5 Clin Pathwl 1992;45:556-560

Chromosome studies of solid tumours

N P Bown

IntroductionCytogenetic analysis of malignant cells veryoften indicates acquired chromosome chan-ges-that is, deviations from the normal con-stitutional 46,XX or 46,XY karyotype. Muchof what we know about these changes isderived from studying leukaemias and lympho-mas. Many of these abnormalities are non-random, and particular chromosome changesare consistently found in particular diseasetypes. These abnormalities may be of chromo-some structure (translocations, deletions,inversions, etc) or of chromosome number(monosomies, trisomies, etc). The recurrentfinding of such specific, clonal karyotypeabnormalities underlies the integral role ofcytogenetics in the clinical investigation of pre-leukaemias, leukaemias, and lymphomas.Depending on the specificity of the particularchromosome abnormality, cytogenetics canprovide powerful confirmatory evidence indiagnosis, and is occasionally of prime impor-tance in defining disease type. As the diseaseprogresses, secondary karyotype changes mayappear; where such clonal evolution pathwaysare well characterised-for example, in chronicmyeloid leukaemia'-cytogenetics can be usedto stage the disease and monitor malignantprogression.

For both lymphoid and myeloid disorders,significant correlations have also been estab-lished between the chromosome changes seenin the bone marrow at presentation and thepatient's prognosis in terms of response totreatment and survival.2 4Over the past two decades, cytogenetic

studies in haematological malignancies havealso contributed to profound advances in theunderstanding of oncogenesis: the existence ofrecurring chromosome breakpoints implies thepresence, at discrete loci, of genes that areimportant in the development of cancer. Inseveral diseases molecular studies have eluci-dated the mechanisms whereby chromosomerearrangements result in oncogene activationor alteration.5 6

The cytogenetic investigation of solidtumours is at a much less advanced stage. Thereasons for this are largely technical: manytumours contain areas of necrosis or show lowmitotic activities; and tissue cultures often failto accumulate sufficient numbers of mitoticcells of suitable quality for chromosome analy-sis. Problems also arise in disaggregating thematerial to produce the cell suspensionsrequired for direct harvests of metaphase cellsand for short term cultures. Improvementshave been reported in the yield of metaphase

cells following tumour disaggregation withenzymes such as collagenase.7The cytogenetics of solid tumours is cur-

rently the subject of intense investigation, andthe rate of accumulation of data now makes asingle brief review impossible. What follows isheavily biased towards tumours occurring inchildren, reflecting the author's research inter-est. It is not yet clear whether karyotype studiesof solid tumours will prove, like leukaemiacytogenetics, to be of routine clinical use, butseveral disease specific chromosome aberra-tions have now been described (table). Corre-lations with prognosis and links with molecularchanges of oncogenic importance have alsobeen established.

Sandberg and Turc-Carel89 have indicatedseveral areas in which solid tumour cyto-genetics has already proved valuable:(a) Defining subsets within a histologicallyhomogenous tumour type-for example, thoseshowing i(5p) or del(5q), + 7, - 9 or del (9p)in transitional cell carcinoma of the bladder.(b) Suggesting aetiological connectionsbetween histologically diverse tumours-forexample, virtually all germ cell tumours of thetestis, both teratomas and seminomas, havebeen found to show i(12p).(c) Highlighting known histological subtypes-for example, the t(12; 16) translocationspecific to the myxoid form of liposarcoma.(d) Suggesting a primary tumour site when aspecific karyotype change is found in a meta-static tumour or in bone marrow.When considering chromosome changes in

solid cancers, it is possible, as in leukaemiasand lymphomas, to distinguish between thoseof primary aetiological importance and sec-ondary changes appearing in the course oftumour progression. For a given tumour typethe former are recognised either because theyare seen as the sole karyotype abnormality orbecause they recur as a common link betweenmore complex abnormal karyotypes whichmay include multiple secondary rearrange-ments.

Sequential cytogenetic changes have beenrecorded in a variety of solid tumours as theydevelop from relatively benign forms to showmore malignant, invasive, or metastatic behav-iour. For example, Nowell has summarised thecurrent knowledge of cytogenetic evolution inmelanocytic tumours and in colon cancer.10 Inthe former, karyotypically abnormal clones aremore common as the disease progresses fromcommon naevus, through dysplastic naevus, toinvasive and metastatic melanoma, with 9pand 1Oq rearrangement in early stages and

Department ofHumanGenetics, University ofNewcastle upon Tyne,19-20 ClaremontPlace, Newcastle uponTyne NE2 4AAN P BownCorrespondence to:N P BownAccepted for publication12 December 1991

556

on 13 June 2019 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.7.556 on 1 July 1992. D

ownloaded from

Chromosome studies of solid tumours

Examples of specific primary changes in solid tumours

Tumour Chromosomal abnormality

Meningioma Monosomy 22/del(22q)Mixed salivery gland tumour t(3;8)(p2l;ql2)Lipoma Translocations of 12ql3-14Myxoid liposarcoma t(12;16)(ql 3;ql 1)Small cell lung carcinoma del(3)(pl4p23)Bladder carcinoma i(5p)Breast carcinoma t or del (16q)Testicular teratoma/seminoma i(l2p)Neuroblastoma t or del (lp)

DMs/HSRsEwing's sarcoma and related tumours t(11;22)(q24;ql2)Alveolar rhabdomyosarcoma t(2;13)(q35-37;ql4)Synovial sarcoma t(X;l8)(pl l;ql 1)Retinoblastoma i(6p)

Constitutional del(l3ql4)Wilms's tumour Constitutional del(llpl3)

del-deletion, q-chromosome long arm, p-short arm, t-translocation, i-isochromosome,DM-double minute chromosome, HSR-homogeneously staining region.

alterations of chromosomes 1, 6, and 7 occur-ring at more advanced stages. For colontumours, the progression from colonic polypsto advanced colon cancer is marked by foursuccessive genetic lesions: mutation in a rasoncogene; partial deletion of 5q; deletions ofchromosome 17; and loss of chromosome 18.

Difficulties can arise in the interpretation ofcomplex chromosome changes in tumourswhich have been subjected before biopsy totherapeutic irradiation or drugs, because bothagents can induce chromosome rearrange-ments. " 12 A recent report described serialcytogenetic studies of a paediatric astrocytomawhich evolved to a glioblastoma multiforme.'3At diagnosis, trisomy 1 q was the sole abnor-mality, but after chemotherapy and radio-therapy multiple reciprocal translocations,involving chromosome breakpoints implicatedas "hot-spots" in radiation induced chromo-some damage, were detected.The prognostic implications oftumour cyto-

Figure 1 Neuroblastoma.Metaphase cell showing,among other abnormalities,rearrangement of the shortarm of chromosome I(lp+) typical of thistumour. Inset: normalhomologue on the left,rearranged tp on the right(G-banding by trypsin andLeishman's stain).

genetics are of intense clinical interest, withneuroblastoma being perhaps the best charac-terised tumour in this respect. Cells fromneuroblastomas frequently show deletions or

rearrangements of material from the distalshort arm ofchromosome 1 (fig 1). Karyotypesshowing hyperdiploidy with 50-77 chromo-somes per cell and normal chromosomes 1 are

strongly associated with stage I, II, or IV-sdisease in patients under 1 year of age, and a

favourable outlook, while hypo- or pseudodi-ploidy with lp involvement is more common inchildren over 1 year, with very poor prognosisstage III or IV tumours.'4 15Another highly significant clinical indicator

in neuroblastoma is amplification of the onco-

gene N-myc, located at 2p23-24. In one studypatients with one, three to ten, or more than 10copies of this gene in their tumour cells showedprogression free survival at 18 months of 70%,30%, and 5%, respectively.'6 Simultaneous useof flow cytometry to measure DNA ploidyindex, and Southern blot hybridisation tomeasure N-myc copy number, has been pro-posed as a means to distinguish betweenprognostic subsets in neuroblastoma.'7 How-ever, in at least one study the presence or

absence of a cytogenetically visible lp rear-rangement emerged as a more accurate pre-dictor of tumour progression than N-mycamplification.'8Gene amplification may be visible at a

cytological level in the form of "double min-ute" chromosomes (DMs) and "homogene-ously staining regions" (HSRs).'9 These havebeen frequently reported in neuroblastomaand brain tumours. DMs are paired, acentricchromosome bodies 0-3-0{5 gum in size; HSRsare integrated into the chromosomes as blocksof amplified genes, with these regions failing to

_ a

3z" 1 1*:

P li,.i.,

I 1 1ps

557

11

jp+..o-

on 13 June 2019 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.7.556 on 1 July 1992. D

ownloaded from

Bown

llq+ ~*~

t; 522q-

22

....

3

U.0

11 22

show differential chromosome banding.Neuroblastoma is one of the small round cell

tumours of childhood, along with Ewing'ssarcoma, peripheral neuroepithelioma, rhab-domyosarcoma and malignant soft tissue lym-phoma. The histopathological distinctionbetween these entities can be problematic, andthis has prompted the search for diagnosticallyuseful chromosome aberrations. It was alsohoped that cytogenetics might help to eluci-date the histogenesis of these tumours, espe-cially Ewing's sarcoma, in which the cell oforigin remains unclear. As well as the neuro-blastoma chromosome changes already descri-bed, two particularly important translocationshave emerged within the group of small roundcell tumours.2021 First, a translocationt(11;22)(q24;ql2) has been reported in over

80% of Ewing's sarcomas (fig 2) and has alsobeen found in neuroepithelioma and Askin'stumour, lending support to an embryonicneural crest origin for Ewing's sarcoma. Sec-ond, a t(2;13) (q35-7;ql4) has repeatedlybeen observed as a specific marker for alveolarrhabdomyosarcoma (fig 3).22 (Highly complexkaryotype rearrangements have been recordedin osteosarcoma cells,23 but no specific chan-

ew_.cc0

*. .s@..

Figure 3 Alveolar rhabdomyosarcoma. Partial karyotype of tumour ceUl showingt(2;13) (q36;ql 4).

ges have yet been identified.) The chromoso-mal characterisation of the small round celltumours is therefore beginning to reveal linksbetween disease types, as well as having a moreimmediate practical impact in diagnosis.24 25

Similar hopes that cytogenetics might helpin the classification of brain tumours have notyet been fulfilled. The specific aberration asso-ciated with meningiomas-loss of part or all ofchromosome 22-was first identified evenbefore the development of chromosome band-ing techniques. More recent investigation ofthe gliomas (astrocytoma, medulloblastoma,glioblastoma multiforme, ependymoma, etc),has identified frequent gains of chromosome 7and losses of 10, 22, and either X orY, but onlytentative proposals have so far emerged forspecific associations between chromosomechanges and histopathological subtypes.26Results from two series of paediatric intra-cranial tumours suggest that the karyotypechanges in these cases may differ from adulttumours, with deletions, unbalanced trans-locations, and isochromosome 17q predom-inating, rather than the numerical changescommonly seen in the adult tumours.27 28The clinical or diagnostic interpretation of a

chromosome result from a solid tumour can beproblematic in several respects. First, whenonly normal karyotypes are detected in directpreparations or in tumour cultures, it is usuallyunclear whether these represent tumour cellswith, presumably, submicroscopic geneticlesions, or stromal fibroblasts-that is, themalignant cells may be present but not inmitosis. Second, it is now known that severalnormal (non-neoplastic) human tissues showclonal chromosome changes-for example,gains of chromosome 7 and losses of sexchromosomes in both brain29 and kidney.30The possibility of constitutional tissue specificmosaicism must therefore be borne in mindwhen interpreting aneuploidies in solid tumourkaryotypes. Third, there are now many reportsof chromosomal abnormalities in non-malig-nant tumours. In one series of 109 non-malignant solid tumours of various types,thyroid and parotid adenomas, meningiomas,acoustic neuromas etc, 37% showed clonalnumerical or structural chromosome chan-ges." This breakdown of the previously wellestablished correspondence between acquiredclonal chromosome change and malignancyseems to be confined to solid cancers; inhaematological diseases the traditional dogmaappears to remain valid.Another intriguing cytogenetic pattern

which has emerged, particularly from studiesof skin tumours, has been the phenomenon ofmultiple independent clones showing unre-lated chromosome changes. This situation hasbeen described in, for example, basosquamouspapilloma,32 basal cell carcinoma,33 and squa-mous cell carcinoma.34 Several groups havecited these observations in support of the"field cancerisation" model of polyclonalinitiation of epithelial tumours; others urgecaution in relating in vitro karyotype findingsto the tumour in vivo."What has the identification of tumour rela-

Figure 2 Ewing'ssarcoma. G-bandedmetaphase ceUl and partialkaryotype showingt( 1;22) (q24;q12)translocation.

3

U

558

t(2',13)(q36; q14)

on 13 June 2019 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.7.556 on 1 July 1992. D

ownloaded from

Chromosome studies of solid tumours

ted chromosomal abnormalities shown aboutthe actions of oncogenes in solid cancers? As ageneral rule the chromosome rearrangementsfound in leukaemias and lymphomas seemmainly to bring about a potentiation of onco-gene function either by increasing gene expres-sion or by modifying the gene product to amore active form.36 37 In solid tumours, on theother hand, the predominant mechanismsseem to be either oncogene amplification, suchas c-neu in breast carcinoma and N-myc inneuroblastoma, or the loss or inactivation of"anti-oncogenes".3 " Also known as "tumoursuppressor genes", the normal function ofthese genes is to inhibit cell growth andpromote differentiation. Their loss or inactiva-tion is therefore a critical event in tumourinitiation and loss of heterozygosity-evidentat either the cytogenetic or molecular level-has now been reported in a wide range ofhuman solid cancers. This model for oncoge-nesis has been most fully explored for retin-oblastoma and Wilms's tumour,39 40 in both ofwhich inherited constitutional deletions maypredispose to tumour development.The possible role of oncogenes in a number

of recurring chromosomal aberrations, such asthe t(I 1;22) and t(2;13) rearrangementsdescribed above, remains to be shown.The interaction between cytogenetics and

molecular genetics continues to generateimportant technical advances. Perhaps themost relevant to the study of solid tumours arethe new methods of fluorescent (non-isotopic)in situ hybridisation which permit both "chro-mosome painting" and "interphase cytoge-netics". In the former whole chromosomeDNA libraries are hybridised to metaphasespreads to elucidate structural rearrangements.The latter technique involves hybridising chro-mosome specific probes to centromeric repeti-tive DNA sequences, allowing the number ofcopies of a particular chromosome to beidentified in poor quality metaphases or ininterphase nuclei. Used in appropriatecircumstances, both approaches greatlyenhance the karyotype information which maybe obtained from cytogenetic preparations,4'and recent reports have described the applica-tion of these methods to solid tumours. Selleriet al,42 for example, have used a panel ofcosmid clones from distal llq to localise thebreakpoints of the t(11;22) seen in Ewing'ssarcoma and related tumours. Other workershave used panels of centromere specific satel-lite DNA probes to identify monosomies andtrisomies in brain tumour interphase nuclei.43Techniques have also been developed for com-bined cell morphology, immunophenotyping,and chromosome in situ hybridisation,44 45although these have yet to be applied to largeseries of solid tumours.The potential clinical contribution of solid

tumour cytogenetics to diagnosis and definingprognostic subgroups has therefore been estab-lished. The number of United Kingdom cyto-genetics centres involved in this research isincreasing, and new disease specific karyotypeabnormalities continue to be reported, permit-ting the identification of sites of oncogenic

importance within the genome and acting as aspringboard for molecular studies. This is arapidly evolving area in which active collabora-tion among pathologists, oncologists, andcytogeneticists is producing both clinicallyimportant information and insights into thefundamental biology of solid cancers.

Thanks are due to Dr A J Malcolm for his comments on themanuscript.

1 Alimena G, De Cuia MR, Diverio D, Gastaldi R, Nanni M.The karyotype of blastic crisis. Cancer Genet Cytogener1987;26:39-50.

2 Schouten HC, SangerWG, Weisenburger DD, Anderson J,Armitage JO. Chromosomal abnormalities in untreatedpatients with non-Hodgkin's lymphoma: associationswith histology, clinical characteristics, and treatmentoutcome. Blood 1990;75:1841-7.

3 Secker-Walker LM, Chessells JM, Stewart EL, SwansburyGJ, Richards S, Lawler SD. Chromosomes and otherprognostic factors in acute lymphoblastic leukaemia: along-term follow up. B J Haematol 1989;72:336-42.

4 Yunis JJ, Lobell M, Arnesen MA, et al. Refined chromosomestudy helps define prognostic subgroups in most patientswith primary myelodysplastic syndrome and acute myelo-genous leukaemia. BJ Haematol 1987;68:189-94.

5 Rowley JD. The Philadelphia chromosome translocation. Aparadigm for understanding leukaemia. Cancer 1990;65:2178-84.

6 Haluska FG, TsujimotoY, Croce CM. Oncogene activationby chromosome translocation in human malignancy. AnnRev Genet 1987;21:321-45.

7 Trent J, Crickard K, Gibas Z, et al. Methodologic advancesin the cytogenetic analysis of human solid tumours.Cancer Genet Cytogenet 1986;19:57-66.

8 Sandberg AA, Turc-Carel C. The cytogenetics of solidtumours. Relation to diagnosis, classification and pathol-ogy. Cancer 1987;59:387-95.

9 Sandberg AA, Turc-Carel C, Gemmill RM. Chromosomesin solid tumours and beyond. Cancer Res 1988;48:1049-59.

10 Nowell PC. Cytogenetics of tumour progression. Cancer1990;65:2172-7.

11 Barrios L, Miro R, Caballin MR, et al. Cytogenetic effectsof radiotherapy. Breakpoint distribution in induced chro-mosome aberrations. Cancer Genet Cytogenet 1989;41:61-70.

12 Bown NP, Roberts DF. Chromosome damage. In: D MDavies, ed. Adverse drug reaction textbook. Oxford: OxfordMed Publications, 1991:46-61.

13 Sawyer JR, Swanson CM, Chadduck WM, Roloson GJ.Evolution of tumour chromosome abnormalities aftertherapy in a paediatric astrocytoma. Cancer Genet Cytoge-net 1991;53:119-23.

14 Hayashi Y, Hanada R, Yamamoto K, Bessho F. Chromo-some findings and prognosis in neuroblastoma. CancerGenet Cytogenet 1987;29:175-7.

15 Kaneko Y, Kanda N, Maseki N, et al. Different karyotypicpatterns in early and advanced stage neuroblastomas.Cancer Res 1987;47:311-8.

16 Brodeur GM, Seeger RC. Gene amplification in humanneuroblastomas: basic mechanisms and clinical implica-tions. Cancer Genet Cytogenet 1986;19: 101-1 1.

17 Bourhis J, DeVathaire F, Wilson GD, et al. Combinedanalysis of DNA ploidy index and N-myc genomiccontent in neuroblastoma. Cancer Res 1991 ;51:33-6.

18 Christiansen H, Lampert F. Tumour karyotype discrim-inates between good and bad prognostic outcome inneuroblastoma. Br 7 Cancer 1988;57: 121-56.

19 Cowell JK. Double minutes and homogeneously stainingregions: gene amplification in mammalian cells. Ann RevGenet 1982;16:21-59.

20 Whang-Peng J, Triche TJ, Knutsen T, et al. Cytogeneticcharacterisation of selected small round cell tumors ofchildhood. Cancer Genet Cytogenet 1986;21:185-208.

21 Donner LR. Cytogenetics and molecular biology of smallround-cell tumors and related neoplasms. Current status.Cancer Genet Cytogenet 1991;54:1-10.

22 Wang-Wuu S, Soukup S, Ballard E, Gotwals B, Lampkin B.Chromosomal analysis of sixteen human rhabdomyo-sarcomas. Cancer Res 1988;48:983-7.

23 Womer RB. The cellular biology of bone tumors. ClinOrthop Rel Res 1990;262:12-21.

24 Brichard B, Ninane J Gosseye S, Verellen-Dumoulin C,Rhodhain J, Cornu G. A disseminated alveolar rhabdo-myosarcoma in a 9-year-old boy disclosed by chromoso-mal translocation (2;13)(q35;ql4). Paediatr Haem Oncol1991;8:215-19.

25 Fletcher JA, Kozakewich HP, Hoffer FA, et al. Diagnosticrelevance of clonal cytogenetic aberrations in malignantsoft-tissues tumors. N Engl _J Med 199 1;324:436-42.

26 Jenkins RB, Kimmel DW, Moertel CA, et al. A cytogeneticstudy of 53 human gliomas. Cancer Genet Cytogenet1989;39:253-79.

27 Bigner SH, Mark J, Friedman HS, Biegel JA, Bigner DD.Structural chromosome abnormalities in human medul-loblastoma. Cancer Genet Cytogenet 1988;30:91-101.

559

on 13 June 2019 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.7.556 on 1 July 1992. D

ownloaded from

Bown

28 Griffin CA, Hawkins AL, Packer RJ, Rorke LB, EmanuelBS. Chromosome abnormalities in pediatric braintumours. Cancer Res 1988;48:175-80.

29 Heim S, Mandahl N, Jin Y, et al. Trisomy 7 and sexchromosome loss in human brain tissue. Cytogenet CellGenet 1989;52:136-8.

30 Elfving P, Cigudosa JC, Lundgren R, et al. Trisomy 7,trisomy 10, and loss of theY chromosome in short-termcultures of normal kidney tissue. Cytogenet Cell Genet1990;53: 123-5.

31 Teyssier JR, Ferre D. Frequent clonal chromosome changesin human non-malignant tumours. Int Cancer1989;44:828-32.

32 Mertens F, Heim S, JinY, et al. Basosquamous papilloma. Abenign epithelial skin tumor with multiple cytogeneticclones. Cancer Genet Cyrogenet 1989;37:235-9.

33 Kawasaki R, Caldeira LF, Andre FS, et al. Multiplecytogenetic clones in a basal cell carcinoma. Cancer GenetCytogenet 199 1;54:33-8.

34 Heim S, JinY, Mandahl N, et al. Multiple unrelated clonalchromosome abnormalities in an in situ squamous cellcarcinoma of the skin. Cancer Genet Cytogenet 1988;36:149-53.

35 Atkin NB, Baker MC. Squamous cell carcinomas of thehead, neck and skin. Monoclonal or polyclonal origin?Cancer Genet Cytogenet 1991;54:135-6.

36 Rowley JD. Molecular cytogenetics: Rosetta stone forunderstanding cancer. Cancer Res 1990;50:3816-25.

37 Miller DM, Blume S, Borst M, et al. Oncogenes, malignanttransformation and modern medicine. Am Med Sci1990;300:59-69.

38 Friend SH, DryjaTP,Weinberg RA. Oncogenes and tumoursuppressing genes. N EngliJMed 1988;318:618-22.

39 Cowell J, Pritchard J. The molecular genetics of retino-blastoma and Wilms' tumour. CRC Crit Rev OncolHaematol 1987;7:153-68.

40 Cowell JK. The genetics of retinoblastoma. B Cancer1991;63:333-6.

41 Nederlof PM,Van der Flier S, Raap AK, et al. Detection ofchromosome aberrations in interphase tumor nuclei bynon radioactive in situ hybridization. Cancer Genet Cywoge-net 1989;42:87-98.

42 Selleri L, Hermanson GG, Eubanks JH, Lewis KA, EvansGA. Molecular localization of the t(l1;22)(q24;ql2)translocation of Ewing sarcoma by chromosomal in situsuppression hybridization. Proc Nail Acad Sci USA1991;88:887-9 1.

43 Arnoldus EPJ, Noordermeer IA, Boudewijn Peters AC, etal. Interphase cytogenetics of brain tumours. GenesChromosomes and Cancer 1991;3:101-7.

44 Knuutila S, Teerenhovi L Immunophenotyping of aneu-ploid cells. Cancer Genet Cytogenet 1989;41:1-17.

45 Perez A, Sole F, Caballin R, Tarrida N, Woessner S.G-banding improvement for the MAC method. CancerGenet Cytogenet 1991;51:41-4.

560

on 13 June 2019 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.7.556 on 1 July 1992. D

ownloaded from