_J Clin Pathol 1997;50:553-558

Numerical chromosomal aberrations in Hodgkin'sdisease detected by in situ hybridisation on

routine paraffin sections

James H Pringle, Jacqueline A Shaw, Angela Gillies, Ian Lauder

AbstractAims-To visualise directly numericalchromosomal aberrations and polyploidyin both Hodgkin and Reed Sternberg(HRS) cells and background cells fromcases of Hodgkin's disease using in situhybridisation.Methods-Non-isotopic DNA in situ hy-bridisation was applied to interphase cellnuclei of Hodgkin's disease within routineparaffin embedded tissue sections. Two asatellite DNA probes, specific for chromo-somes 3 and 12, were used to evaluate thefeasibility of this approach. Double label-ling with immunocytochemical detectionof the CD30 antigen was used to identifyHRS cells. Cytogenetic normal diploidand triploid placental tissue served ascontrols.Results-The eight cases ofHodgkin's dis-ease investigated displayed frequent poly-somy, while the majority of backgroundcells showed disomy signals.Conclusions-Numerical chromosomalaberrations were detected in HRS cellsfrom eight cases ofHodgkin's disease by insitu hybridisation. These data show that inHodgkin's disease HRS cells frequentlydisplay polyploidy compared with back-ground cells and are, therefore, probablythe only neoplastic component in this dis-ease. Correlations between polysomy andtumour type or grade could not be madefrom these data owing to the limitednumber of cases examined and to prob-lems with interpreting data from trun-cated nuclei.(7 Clin Pathol 1997;50:553-558)

Keywords: Hodgkin's disease; chromosome aberrations;polyploidy; in situ hybridisation

Department ofPathology, Universityof Leicester, RobertKilpatrick ClinicalSciences Building,PO Box 65,Leicester RoyalInfirmary, LeicesterLE2 7LX, UK

Correspondence to:Dr Pringle.

Accepted for publication8 April 1997.

Many different tumours are characterised byspecific chromosomal aberrations, which can

be useful for defining those with a similar bio-logical background. The detection of particularchromosomal aberrations can also be usefulboth for diagnosis and assessment ofprognosis. 1

Hodgkin's disease is characterised by a smallnumber of putative malignant cells, the mono-nuclear Hodgkin and multinucleated ReedSternberg (HRS) cells. These cells usuallycomprise less than 2% of the tumour mass,among a background of lymphocytes, plasmacells, histiocytes, neutrophils, eosinophils, and

stromal cells. The lineage ofHRS cells remainselusive and the histopathological designation ofHodgkin's disease is likely to encompass anumber of distinct immunophenotypic andimmunogenetic entities. Two key subtypes ofthe disease, nodular sclerosing and mixedcellularity, both characterised by expression ofthe CD30 and CD15 antigens,2 were consid-ered in this study.

Cytogenetic studies of Hodgkin's diseasehave presented particular problems owing to:the rarity ofHRS cells in cellular material frombiopsied lymph nodes; HRS cells having lowmitotic indices; and the typically complexkaryotypes of these cells and difficulties withbanding.3 Teerenhovi et al were the first todemonstrate successfully that clonal chromo-somal abnormalities are carried by HRS cellsexpressing the CD 15 and CD30 antigens.4Other cytogenetic studies of Hodgkin's diseasehave reported cytogenetic changes typical of Bcell lymphomas,5 T cell malignancies,6 andnumerous other chromosomal abnormalities.7Numerical abnormalities, both gain and loss ofchromosome copy number, have been identi-fied for virtually all chromosomes. Structuralchanges (for example, translocations and dele-tions) are also common, although no singleabnormality predominates. One study thatidentified frequent loss or deletion of chromo-some 4q25-4q279 is noteworthy, as loss of thisregion has not been reported previously inother human malignancies. However, no clearcorrelations have been determined between thenumerous chromosomal abnormalities andeither clinicopathological features orprognosis.7 This might be explained by the factthat the metaphases analysed may not neces-sarily be derived from the malignant cells ormight reflect distinct genetic entities.

Non-isotopic in situ hybridisation (NISH) isa powerful method for discriminating numeri-cal chromosomal changes in interphase cellnuclei in solid tumours.10 The method can beused on tissue sections, allowing combinedcytogenetic and histological analyses.'1 12 Toresolve some of the problems associated withcytogenetic studies of Hodgkin's disease, weused NISH with chromosome specific repeti-tive DNA probes to interphase cell nuclei ofnodular sclerosing and mixed cellularity Hodg-kin's disease within routine paraffin waxembedded tissue sections. In situ hybridisationwas also carried out after immunocytochemicaldetection of the CD30 antigen to identify HRScells. Chromosome specific a satellite DNA

553

on February 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.7.553 on 1 July 1997. Dow

nloaded from

54Pringle, Shaw, Gillies, Lauder

Table 1 Tumour and patientfeatures for eight cases ofHodgkins's disease

Age HistopathologicalCase Sex (years) diagnosis

1 Male 18 NS grade 12 Male 27 NS grade 13 Male 1 6 NS grade 14 Female 32 NS grade 25 Female 5 MC6 Female 39 NS grade 17 Male 22 NS grade 28 Female 21 NS grade 1

MC, mixed cellularity; NS, nodular sclerosing.

probes from chromosomes 3 and 12 were usedto illustrate numerical chromosomal aberra-tions.

MethodsHISTOLOGICAL MATERIALEight histologically unequivocal cases ofHodg-kin's disease were selected from the pathologyarchives of Leicester University and theLeicester Royal Infirmary (table 1). Formalinfixed paraffin wax embedded biopsy material,taken for histological evaluation, was studied ineach case. None of the patients had undergoneany therapy at the time of biopsy. Normal dip-loid and cytogenetic triploid placental tissuesserved as controls.

DNA PROBES AND PROBE LABELLINGPericentromeric a satellite DNA probes spe-cific for chromosomes 3 and 12 (D3Z1 13 andD12Z3'4) were selected for study. Recom-binant DNA clones were obtained from theAmerican Type Culture Collection (ATCC,Rockville, Maryland, USA). For in situ hy-bridisation 2 jg of each plasmid DNA waslabelled with digoxigenin-1 1-dUTP (Boe-hringer Mannheim, Germany) by randompriming'5 modified for non-isotopic labelling."

CHROMOSOME IN SITU HYBRIDISATIONThe DNA in situ hybridisation procedure" wasperformed as follows: 5 jm paraffin sectionswere mounted on silane coated slides, dried,heated, and dewaxed. The slides were trans-ferred to 50 mM Tris HCI pH 7.65 anddigested with proteinase K (Boehringer Man-heim) solutions in Tris HCI at concentrationsranging from 25 to 50 jig/ml for one hour at37°C in a humidity chamber. This wasfollowed by treatment with 0.4% paraformal-dehyde, 0.1 M phosphate buffered saline at4°C (using precooled solutions) and ultra purewater (x2) before prehybridisation. The opti-mum concentration of proteinase K was deter-mined for each case to obtain an optimalbalance between morphological preservationand DNA target accessibility.

Prehybridisations were carried out for onehour at 42'C with an overlay of hybridisationbuffer (50% formamide, 2x saline sodiumcitrate (SSC) 10% dextran sulphate, 1 mMTris pH 7.5, 100 jM EDTA pH 8.0). Freshhybridisation buffer containing 500 ng/ml ofdigoxigenin labelled probe and 500 jg/ml soni-cated salmon sperm DNA was then added.This buffer was located over the section using acoverslip. Chromosomal DNA and probe

DNA were denatured by heating for 10minutes at 1 00°C and hybridisations were car-ried out overnight at 42°C in a humidifiedchamber. Post-hybridisation washes at 42°Ccomprised: 3 x 10 minutes in 50% formamide,1 x SSC, followed by 30 minutes in 0.1 x SSC,2 MgCl,, 0.1% Triton X100. In situ signalswere visualised by incubation for 30 minutes ina 1/4000 dilution of an anti-digoxigenin mono-clonal antibody (Boehringer Mannheim) inblocking solution (3% bovine serum albumin,1 x Tris buffered saline (TBS), 0.1% TritonX100) followed by a similar incubation in a1/400 dilution of biotinylated rabbit antimouseantibody (Dako, UK) and a streptavidin-biotinylated peroxidase complex (SABC-Px;Dako) method using diaminobenzidine (DAB)with nickel (diammonium nickel sulphate 6hydrate) enhancement to give dark brownspots at the location of specific in situ hybrids."Sections were counterstained with 0.1% nu-clear fast red, dehydrated, immersed in xylene,and mounted in DPX.

IN SITU HYBRIDISATION WITH CD30IMMUNOCYTOCHEMISTRYParaffin sections (5 jm thick) were mountedon silane coated slides, dried, heated, dewaxed,and immersed in ultra pure water. Antigenretrieval was achieved by heating sectionssubmerged in 10 mM citric acid (pH 6.0) in adomestic microwave oven (850 W) on fullpower for 60 minutes. Slides were allowed tocool to room temperature in the citrate bufferand then rinsed for 2 x 5 minutes in TBS. Sec-tions were incubated with an overlay of a 1/5dilution of normal rabbit serum (Dako) inblocking solution for 10 minutes, followed byincubation in a 1/10 dilution in blocking solu-tion of anti-CD30 antibody (Ber H2; Dako)overnight at 40C. Slides were treated sequen-tially with TBS and biotinylated rabbit anti-mouse antibody (1/400 dilution in blockingsolution; Dako) for 30 minutes; TBS andSABC-AP for 30 minutes; TBS, ultra purewater, 1 x veronal acetate buffer (VAB; 30 mMsodium acetate trihydrate, 30 mM sodium bar-bitone, 100 mM sodium chloride, 50 mMmagnesium chloride hexahydrate, pH 9.2),and fast red developer (50 mg fast red TR salt(F 2768, Sigma); 50 mg napthol ASBIphosphate (N 2250, Sigma) predissolved in 1ml dimethyl formamide and 24 mg levamisole(L 9756, Sigma) in 100 ml VAB pH 9.2) forone hour and then immersed in ultra purewater. Sections were then hybridised with dig-oxigenin labelled probes as described above,starting at the prehybridisation step, slides werethen rinsed in tap water, air dried, immersed inxylene, and mounted in DPX.CD30 only controls were stained in Mayer's

haematoxylin for one minute, cleared in tapwater, and mounted in aqueous mount.

EVALUATION OF NISH RESULTSSections were evaluated using previously de-fined criteria as follows. Haematoxylin andeosin stained sections from each case werechecked to ensure that they were representativeof the lesion and for preservation of its general

554

on February 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.7.553 on 1 July 1997. Dow

nloaded from

Numerical chromosomal aberrations in Hodgkin's disease

Table 2 Combined data for chromosome specific signals in eight cases ofHodgkin'sdisease, diploid controls, and triploid controls

Percentage of cells with number of spots

Case Probe 0-2 3-4 5-6 Ploidy

1 3 46 51.5 2.5 Trisomy/tetrasomy12 66.5 31 2.5 Trisomy/tetrasomy

2 3 27 57 16 Polysomy12 25 55 20 Polysomy

3 3 25 60 15 Polysomy12 27 62 11 Polysomy

4 3 32 61 7 Polysomy12 31 62 7 Polysomy

5 3 61.5 34 4.5 Trisomy/tetrasomy12 75 25 0 Trisomy/tetrasomy

6 3 29 59.5 11.5 Polysomy12 32 62.5 5.5 Polysomy

7 3 67.5 31.5 1 Trisomy/tetrasomy12 62 36.5 1.5 Trisomy/tetrasomy

8 3 60 35.5 4.5 Trisomy/tetrasomy12 56.5 40 3.5 Trisomy/tetrasomy

Diploid control 3 98.0 2.0 0 Disomy12 97.5 2.5 0 Disomy

Triploid control 3 74.5 25.5 0 Trisomy12 67.0 32.0 1 Trisomy

Hodgkin's disease cases 1 to 8, counts for HRS cells only are given.

architecture. Optimal slides for evaluation of insitu signals were then selected. These wereones with well preserved HRS cells in combi-nation with clear, distinct in situ spots. Onehundred HRS cell nuclei and 100 backgroundcells from each case of Hodgkin's disease,selected from random fields, were scored forthe number of in situ spots (0-6) by twoobservers (JHP, JAS). One hundred large stro-mal cells from the cytogenetic controls werecounted by one observer (AG) as these cells

lo. v

were of a similar size to the HRS cells. Thesedata were pooled into three categories: 0-2spots, 3-4 spots, and 5-6 or more spots.

ResultsTable 1 lists the clinical diagnoses of the eightHodgkin's disease cases studied. Seven of theeight cases were of the nodular sclerosing typegrade 1 or 2, and one case was of mixed cellu-larity Hodgkin's disease. In each case, diagnos-tic biopsy material was studied and none of thecases had undergone any therapy at the time ofbiopsy. The reliability of the in situ method wastested by hybridisation to interphase nuclei ofnormal diploid and cytogenetic triploid controltissues prepared from placentas (figs 1 and 2F,and table 2). In each case, it was necessary tofocus through the nucleus to detect all signalsclearly. The diploid control showed 0-2 spotsin 97.5-98% of cells analysed with both chro-mosome probes and the triploid controlshowed 0-3 in situ signals in 93.0% and 91.0%of cells analysed with the chromosome 3 and12 probes, respectively (table 2 and fig 2B).Both the diploid and triploid controls showedconsistent data. These controls illustrate thereduction in counts owing to large cells withtruncated nuclei. The observer variation wasnot significant for controls or Hodgkin'sdisease cases. The quality of the nuclearmorphology in the Hodgkin's and controltissues was reduced by proteinase K pretreat-ment, which is necessary for the in situhybridisation technique. Therefore, nuclearfast red was used on the controls to identify

I.,,.i0:*I

4 ^

A n

*, :4

'.& 4 -4X

IS

4 .

4.

Figure 1 Aoln-isotopiC D)AA in sitrhybridisation (NISH) /0r cLhromosoiSOes 3 l,and 12,fi)r tcza()ases of Hodgki 's disease, normi-iial diploidcontrosls, and triploid controls. (A) CD30if Cytochiiist.y anld 1AUSH zit/h thcD1273 probewwithi a 5,urn sectiO5)i of HodgkinN'.sdisease case 1. HRS cells cani be distinguisishedshow-ing cvtoplastiic stainingfoi CD30 andnuclei witli three or niore spots indicatinlgpo/vSoniy. (B) GD30 inun1unocvochemistry andciANISH with the 1)I 2Z-3 probe zwithini a 5 ,umsectiOni of Hodgkin 's cdisease case 7. HRS nuiicleican be distiniguished showls)ig polVsonivt(C) Diplohid placenttal conitrol .shlowing NlISwHresults fir chrosniosomiie 12. The miajority ol cellssliow two( Inuclear spots. (D) Triploid control1shliozving NTI-SH results osii placental tissue forchnir(siiosonie 3. The niajority iJfcells shOisw threenuclear spots.

C m,

_.. : ..

i. 4

**t

x....

.,Ri f

f;l.-

i.--J.!w,1.:.:*? -.1k,..i-t -i..`ki.

D

d::.* w-S

V..

555

4*- .:::.-..

'T,

*.. A16 .*.

t.. 1. 00

. 0.40'.

*. i-6

00

44

:*

r

*A

:. 4t :-.-,.

F...^u,,

IF--- V

on February 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.7.553 on 1 July 1997. Dow

nloaded from

Pringle, S/hazv, Gillic%s, Laneder

A

0 1 2

- Chr 3Chr 12

3 4 5 6

CCChr 3 HRS cells90 Chr 12 HRS cells

80

70-

60

50

40

30

20

8071 1 2 3 4 5 6

80 E MChr 3 HRS70L cells

70 * L-.(kr 10 LWPCcellnrs z m;cells

120 B

100

80

60 |

401

20

0 1 2 3 4140 D

120

100

80

60

40 |

20 I

0 1160

140 -

120

100

80

60

40

20

F

1 1 O

2 3 4 5 6 0 1

Chromosome number per cell

2

-M Chr 3cellsCh r 1cells

3 4

- Chr 3Chr 12

5 6

background

12 background

5 6

- Chr 3 backgroundcells

l Chr 12 backgroundcells

2 3 4 5 6

Figure 2 DNA probe Jrequency distributionls showilg the num;iber of hybridisation spots per lucleus for Hodgkin 's diseasecases anid conitrols after NISH using the two probes anid 5 lnu tissue sectionis. (A) Diploid conztrol shozwing diploid NISHprofile; (B) cytogenietic triploid control showing triploid profile; (C) case 1 Hodgkin anid Reed Steinlberg (HRS) cell profile;(D) case 1 background cell profile; (E) case 2 HRS cell profile; (F) case 2 backgr-oulnd cell profile.

individual nuclei and no counterstaining was

necessary where CD30 immunostaining was

used to identify HRS cells.The combination of immunocytochemistry

for CD30 and chromosome specific in situ

hybridisation served to identify HRS cells spe-

cifically, facilitating the analysis of in situ spots(fig 1). No significant masking of the in situsignal was observed by the fast red CD30immunostaining. The in situ photomicro-graphs illustrated clearly that the size of CD30positive HRS cells varied as did the number ofin situ spots per HRS cell nucleus. Wherebinucleated cells were clearly distinguishable,these were counted as two distinct nuclei.However, these cells were far less common

than mononuclear HRS cells. Figure 2 showsDNA probe frequency distributions for Hodg-

kin's disease cases 1 and 2 (both HRS andbackground nuclei) and the diploid andtriploid controls. These data show that thenumber of spots per HRS cell nucleus variedmore widely than for the background cells fromthe same case. The pooled in situ data (bothobservers) from the eight Hodgkin's diseasecases, normal diploid controls, and triploidcontrols are summarised in table 2. For allHodgkin's cases, only the HRS cell counts are

presented in table 2 as counts for thebackground cell populations predominantlyshowed disomy for both probes (0-2 spots in>90% of nuclei counted) (figs 2D and 2F). Alarger proportion of these cells showed two

spots than for the diploid control because thesecells were smaller and less prone to nucleartruncation effects. Rare background cells with

90

80

70-

60-

50 -

40

30

20

10

0-

in -

0

a,

60

50 -

40

30

20

10

0 -0 1

556

on February 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.7.553 on 1 July 1997. Dow

nloaded from

Numerical chromosomal aberrations in Hodgkin 's disease

three spots (< 10%) are likely to reflect theclose proximity of two small nuclei rather thangenuine trisomy.Four of the eight Hodgkin's disease cases (2,

3, 4, and 6) showed polysomy with more than65% of HRS cells having three or more in situsignals with both chromosome probes (table2). The remaining Hodgkin's disease cases

appeared to show trisomy/tetrasomy for bothchromosome probes as 5-6 in situ signals were

rare (< 5%). Of these cases, one showed more

frequent trisomy/tetrasomy for the D3Z 1probe than for the D12Z3 probe. These resultsshow that clear numerical chromosome datacan be obtained from the HRS cells and back-ground cells.

DiscussionNISH with chromosomes 3 and 12 specific asatellite DNA probes to paraffin archival tissuesections allowed us to screen eight cases ofHodgkin's disease for the presence of numeri-cal aberrations of the selected chromosomes.Chromosomes 3 and 12 were selected as

numerical aberrations of these chromosomeshad been observed previously in metaphaseanalyses of Hodgkin's disease."6 In combina-tion with immunocytochemistry to mark theCD30 antigen, this approach delineated spe-

cific HRS cells that showed clear numericalchromosomal aberrations, frequently polys-omy, while the background cell populationswere mostly diploid for the two chromosomesstudied. For the in situ studies carried outwithout CD30 detection the tissue architecturewas retained sufficiently to enable histologicalexamination and the identification of HRScells. Hence, it was possible to distinguishdirectly between HRS and background cellsusing both approaches. The analysis of inter-phase nuclei in tissue sections, therefore, offerscertain advantages over other techniques thatuse isolated cells or tissue extracts. Moreover,the background cells give useful informationwith repect to hybridisation efficiency andpatient genetics.

Despite sectioning, which yields sliced nu-

clei, specific aberrations were detected for HRScells from all eight Hodgkin's disease cases

studied, with all cases showing polysomy,although the frequency of polysomy varied.Three cases showed approximately 30% ofHRS cells having trisomy/tetrasomy of bothchromosome 3 and 12. One case appeared toshow more frequent trisomy/tetrasomy for theD3Z1 probe than for the D12Z3 probe andfour cases showed more than 65% ofHRS cellswith three signals or more for both chromo-some probes. It is noteworthy that no signifi-cant under-representation of either chromo-some 3 or 12 was detected in any of the eightHodgkin's disease cases studied. These datasuggest that deletion of these two chromo-somes may not be common in this disease.

It is difficult to distinguish accuratelybetween trisomy, tetrasomy, and higher orderpolysomy, as not all nuclei in the plane of thesection will retain their full chromosome com-

plement. For this reason, these in situ data are

likely to underestimate the true frequency of

chromosomal aberrations. 7 Moreover, thenumber of signals detected per nucleus is likelyto become more variable as the chromosomecopy number increases. Our data for the twocytogenetic controls appear to support thissuggestion. For the diploid control, 98.0% ofnuclei counted showed 0-2 in situ signals (dis-omy), whereas for the cytogenetic triploid con-trol fewer of the nuclei counted (93.0% and91.0% for the chromosome 3 and 12 probes,respectively) showed 0-3 distinct in situ spots(table 2). These data suggest that the frequencyof higher polysomy (trisomy or greater) isincreasingly underestimated by this type ofanalysis.The problem of chromosome retention on

sectioning and, therefore, underestimation ofactual chromosome number may be furthercompounded by the larger size of HRS cellscompared with the background cells. The HRScell chromosome number may vary alsobecause of binucleated and multinucleatedvariants. In an attempt to address this problem,when binucleated HRS cells were clearlydistinguishable, they were counted as twodistinct nulcei. However, these were far lesscommon than mononuclear variants. Despitethese shortcomings, four ofthe eight Hodgkin'sdisease cases studied (2, 3, 4, and 6) showedclearly a high frequency ofpolysomy (> 65% ofHRS nuclei having three or more signals). Theother four cases (1, 5, 7, and 8) showed a lowerbut significant frequency of polysomy and,therefore, are more likely to have trisomy ortetrasomy. These data support previous studieswhere metaphase analyses of Hodgkin's diseaserevealed frequent trisomy and tetrasomy ofboth chromosomes 3 and 12.16 However, thevariation in the number of spots detectedwithin HRS cells for all eight cases might sug-gest that both inter- and intratumoral heteroge-neity are common. As mentioned previously,the effects of nuclear truncation on sectioningare significant. Therefore, all the data shouldonly be interpreted as providing global infor-mation about local differences in ploidy withina tumour.'8Weber-Matthiesen et all9 used simultaneous

fluorescence immunophenotyping and inter-phase cytogenetics to study fresh cell prepara-tions in 30 cases of Hodgkin's disease. Thesestudies showed that HRS cells contained com-plex chromosome aberrations with hyperploidypredominating. Our findings from routine for-malin fixed paraffin embedded diagnosticmaterial for eight cases of Hodgkin's diseasesupport these conclusions. Moreover, ourmethod is applicable to a much wider range ofstored tissue samples. One study, using inter-phase cytogenetics on fresh tissue samples,demonstrated that HRS cells obtained fromdifferent tissue samples from the same patienthave a similar unique abnormal chromosomalpattern, leading the authors to suggest thatHRS cells are clonal.20 However, we believethat it is difficult to establish by cytogeneticmethods that these DNA changes are clonal.Indeed, we would argue that the intratumouralheterogeneity in chromosome copy number

557

on February 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.7.553 on 1 July 1997. Dow

nloaded from

8Pringle,Shaw, Gillies, Lauder

observed by us and previous studies'9 20 wouldsuggest that HRS cells are not always clonal inorigin.The effective demonstration of clonality in

Hodgkin's disease using immunoglobulin or Tcell antigen receptor gene rearrangements hasbeen limited to cases with large numbers ofHRS cells where the clonal band can beassigned to these cells2' or analysis of microdis-sected single HRS cell DNA by the polymerasechain reaction.22 More convincing evidencethat HRS cells are clonal populations has comefrom the analysis of Epstein-Barr virus infec-tion of HRS cells and the demonstration thatthe viral genome is of clonal origin by Southernblotting.23 However, despite numerous studiesof clonality in Hodgkin's disease there stillremains a significant number of cases thatappear to be polyclonal in origin.24 This leadsto the possiblity that Hodgkin's disease starts asa polyclonal proliferation of HRS cells andsubsequently develops into a monoclonalpopulation as the disease progresses.25

In conclusion, we have demonstrated theutility of chromosome specific DNA in situhybridisation for detecting numerical chromo-somal aberrations in routine paraffin embed-ded tissue of Hodgkin's disease. These datashow that in Hodgkin's disease HRS cells pos-sess frequent polysomy compared to back-ground cells and are, therefore, probably theonly neoplastic component in the disease. Nocorrelations between polysomy and eithertumour type or grade could be made fromthese data, owing to both the limited number ofcases examined and problems with interpretingdata from truncated nuclei. If polysomy is acommon feature of all cases of Hodgkin'sdisease, this information would have littleprognostic significance, although it would indi-cate genomic instability of the HRS cells.

We thank Dr P McKeever (department of pathology, Universityof Leicester) for providing material for the cytogentic diploidand triploid controls.

1 Boehm T, Rabbitts TH. A chromosomal basis of lymphoidmalignancy in man. Eurj7 Biochem 1989;185:1-17.

2 Drexler HG. Recent results on the biology of Hodgkin andReed-Sternberg cells. I. Biopsy material. Leuk Lymph 1992;8:283-313.

3 Rowley JD. Chromosomes in Hodgkin's disease. CancerTreat Rep 1982;66:639-43.

4 Teerenhovi L, Lindholm C, Pakkala A, Franssila K, Stein H,Knutilla S. Unique display of a pathologic karyotype in

Hodgkin's disease by Reed-Sternberg cells. Cancer GenetCytogenet 1988;34:305-1 1.

5 Cabanillas F, Pathak S, Trujillo J, Grant G, Cork A, Hage-meister FB, et al. Cytogenetic features of Hodgkin's diseasesuggest a possible origin from a lymphocyte. Blood1988;71:1615.

6 Tilly H, Bastard C, Delastre T, Duval C, Bizet M,Lenormand B, et al. Cytogentic studies in untreated Hodg-kin's disease. Blood 1991;77:1298-304.

7 Ladanyi M, Parsa NZ, Offit K, Wachtel MS, Filippa DA,Jhanwar SC. Clonal cytogenetic abnormalities in Hodg-kin's disease. Genes Chromosom Cancer 1991;3:294-9.

8 Ankathil R, Vijayakumar T, Pillai GR, Nair MK. Chromo-somes in Hodgkin's disease-analysis of involved lymphnodes. Neoplasma 1992;39:245-8.

9 Dohner H, Bloomfield CD, Frizzera G, Frestedt J, ArthurJC. Recurring chromosome abnormalities in Hodgkin'sdisease. Genes Chromosom Cancer 1992;5:392-8.

10 Cremer T, Tesin D, Hopman AHN, Manuelidis L. Rapidinterphase and metaphase assessment of specific chromo-somal changes in neuroectodermal tumour cells by in situhybridisation with chemically modified DNA probes. ExpCell Res 1988;176:199-220.

11 Pringle JH, Homer CE, Warford A, Kendall CH, Lauder I.In situ hybridisation: alkaline phosphatase visualisation ofbiotinylayed probes in cryostat and paraffin sections. Histo-chemJ 1987;19:488-96.

12 Van Dekken H, Kerstens HMJ, Tersteeg TA, VerhofstadAAJ, Vooijs GP. Histological preservation after in situhybridisation to archival solid tumour sections allowsdiscrimination of cells bearing numerical chromosomechanges. _J Pathol 1992;168:317-24.

13 Waye JS, Willard HF. Chromosome specificity of satelliteDNAs: short and long-range organisation of a divergeddimeric subset of human alpha satellite from chromosome3. Chromosoma 1989;97:475-80.

14 Baldini A, Rocchi M, Archidiacono N, Miller OJ, MillerDA. A human alpha satellite DNA subset specific for chro-mosome 12. Am J Hum Genet 1990;46:784-8.

15 Feinberg AP, Vogelstein B. A technique for radiolabellingDNA restriction endonuclease fragments to high specificactivity. Addendum. Analyt Biochem. 1984;132:166-7.

16 Ankathil R, Vijayakumar T, Pillai GR, Nair MK. Chromo-somes in Hodgkin's disease: analysis of involved lymphnodes. Neoplasma 1992;39:245-8.

17 Arnoldus EPJ, Dreef EJ, Noordermeer IA, Verheggen MM,Thierry RF, Peters ACB, et al. Feasibility of in situ hybridi-zation with chromosome specific probes on paraffin waxembedded tissue. J Clin Pathol 1991 ;44:900-4.

18 Beck JLM, Hopman AHN, Feitz WFJ, Schalken J,Schaafsma HE, Van De Kaa CA, et al. Numericalaberrations of chromosomes 1 and 7 in renal cellcarcinomas as detected by interphase cytogenetics. )7 Pathol1995;176:123-35.

19 Weber-Matthiesen K, Deerberg J, Poetsch M, Grote W,Sclegelberger B. Numerical chromosome aberrations arepresent within the CD30+ Hodgkin and Reed-Sternbergcells in 100% of analysed cases of Hodgkin's disease. Blood1995;86: 1464-8.

20 Inghirami G, Macri L, Rosati S, Ying Zhu B, Yee HT,Knowles DM. The Reed-Sternberg cells of Hodgkin diseaseare clonal. Proc NatlAcad Sci USA 1994;91:9842-6.

21 Weiss LM, Strickler JG, Warnke RA, Sklar J. Immunoglobu-lin gene rearrangements in Hodgkin's diseases. Hum Pathol1986;17: 1009-14.

22 Delabie J, Tierens A, Weisenburger DD, Chan WC. Nodu-lar sclerosis Hodgkins disease-lineage and clonality analy-sis using a single cell assay. FASEBJ 1995;9:272A.

23 Weiss LM, Movahed LA, Warnke RA, Skalr J. Detection ofEpstein-Barr viral genomes in Reed-Sternberg cells ofHodgkin's disease. N EnglJ Med 1989;320:502-6.

24 Stein H, Hummel M, Ziemann K, Lammert H. Hodgkin'sdisease-its nature and clonality [abstract]. Exp Hematol1995;8:895.

25 Chan WC, Delabie J. Hodgkin's disease-lineage and clon-ality. Am _J Clin Pathol 1995;4:368-370.

558

on February 14, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.50.7.553 on 1 July 1997. Dow

nloaded from