial, and muscle differentiation



PNET, desmoplastic PNET

PNET with

PNET with

ND

Multiple i( I7q) 6 DP 46,XY,i( 17q),dmin 58 2

Diploid, tetraploid with 13 I ,2 i( 174

add(6)(p22),de1(6q) 46.XX,de1(5)(q3), 30 DP

Diploid, add(3)(q26)

Complex

ND

2 cells with + I I

45,XY,-8,i( I7q)

N D

ND 46,XY 35 I

12 I

20 DP

4 DP

29 I

"DP = direct preparation. bND = no metaphases were obtained, or karyotypes from 7 to 14 day cultures were normal. 'NOS = not otherwise specified (Rorke et al., 1985).

with a Zeiss Axiophot microscope equipped with a dual or triple band pass filter set. A minimum of 50 interphase nuclei were examined for each probe. Images were captured and printed with an Oncor imaging system (Oncor). T h e specific probes used for each patient varied according to the amount of material obtained, and the reagents available at the time when these studies were initiated. In cases with limited material, we evaluated loss or gain, respectively, at the D17S34 and D17S7.5 loci first, and then expanded the studies with additional probes as needed. In 31 of the 34 cases, we were

able to analyze at least one probe each from the short and long arms of chromosome 17.

RESULTS

Cytogenetic Studies

As shown in Table 1, karyotypes were obtained for 24 of the 34 cases. For the remaining 10 cases, there were no metaphase cells for analysis, or normal karyotypes were obtained after 1 to 2 weeks in culture and were not thought to be informative.

One tumor (case 94-07-POO1) had an apparently

FISH ASSAY FOR i(17q) IN PNETs 89

normal karyotvpe, 46,XY, whereas 23 cases had abnormal karyotypes. The karyotypes demonstrated numerical and structural changes, the latter consisting primarily of unbalanced translocations and deletions. T h e most common abnormalities among these PNETs were structural changes of chromosome 17, which were observed in 13 of the 24 (54%) cases. T e n of these 13 cases contained one or more copies of an i(17q) (Fig. 1). T h e second most frequent abnormality was additional copies of chromosome 7, which were seen in 5 cases. Four of the 5 cases with extra chromosomes 7 were PNETs with an i(17q).

Abnormalities of three other chromosomes, including chromosomes 6, 8, and 1 1 , were observed in several PNETs in this study, either as a solitary change or in association with additional cytogenetic findings. Numerical or structural changes of chromosomes 6 and 11 were present in 6 cases each. Monosomy (92-155), deletion (94-15 and 94-72), or translocation (92-135 and 94-72) involving chromosome 6 was observed in a total of 4 cases, whereas extra copies of chromosome 6 were present in 2 cases (93-41 and 93-60). For chromosome 1 1 , trisomy was seen in 2 cases (93-150 and 94-106), loss of 1 1 was observed in 2 cases (92-162 and 93-18), and structural abnormalities of chromosome 1 1 were found in 2 cases (92-139 and 94-6). Loss of chromosome 8 was observed in 4 cases (92-162, 93-18, 94-6, and 94-108); however, case 93-41 had 2 extra copies of chromosome 8.

Previous studies have suggested that structural or numerical abnormalities of chromosome 1 or 22 and double minute chromosomes are frequent in PNETs (Bigner et al., 1988; Griffin et al., 1988; Vagner-Capodano et al., 1989). In the present study, double minute chromosomes were present in 1 case (94-59), translocations involving chromosome 1 were observed in 2 cases (92-135 and 92- 149), and a translocation (92-149) or monosomy (93-60) of chromosome 22 was seen in 1 case each.

FISH Analysis of Chromosome 17 in PNETs

In this two-color FISH assay, we simultaneously hybridized a probe for the centromeric region of chromosome 17 with a probe which mapped to 17p or 17q in a successive series of experiments designed to evaluate the integrity of chromosome 17 in the tumor. This strategy allowed us to control for variations in ploidy commonly seen in these tumors, and it was not markedly affected by the ten- dency of the chromosome 17 centromeres to associate in the interphase nucleus. A summary of the FISH results is shown in Table 2. The data pre-

sented indicate the modal number of signals seen in 50-200 interphase nuclei for the chromosome 17 a-satellite probe compared to the test probe. As expected, there was cell-to-cell variation for many of the probes, which was more pronounced in the tetraploid tumors (data not shown). Analysis of additional cells in these cases did not reduce the variability in results; however, we were able to inter- pret whether there was loss or gain of a probe in most cases. A comparison of the cytogenetic and FISH interpretations of the status of chromosome 17 for the 34 PNETs is shown in Table 3.

T h e chromosome 17 FISH results were consistent with the cytogenetic findings in 9 of the 10 tumors with apparently normal chromosomes 17. In the remaining case (92-155), there was an increase in the number of signals for the D17S75 probe, which maps to 17q25, in approximately half of the cells. We were able to analyze only 5 metaphase cells from this P N E T and thus were not able to determine whether there was an unbalanced translocation involving the distal long arm of 17, as suggested by the FISH experiments.

Ten of the 24 cases for which we obtained karyotypes had one or more copies of an i(17q). Rep- resentative FISH results for case 94-70 are shown in Figure 2. In all 10 cases, the interphase FISH analysis was consistent with the loss of the entire 17 p arm. In 8 of the 10 cases, a relative increase in the number of signals for each of the 17q arm probes was also observed. Cases 92-139 and 93-41 each had 1 normal 17 and 1 i(17q). However, for case 92-139 there were only 2 copies of the HERZINE U locus instead of the expected 3 copies. In case 93-41, there were 2 copies of the D17S33 locus and one population of cells with 2 copies of the HER2INEU locus. A second population of cells had 3 copies of the HER2 locus. These results may have been due to a decreased hybridization efficiency for the D17S33 or HER2 probe compared to the a-satellite probe; alternatively, these data suggest that there are different 17 chromosome breakpoints in tumors with an apparent i(17q).

Four tumors had complex karyotypes with suspected structural abnormalities of chromosome arm 17p. Case 92-129 was near-tetraploid and, in 2 karyotypes, had a derivative chromosome 17 which we suspected was an i( 17q), but we could not confirm this. T h e interphase FISH analysis demonstrated an increase in signal number for the HERZI NEU and D17S75 cosmids compared to the centromere probe, but we did not see a concomitant decrease in the number of signals for the 3 p-arm probes examined. In fact, in some cases it

Figu

re I

. R

epre

sent

ativ

e ka

ryot

ype

from

PN

ET

94-5

9, d

emon

stra

ting

the

i( I7

q) a

nd d

oubl

e m

inut

e ch

rom

osom

es as

the

only

chr

omos

omal

abn

orm

aliti

es.

FfSH ASSAY FOR i(l7q) IN PNETs

TABLE 2. FISH Results for 34 PNETs

91

Chromosome arm I7p loci Chromosome arm 17q loci i( 174

D I7534 TP53 Dl75122 D17S33 HERZlNEU D17S75 D 17s I22MPO Case no. (17~13.3) (17~13.1) (17~1 1.2) (17ql I ) (17ql I.2-ql2) (17q25) (17pl 1.2,17q21.3-q23)

92- I29 92- I 35 92- I 39 92- I42 92- I48 92- I49 92- I55 92- I 56 92- I62 93- I8 93-23 93-32 93-4 I 93-60 93-64 93-88 93- I50 94-6 94- I5 94- I 7 94- I9 94-40 94-59 94-69 94-70 94-72 94-82 94-92 94-101 94- I06 94- I08 92-04-1 I0 93-0 I -A047 94-07-Pool

3:4. 44” 2 2 2 1 2 1 3 2 2 2 2:2

33.44 4 3 5 3

34 ,44 2 2 21 3 3 2 2 2: 2

3 2 5 5

32. 4:2 2 2 2 1 2: I

3:2, 4 2 2 1 4 4 2 2 3 2 6:4 2:2 2: I 5:4 2 2 22

3:4 33 2 2

2 2 3 2

3:3, 4 4 3:3, 4 4 4 3

33 4 3 2 2 2 2

33

3 2 , 4 2

22 .44

2 2 33

4 5

c<q

2 2

4 6

2: 3

3 4

2 2

3 4 2 2 2 2 2 1 33 2 2 2 2 4 4

c<q 33,3:4,44

22 ,23 33 2 2

c<qb

45, 5:5 3 4

35 25 .26

2 3 2 2

22, 2:3 33 .44

26 .27

3:3

2: 3

2.2

4 6

2 3

4 7 2 3

23, 2:4 2:2

32 ,33 7:lO 2 2 2 3

22

2 2 2 4 5 5 2 6 2 2

1:3

The modal number of signals for the a-sate1lite:test probe is shown for each l7p or 17q locus examined. The tatio of the D I75 I 22 signals to the MPO signals is indicated for the i( 179) probe. bA modal number could not be established; however, there were more signals for the q-arm probe than the centromere (a-satellite) probe.

appeared that there may have been an increase in signal number for some cells with the D17S34 and ‘IF53 probes. Results with the i(17q) probe were consistent with the presence of 2 normal chromosomes 17 and 1 or 2 copies of an i( 17q). Overall, we did not consider these results to be consistent enough to categorize the tumor as having an i( 17q), although there clearly is an extra copy of 17q in these cells.

Case 92-149 also had a complicated karyotype with a number of structural abnormalities. Some cells appeared to have a deletion in 17p. T h e FISH results for D17S34 and HER2INEUdid not confirm the cytogenetic findings, however. One possible

explanation for these results is that the chromosome 17 deletion is interstitial and does not involve the D17S34 locus. Another likely hypothesis is that the chromosome arm 17p material is contained in a marker chromosome which we could not identify in the karyotypes.

Two PNETs were recurrent tumors from patients who had been treated with radiation and chemotherapy. T h e karyotypes from these tumors demonstrated abnormalities of almost every chromosome pair, including chromosome 17. In case 93-32, some of the karyotypes contained a deletion of the entire 17p arm, and yet FISH results with 3 probes from 17p were consistent with 2 copies of

92 BIEGEL ET AL.

TABLE 3. Comparison of Cytogenetic and FISH Results for Chromosome I7

Cytogenetic determination FISH interpretation Case no. of chromosome I7 of chromosome 17

92- I29 92- I35 92- I39 92- I42

92- I48 92- I49 92- I55

92- I56 92- I62 93- I 8 93-23 93-32 93-4 I 93-60 93-64 93-88 93- I50 94-6 94- I 5 94- I 7 94- I 9 94-40 94-59 94-69 94-70 94-72 94-82 94-92 94-101 94- I06 94- I 08 92-04-1 I0

93-0 I -A047 94-07-PO0 I

Normal I ~ s , add( I7p) 17.17 17, i(17q) ND"

ND 17, ? del( I7p) 17,17

ND 17, I7,i( I7q),i( I7q) 17, I7,i( I7q),i( I7q) ND ? del( I7p) 17, i(17q) 17.17, I7 ND ND 17,17 17, I7,i( I7q) 17, I7/17,17,I7, I7 17, I7,i( I7q),i( l7q) 17.17 ND I7,i( I7q) multiple i( I7q) I7,i( I7q)/ 17, I 7,i( 17q),i( I7q) 17,17 17.17 Complex ND 17,17 I7,i( I7q) ND

ND 17.17

Relative increase in 17q 17,17 17, i( I7q) Deletion distal I7p,

17,17, del 17p 17,17 I7,I7, some cells with

trisomy 17q 3-4 copies of I 7 Loss of I7p, gain of 17q Loss of I7p, gain of 17q 3-4 copies of I7 N o 17p deletion 17, i( I7q) I7,17,17 17,17 17,17 17,17 17, I7,i( I7q) 5 copies of I 7

deletion proximal 17q

17. I7,i( I7q),i( I7q) 17.17 I7.i( I7q) I7,i( I7q) 17, I7,i( I7q),i( I7q) I7,i( I7q) 2-4 copies of 17 17,17 3 copies of I 7 4 copies of 17, 2-3 i( I7q) 17.17 I7,i( I7q) 5 copies of 17, one with

del 17p No del 17p 17.17

"ND = not determined.

17p and 2 copies of the 17 centromere. T h e karyotypes were obtained from cells that had been cul- tured for 14 days, whereas the FISH studies were done on direct preparations or 24-hour cultures. Therefore, if the cells with the 17p deletion had a growth advantage in culture, they could have been missed in the FISH studies. In the second recurrent case, 94-92, the FISH results were consistent with the presence of 3 No. 17 chromosomes. In- terestingly, one chromosome 17 appeared to have a deletion involving D17S34. T h e results for the 17q cosmids were inconsistent, because there appeared to be an extra copy of the D17S33 locus, but an apparent loss of the D17S75 locus in some cells.

We were unable to obtain informative karyo-

types from 10 of the 34 cases. Three patients had apparently normal chromosomes 17 by FISH (92- 156, 93-64, 93-88). Two of the patients (93-64 and 93-88) had the normal diploid number, whereas one patient (92-156) appeared to have trisomy or tetrasomy for chromosome 17. Case 93-01-A047 did not have a deletion at the D17S34 locus, and there was not enough material to test any of the q-arm probes. Two of the 10 cases (94-40 and 94- 101) had FISH results consistent with an i(l7q). Case 94-40 had one normal 17 and an i( 17q), and case 94-101 had 4 normal No. 17 chromosomes and 2 or 3 copies of an i(17q).

Deletions at one or more loci on chromosome 17 were detected in the remaining 4 cases. Case

FISH ASSAY FOR i(17q) IN PNETs 93

92-04110 appeared to have 5 No. 17 chromosomes, one of which was deleted a t D17S34. There was not enough material to test any of the q-arm probes to see whether the case had an i(17q). Case 92-148 had 2 normal No. 17 chromosomes and one deleted chromosome 17; the deletion appeared to involve the entire p-arm in this case. Case 92-142 appeared to be deleted at the D17S34 locus, as well as for HERZ. Finally, case 93-23 had an extremely complicated series of results, with populations of cells that had extra signals for the D17S34, D17S33, and HER2 loci, but apparent loss at D17S122. There was also a large amount of cell-to-cell variability within the tumor for each probe examined. Over- all, these results demonstrated that the tumor was in the triploid to tetraploid range, but we could not categorize it with respect to what chromosome 17 abnormalities may have been present in the tumor.

DISCUSSION

Cytogenetic studies of pediatric PNETs of the CNS have demonstrated that an i( 17q) is the most common structural abnormality (Bigner et al., 1988; Griffin et al., 1988; Biegel et al., 1989; Chadduck et al., 1991; Karnes et al., 1992; Neu- mann et al., 1993). In the present study, we report the largest series of PNETs examined to date. Ten of the 34 PNETs in this series were found to contain one or more copies of an i( 17q) by standard cytogenetics. In each case, the karyotypic findings were confirmed by FISH. An additional 2 cases were determined to have an i( 17q) by FISH alone. Three PNETs had deletions involving the distal region of 1 7 ~ 1 3 , as detected by FISH with the probe for D17S34. Thus, a total of 15 of the 34 (44%) PNETs had deletions of the short arm of chromosome 17. Whereas the percentage of tumors that contain 17p deletions is comparable to the number of PNETs found to have an i(17q) or 17p deletion by standard cytogenetics, we were able successfully to evaluate a much larger number of tumors with the interphase FISH assay. T h e fact that the FISH assay is rapid and does not depend on obtaining metaphase chromosomes will also be helpful in the use of this assay in a clinical diag-

Figure 2. FISH with chromosome 17-specific probes to PNET case 94-70. The a-satellite probe is in red and the cosmid probes are in green. A One hybridization signal is observed for the Dl7534 locus compared to the 2 centromere signals. The tetraploid cell in B contains 2 green and 4 red signals, also consistent with loss of distal 17p. C Hybridization with a probe from the long arm of chromosome 17. D I7S75, demonstrates 3 green and 2 red signals and is consistent with the presence of one normal chromosome 17 and one i(17q) in the tumor.

94 BIEGEL ET AL.

nostic setting. Hybridization with the D17S34 and D17S7.5 probes on 1 7 ~ 1 3 . 3 and 17q25, respectively, will detect an i( 17q) in most cases. Finally, the ability to screen a series of PNETs rapidly to identify smaller deletions of 17p will be particularly helpful in positional cloning efforts designed to identify a tumor suppressor gene.

T h e interphase FISH assay which we employed in the present study was specifically designed to take into account the fact that, concomitant with deletion of 17p, tumors with an i(17q) also have an increase in copy number for loci on 17q. Because an increased dosage of genes that map to 17q may play a role in the initiation or progression of PNETs, we were interested in determining whether there were tumors with extra copies of 17q that did not have a deletion of 17p. One case, 92- 155, had cells with apparent trisomy for 17q, as detected by FISH with a probe for D17S75. Case 92-129 also demonstrated a relative increase in the number of signals with the HERZINEU and D17S75 probes, and i t contained an increase in copy number for the myeloperoxidase locus detected with the commercial i(17q) probe. T h e FISH results with D17S34 and TP5.? suggested that there was an extra copy of distal 17p in some cells. However, the number of centromere signals was not consistent, and we could not rule out the possibility of an i(17q) with an additional unbalanced translocation of 17p. Overall, these results suggest that i t is the loss of 17p, and not an increased dosage for genes on 17q, that is responsible for the development of PNETs.

T h e formation of an i(17q) is not understood, but has been proposed to result from several different mechanisms, including misdivision laterally across the centromere, or an isochromatid break and exchange. Studies of i(Xq) chromosomes have demonstrated that there is variation in the number of repetitive and unique sequences retained around the centromere in different isochromo- s o m a Xq, and that many of the chromosomes which appear to be monocentric by standard cytogenetics are in fact dicentric (Sharp et al., 1990). In our series of PNETs, t h e morphology of the i( 17q)s suggested that some of them were dicentric. In several tumors, the interphase FISH studies with the a-satellite probe demonstrated that one of the a-satellite signals appeared to be larger or more diffuse than the other. However, this was quite variable, and we could not correlate the morphology of the i(17q) in metaphase with what the a-satellite signal looked like in interphase. In 2 PNETs, each with 1 normal 17 and 1 i( 17q) (cases 93-41 and

92-139), the number of signals for the proximal 17q markers D17S33 and HERZ, respectively, were equal to the number of signals for the a-satellite probe. One possible explanation for these results is a decreased hybridization efficiency of the 17q arm probes compared to the centromere probe. Alter- natively, this suggests that there is heterogeneity in the chromosome 17 breakpoints which result in the formation of the i(17q). Arnoldus et al. (1991b) have shown that there is a propensity for the chromosome 17 centromeres to associate in the interphase nucleus. This phenomenon accounts for the high frequency of monosomy 17 cells seen, even in control samples, when the peri-centromeric a-satellite probes are used for determination of numerical changes of chromosome 17 by interphase FISH analysis. It is possible that this association extends into the proximal portion of the p- or q-arm as well. However, given the high resolution for interphase mapping of probes, and the fact that D17S33 clearly maps below the centromere and proximal to HERZ (data not shown), this explanation is unlik- ley to account for the apparent heterogeneity in the breakpoints seen in the tumors with an i(l7q). A final possibility is that there was a t(17;17) that resulted in a chromosome that looked like an i( 17q). Additional molecular and cytogenetic studies with sequences isolated from the centromeric region of chromosome 17, and with polymorphic markers from the p- and q-arms, will be helpful in elucidating the mechanisms involved in the gener- ation of an i(17q) or t(17;17) in these cases.

In addition to the presence of an i(17q), several other chromosomes appear to be non-randomly involved in PNETs. We previously reported the non-random involvement of chromosomes 1, 6, 7, 11, and 16; the loss of a sex chromosome; and the presence of double minutes in these tumors (Big- ner e t al., 1988; Biegel et al., 1989, 1992) . We have confirmed our previous findings that abnormalities of chromosomes 6, 7, and 11 are frequent changes in PNETs. Loss of all or part of chromosome 6 has been noted as a single change in several PNETs (Griffin et al., 1988; Biegel et al., 1989; Stratton et al., 1991), suggesting that it is another primary cytogenetic event. Loss of heterozygosity studies have confirmed the deletion of chromosome arm 6q (Raffel et al., 1990) in some PNETs. Interestingly, deletion of chromosome 6 is usually found in tumors that do not contain an i( 17q), suggesting that it may define a subset of PNETs which arise through a different initiating cytogenetic event.

Trisomy 7 is the most common secondary

FISH ASSAY FOR i(l7q) IN PNETs 95

change in PNETs and was present in 4 of the first 22 cases (Biegel et al., 1989) and in 5 of the 24 cases now reported. Contrary to the loss of chromosome 6, trisomy 7 is often seen in tumors with an i( 17q). Loss of a sex chromosome has also been demonstrated in tumors with an i(17q) (Biegel et al., 19891, although, based on our present results, this does not appear to be a frequent change.

Numerical and structural abnormalities of chromosomes 8 and 11 are relatively common in PNETs, although the types of rearrangements seen are variable (Bigner et al., 1988; Griffin et al., 1988; Biegel et al., 1989; Callen et al., 1989; Karnes et al., 1992; Neumann et al., 1993). Loss of heterozygosity for probes on chromosome 11 was seen in 3 of 11 medulloblastomas/PNETs by James et al. (1990) and Fults et at. (1992), and in 1 of 7 patients screened by Raffel et al. (1990). Microsat- ellite analysis of chromosome 11 did not reveal loss for tyrosine hydroxylase or the D11S490 locus in 15 informative PNETs in a study reported by Al- brecht et al. (1994). Unfortunately, these studies were limited in terms of the numbers of patients examined and the number and location of the chromosome 11-specific probes tested. Additional com- bined cytogenetic and molecular studies are re- quired for determining the frequency and nature of the chromosome 11 rearrangements in PNETs before their significance can be understood.

Several studies have suggested that the loss of chromosome 22 may be significant for the development of medulloblastomas or PNETs (Vagner- Capodano et al., 1989; James et al., 1990). We have not observed the loss of chromosome 22 as a single change in PNETs of the central nervous system in our studies. In addition, Thomas and Raffel (1991) did not detect LOH for chromosome 22 in 20 informative patients. It is possible that some of the reported PNETs with monosomy 22 are of a different histologic subtype, such as a rhabdoid or atypical teratoid tumor (Biegel et al., 1990), or that abnormalities of chromosome 22 define another subset of PNETs which may be less frequent. T h e detailed analysis of candidate genes for these tumor types will eventually resolve these questions.

ACKNOWLEDGMENTS

The authors thank Dr. Barry Barnoski for read- ing the manuscript, Beatrice Sellinger and Yvonne Tatsumura for technical assistance, and Kathy Bonner for obtaining the clinical specimens from the patients at T h e Children’s Hospital of Phila- delphia. We are grateful to the following individu-

als for contributing specimens to this study: Drs. Anne-Christine Duhaime, Luis Schut, Philip Mon- teleone, Steven Halpern, Naryan Shah, Pamelyn Close, Jose Bonnin, and Theodore Pysher. This work was supported by grants from the National Institutes of Health (CA46274) (J.A.B.) and Ron- ald McDonald Children’s Charities (J.A.B.) and was facilitated by the Children’s Cancer Group, which is supported in part by a grant from the Na- tional Institutes of Health (CA13539). Tissue samples 92-04-110, 93-01-A047, and 94-07-PO01 were provided by the Cooperative Human Tissue Net- work, which is funded by the National Cancer In- stitute. Other investigators may have received samples from these same tissues.

REFERENCES Albrecht S, von Deimling A, Pietsch ’r, Giangaspero F, Brandner S,

Kleihues P. Wiestler OD (1994) Microsatellite analysis of loss of heterozygosity on chromosomes 9p, l l p and 17p in medulloblastomas. Neuropathol Appl Neurobiol 2074-81.

Arnoldus EPJ, Noordcrmeer IA, Boudewijn Peters AC, Voormolen JHC, Bots GTAM, Raap AK, van der Ploeg M (1991a) Interphase cytogenetics of brain tumors. Genes Chromosom Cancer 3: 101- 107.

Arnoldus EPJ, Noordermeer IA, Peters ACB, Raap AK, van der Ploeg M (1991 b) Interphase cytogenetics reveals somatic pairing of chromosome 17 centromeres in normal human brain tissue. but no trisomy 8 or sex-chromosome loss. Cytogenet Cell Genet Sh: 214-216.

Arnoldus EPJ, Wolters LBT, Voormolen JHC. van Duinen SG. Raap AK, van der Ploeg bl, Boudewijn Peters A C (1992) Inter- phase cytogenetics: a new tool for the study of genetic changes in brain tumors. J Neurosurg 76997-1003.

Biegel JA. Rorke LB, Packer RJ, Sutton LN, Schut L, Bonner I<. Emanuel B (1989) Isochromosome 17q in primitive neuroectoder- ma1 tumors of the central nervous system. Genes Chromosom Cancer 1: 139-147.

Biegel JA, Rorke LB, Packer RJ, Emanuel BS (1990) Monosomy 22 in rhabdoid or atypical teratoid tumors of the brain. J Neurosurg 73:710-714.

Biegel JA, Burk CD, Barr FG, Emanuel BS (1992) Evidence for a 17p tumor related locus distinct from pS3 in pediatric central nervous system primitive neuroectodermal tumors. Cancer Res 52:3391-3395.

Bigner SH, Mark J. Friednian HS, Biegel JA, Rigner DD (1988) Structural chromosomal abnormalities in human medulloblas- coma. Cancer Genet Cytogenet 30:Yl-101.

Callen DF, Cirocco L, Moore L (1989) A der(ll)t(8;11) in two medulloblastomas. Cancer Genet Cytogenet 38:255-260.

Chadduck WM, Boop FA, Sawyer JR (1991) Cytogenetic studies of pcdiatric brain and spinal cord rumors. Pediatr Neurosurg 92:57- 65.

Chung R, Whaley J, Kley N, Anderson K, Louis D, blenon A, Heetich C, Freiman R, Hedlcy-Whyte ET. Martuza R, Jenkins R, Yandell D, Seizinger BR (1991) TP53 gene mutations and 17p deletions in human astrocytomas. Genes Chromosom Cancer 3:323-331.

Cogen PH (1991) Prognostic significance in molecular genetic mdrk- ers in childhood brain tumors. Pediatr Neurosurg 17:245-250.

Felix CA, Slavc I , Dunn M. Strauss EA, Phillips PC, Rorke LB, Sutton LN, Bunin GR. Biegel JA (1995) p53 mutations in pcdiatric brain tumors. MPO. in press.

Fults D, Petronio J, Noblett BD. Pedone CA (1992) Chromosome 1 lp15 deletions in human malignant astrocytomas and primitive neuroectodermal tumors. Genomics 14:799-801.

Griffin CA, Hawkins At,, Packer RH, Rorke LB, Emanuel BS (1988) Chromosome abnormalities in pediatric brain tumors. Can- cer Res 48:175-180.

ISCN (1991) Cancer Cytogenetics, Supplement to An International System for Cytogenetic Nomenclature VI. Basel. Switzerland: S. Karger, 54 pp.

96 BIEGEL ET AL.

James CD, He J, Carlbom E, Mikkelsen T, Ridderheim P-A, Cave- nee WK, Collins VP (1990) Loss of genetic information in central nervous system tumors common to children and young adults. Genes Chromosom Cancer 294-102.

Karnes PS, Tran TN, Cui MY, Raffel C , Gilles FH, Barranger JA, Ying KL (1992) Cytogenetic analysis of 39 pediatric central nervous system tumors. Cancer Genet Cytogenet 59: 12-19.

Kuwano A, Ledbetter SA, Dobyns WB, Emanuel BS, Ledbetter DH (1991) Detection of deletions and cryptic translocations in Miller-Dieker syndrome by in situ hybridization. Am J Hum Genet 49:707-7 14.

Lehrach H et al. (1990) In Davis KE, Tilghman SM (eds): Genome Analysis Volume 1: Genetic and Physical Mapping. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, pp 39-81.

Lichter P, Cremer T , Borden J , Manuelidis L, Ward DC (1988) Delineation of individual human chromosomes in metaphase and interphase cells by in situ suppression hybridization using recom- binant DNA. Hum Genet 80:224-236.

Matsumura K, Kallioniemi A, Kallioniemi 0, Chen L, Smith HS, Pinkel D, Gray J. Waldman FM (1992) Deletion of chromosome 17p loci in breast cancer cells detected by fluorescence in situ hybridization. Cancer Res 523474-3477.

McDonald JF, Daneshvar L, Willert JR, Matsumura K, Waldman F, Cogen PH (1994) Physical mapping of chromosome 17~13.3 in the region of a putative tumor suppressor gene important in medulloblastoma. Genomics 12229-232.

Mertens F, Johansson B, Mitelman F (1994) Isochromosomes in neoplasia. Genes Chromosom Cancer 10221-230.

Molenaar WM, Jansson DS, Gould VE, Rorke LB, Franke WW, Lee VM-Y, Packer RJ, Trojanowski JQ (1989) Molecular markers

of primitive neuroectodermal tumors and other pediatric central nervous system tumors. Lab Invest 61:635-643.

Neumann E, Kalousek DK, Norman MG, Steinbok P, Cochrane DD, Goddard K (1993) Cytogenetic analysis of 109 pediatric central nervous system tumors. Cancer Genet Cytogenet 71:40-49.

Raffel C, Gilles FE, Weinberg KI (1990) Reduction to homozygos- ity and gene amplification in central nervous system primitive neuroectodermal tumors of childhood. Cancer Res 50:587-591.

Raffel C, Thomas GA, Tishler DM, Lassoff S, Allen JC (1993) Absence of p53 mutations in childhood central nervous system primitive neuroectodermal tumors. Neurosurgery 33:301-306.

Rorke LB, Gilles FH, Davis RL, Becker LE (1985) Revision of the WHO classification of brain tumors for childhood brain tumors. Cancer 56(Suppl): 1869-1886.

Saylors RL, Sidransky D, Friedman HS, Bigner SH, Bigner DD, Vogelstein B, Brodeur GM (1991) Infrequent p53 gene mutations in medulloblastomas. Cancer Res 51:4721-4723.

Sharp CB, Bedford HM, Willard H F (1990) Pericentromeric struc- ture of human X “isochromosomes”: evidence for molecular heterogeneity. Hum Genet 85330-356.

Stratton MR, Darling J , Cooper CS, Reeves BR (1991) A case of cerebellar medulloblastoma with a single chromosome abnormality. Cancer Genet Cytogenet 53:lOl-109.

Thomas GA, Raffel C (1991) Loss of heterozygosity on 6q, 16q, and 17p in human central nervous system primitive neuroectodermal tumors. Cancer Res 51:639-643.

Vagner-Capodano AM, Gentet JC, Choux M, Lena G, Garbarelli D, Bernard JL, Raybaud C (1989) Chromosomal abnormalities in 16 pediatric brain tumors. Pediatr Neurosci 1 4 150-160.

Documents

Isochromosome 17q demonstrated by interphase fluorescence in situ hybridization in primitive neuroectodermal tumors of the central nervous system