the genetic analysis of cancer

Journal of Internal Medicine 1998; 243: 413–417

© 1998 Blackwell Science Ltd 413

Introduction

In a review article entitled ‘How Cancer Arises’ thatappeared in Scientific American at the end of 1996,Robert Weinberg stated ‘how cancer develops is nolonger a mystery’ [1]. Of course, Weinberg was refer-ring to the incredible increase in information accu-mulated mainly during the last 20 years about themolecular processes involved in oncogenesis.However, to claim that this is no longer a mystery iscertainly oversimplifying the situation. Whilst we nowknow that cancer is a complex of genetic diseases

arising from mutations occurring in a wide variety ofgenes, the issues of which combination of genes isinvolved for any given tumour, what determines theirtissue specificity, and the order in which mutationsoccur are mysteries that are likely to remain with usfor a long time to come. As more oncogenes andtumour suppressor genes are isolated, understandingthe way in which these genes are regulated and howthey interact with each other to produce the cancerphenotype becomes increasingly more difficult. Givenour present rudimentary or almost complete lack ofknowledge on the function of many of the genes

The genetic analysis of cancer

P. L. PEARSON & R. B. VAN DER LUIJTFrom the Department of Human Genetics, Utrecht University, Utrecht, the Netherlands

MINISYMPOSIUM MEN & VHL

Abstract. Pearson PL, Van der Luijt RB (UtrechtUniversity, Utrecht, the Netherlands). The geneticanalysis of cancer (Minisymposium: MEN & VHL). JIntern Med 1998; 243: 413–17.

During the past two decades an overwhelmingamount of knowledge has been acquired on the mol-ecular genetics of human cancer. It is now evidentthat cancer is essentially a genetic disease, arisingfrom inherited and/or somatically acquired muta-tions at different genetic loci, and that tumourigene-sis is a multistep process. Gene mapping studies ofinherited cancer syndromes have resulted in theidentification of many genes implicated in the initia-tion of tumours. Importantly, alterations of the samegenes were also found to play a role in the develop-ment of common, non-familial tumours. The genesinvolved belong to distinct functional classes, andinclude proto-oncogenes and tumour suppressorgenes, which are regulators of cellular growth and

proliferation, cell adhesion and programmed celldeath. Another class of cancer susceptibility genesconsists of DNA repair genes, which are involved inmaintaining genomic stability. In unravelling thegenetic basis of cancer, the localization and identifi-cation of genes involved in tumourigenesis can beconsidered as the ‘easy’ part; determination of thenormal physiological function of these genes andtheir precise role in tumourigenesis has proved to bemuch more difficult. In this review, we highlightsome of the major breakthroughs in the field of can-cer genetics, and discuss recent insights in the puta-tive role of proto-oncogenes, tumour suppressorgenes and DNA repair genes in the initiation and pro-gression of cancer. Also, we point to some of thechallenges to be faced in the coming years.

Keywords: cancer genetics, DNA repair genes, hered-itary cancer, multistep process, proto-oncogenes,tumour suppressor genes.

P. L. PEARSON & R. B. VAN DER LUIJT: THE GENETIC ANALYSIS OF CANCER414

© 1998 Blackwell Science Ltd Journal of Internal Medicine 243: 413–417

involved, current interpretations on how cancer aris-es are certainly grossly oversimplified. Whilst we havebeen doing a reasonable job of determining whichgenes are involved, the much more difficult task ofdefining their function in the cascade of cell regula-tory changes leading to cancer still lies ahead.

The study of inherited forms of cancer provides valu-able starting points for dissecting which genes areinvolved in the initiation of cancer and providesinsights into their interactions with other genes. Amajor factor limiting the meeting of these aims is howthe necessary collections of patient material can bemade available to the scientific community in an effi-cient and timely fashion. Workshops such as the min-isymposium on multiple endocrine neoplasia type 1(26–28 June 1997; Noordwijkerhout, the Netherlands)are important forums for scientists to meet andexchange ideas with a small core of extremely hard-working and enthusiastic physicians who recognize theneed to assemble materials from families with inheritedcancers and who are also willing to collaborate withbasic research scientists in isolating the genes con-cerned. Often the final isolation of such genes dependson the availability of a small number of crucial patientswho provide the breakpoint or genetic recombinationdata to narrow a chromosome region down to levelswhere gene hunting or investigation of candidate geneswithin the region becomes a feasible proposition.

The modern era of cancer genetics started with thein vitro experiments of Harris et al. [2] in which a loss oftransformed phenotype following cell fusion betweencancer and normal cells was observed and it was con-cluded that cancer was a recessive phenotype. This wassubsequently confirmed in many other in vitro studiesduring the following decade, using single human chro-mosome somatic cell hybrids as the ‘normal’ fusionpartner and demonstrating that suppression of atransformed phenotype was frequently chromosome-specific for the transformed phenotype concerned, sug-gesting involvement of different genes. This was aconundrum since the pattern of familial inheritancefor the majority of inherited tumours was overwhelm-ingly autosomal dominant. In 1971 Knudson present-ed his ‘two-hit’ hypothesis for the origin of cancerbased on an epidemiological survey of retinoblastoma[3]. The hypothesis predicted that an initial mutationwas already present and transmitted in the germ-lineof families exhibiting hereditary cancer predispositionand that a second mutation occurred somatically inthe tissue giving rise to the cancer and resulting in

mutation of both alleles of a single gene. However, itwas another 12 years before Cavenee et al. (1983) for-mally demonstrated the mechanisms which result inthe occurrence of the second mutation including lossof heterozygosity resulting from chromosome deletionor non-disjunction [4]. This provided an explanationfor the apparent dominant inheritance of cancer infamilies: there is either a high mutation frequency, or anormal mutation frequency in tissues with a highturnover, such as retina or colon, or a combination ofthe two, resulting in the occurrence of the secondmutation in almost all cases carrying the germ-linemutation and producing genetic segregation approxi-mating the 50% expected of dominant disorders. Theirimportant conclusion was that although there is apseudo-dominant inheritance at the cancer phenotypelevel, at the cellular level it is recessive.

Two years later, in 1985, another category ofgenes involved in oncogenesis was recognized follow-ing transformation of mouse 3T3 cells with DNAderived from various human solid tumours and iso-lating the human gene sequences concerned [5].These genes, referred to as oncogenes, requiredmutation in only a single allele to exert their influ-ence and they therefore fulfil the requirements oftrue dominant inheritance. In retrospect we nowrealize that the cancer susceptibility genes first notedby Harris et al. [2] and predicted by Knudson [3] aretumour suppressor genes which require a completeloss of function by inactivation of both alleles, inorder to initiate oncogenesis. The oncogenes detectedlater on, harbour genetic changes which result in again of function as expected from dominant muta-tions. It now appears that the normal function of themany tumour suppressor genes and oncogenes iden-tified to date cover all aspects of cellular control andtissue development varying from embryonic develop-ment (mainly oncogenes) to aspects of signallingpathways involving cell contact and interaction,transcription, cell cycle control and programmed celldeath (both oncogenes and tumour suppressorgenes). Additionally, both DNA mismatch andnucleotide excision repair genes have also been impli-cated in oncogenesis and their loss of function leadsto greatly increased mutation rates, presumably forall genes but with particular importance to thoseinvolved in oncogenesis. We may conclude that can-cer is a polygenic disorder arising from an accumula-tion of mutations in several genes which can belongto different functional categories.

MINISYMPOSIUM: MEN & VHL 415


The mutation which initiates the chain of geneticchanges leading to particular types of cancerappears to be gene- and tissue-specific. It is thismutation which can be tracked in families withinherited tumours and, with the exception of theoncogenes MET and RET, the genes involved aretumour-suppressor or DNA mismatch-repair genes.Subsequent mutations occur in both tumour sup-pressor and oncogenes. Although some genes, suchas P53 and HRAS, are mutated in many differenttypes of tumour and clearly play a wide-ranging rolein oncogenesis, apart from the initiating gene muta-tion, exactly which mix and match of subsequentmutations gives rise to a particular type of cancer isdifficult to predict. Kinzler & Vogelstein [6, 7] haverecently proposed an alternative cancer susceptibilitygene definition, namely that of dividing the genesinto ‘gatekeeper’ and ‘caretaker’ genes. Gatekeepergenes are those directly controlling cellular prolifera-tion. Their characteristics include: inhibition of

tumour growth or promotion of cell death, tissuespecificity, with one or a few gatekeeper genes foreach tissue. For many inherited cancers, oncogenesisis initiated with one additional, somatically acquiredmutation involving the second gatekeeper gene allele.Many tumour suppressor genes are assumed to func-tion as gatekeeper genes for a specific cell or tissuetype. Examples of gatekeeper genes include theretinoblastoma [8], adenomatous polyposis coli[9,10] and most likely also the multiple endocrineneoplasia (MEN) type 1 [11], and von Hippel–Lindau[12] genes, which are involved in tumours of theretina, intestinal tract, parathyroid glands and kid-neys, respectively. However, in our view, apart fromtumour suppressor genes, proto-oncogenes mayoccasionally also act as gatekeeper genes for specifictissues. Their role in familial cancer syndromes clear-ly shows that germ-line mutations of proto-onco-genes can also directly initiate oncogenesis: inheritedmutations of the RET and MET proto-oncogenes are

P53 (?)BRCA1 (?)BRCA2 (?)

CARETAKER PATHWAY

GKGKCTCT

GKGK

GK

4

2

1

3

1

21

Genetic mutation or 'hit'

GATEKEEPERPATHWAY

GATEKEEPERPATHWAY

NE

OP

LAS

TIC

CE

LL

NO

RM

AL

CE

LL

MA

LIG

NA

NT

TU

MO

URFurther

accumulationof mutations

Tumour progressionTumour initiation

MSH2

DNA REPAIR GENESas caretakers (CT):

ATM (?)MLH1 P53 (?)PMS1 BRCA1 (?)PMS2 BRCA2 (?)

MEN1

TUMOUR SUPPRESSORGENES as gatekeepers(GK):

VHLRBAPC

RET

PROTO-ONCOGENESas gatekeepers (GK):

MET

Fig. 1 Schematic representation of the different pathways to tumour initiation, involving gatekeeper (GK) and caretaker (CT) genes. In thepathway in which a proto-oncogene acts as a gatekeeper (top), only one mutation is required to initiate oncogenesis. In contrast, if a tumoursuppressor gene functions as a gatekeeper (middle), mutations of both gatekeeper alleles are required. Finally, in the caretaker pathway(bottom) a total of four mutations, resulting in the inactivation of two alleles of a caretaker gene (causing genetic instability) and two allelesof a gatekeeper gene (leading to loss of growth control), are necessary to drive oncogenesis. Modified from Kinzler & Vogelstein [7].

P. L. PEARSON & R. B. VAN DER LUIJT: THE GENETIC ANALYSIS OF CANCER416


associated with MEN type 2 and hereditary papillaryrenal carcinoma, respectively [13, 14]. Comparedwith tumour suppressor genes, in which inactivationof both alleles is required for tumour initiation, in thecase of proto-oncogenes only a single mutation is suf-ficient to initiate oncogenesis (see Fig. 1). Caretakergenes are involved in maintaining genetic stabilityand preventing undue mutation. They include theDNA mismatch repair genes (germ-line mutations ofeither MLH1, MSH2, PMS1, or PMS2 in hereditarynon-polyposis colorectal cancer), nucleotide-excisionrepair genes (ERCC2, 3 and 5 mutations in xeroderma pigmentosum), and probably also theATM gene (involved in ataxia telangiectasia). In con-trast to gatekeeper genes, inactivation of a caretakergene in itself does not promote tumour initiationdirectly.

Recent discussions on the possible functions of thebreast cancer susceptibility genes BRCA1 and BRCA2[7, 15] have considered the binding of the BRCA1gene product to RAD51, a well-defined DNA doublestrand break repair enzyme, and the increased radia-tion sensitivity of embryonal cells derived from Brca2(the murine homologue of BRCA2)-deficient ‘knock-out’ mice. These observations have raised the possi-bility that the BRCA genes may function ascaretakers, apart from their assumed roles as gate-keepers in breast epithelial cells in families withBRCA1 or BRCA2 germ-line mutations. If BRCA1and BRCA2 do indeed act as caretakers under somecircumstances, the inherited forms of breast cancerwould need to be initiated by at least one additionalmutation that inactivates the remaining caretaker(BRCA1 or BRCA2) allele, and two alleles of a differ-ent gatekeeper gene for breast cancer that stillremains to be identified (see fig. 1). Such discussionshighlight one of the major problems facing molecu-lar biologists, namely that the potential role of manytumour susceptibility genes as either gatekeepers orcaretakers cannot be reliably assessed because theirfunctions are currently so ill-defined. Many cancersusceptibility genes, e.g. BRCA1 and BRCA2, arelarge and have many potentially functional domainswith disparate functions. Currently BRCA1, besidesits association with RAD51, is believed to contain atranscription activation domain known to be com-promised by mutations. It can inhibit cell prolifera-tion when over expressed and localizes to thesynaptonemal complex during meiosis (possibly alsoa RAD51 association). Some or all of these functions

may be involved in oncogenesis and much work onfunctional analysis must take place before we caninterpret gene function and interaction reliably.Interestingly, in a fashion similar to the BRCA genes,P53, another enigmatic cancer gene, may functionboth as a gatekeeper and a caretaker. Germ-line P53mutations initiate oncogenesis in patients withLi–Fraumeni Syndrome, suggesting a gatekeeperfunction. Somatic mutations and/or loss of the P53gene are common events during progression of awide variety of both familial and sporadic tumours.Normal P53 has been shown to function as a cellcycle checkpoint, determining cell fate (programmedcell death or survival) following DNA damage. P53-mediated cell cycle arrest is believed to occur to allowmore time to perform DNA damage repair. If therepair fails, programmed cell death or apoptosisoccurs, thereby preventing survival of cells with pro-found DNA damage. In tumour cells, the P53-medi-ated cellular response to DNA damage is frequently‘knocked out’, allowing tumour cells to escape thiscrucial checkpoint and imparting an immortality totumour cells not found in normal tissue. The role ofP53 as ‘guardian of the genome’ maintaininggenomic stability, clearly points towards a caretakerfunction.

With respect to the MEN-1 and VHL genes,although these genes are considered as gatekeepers,at present it cannot be excluded that they may (also)function as caretakers.

In many respects the easy genetic work, that ofmapping and isolating highly penetrant cancer sus-ceptibility genes, has now been largely completed. Itis inevitable that for the cancer genes alreadymapped, the emphasis will shift from genetics tofunctional analysis. However, there are many othercancers in which the initiating genetic componentshave still not been defined, probably because of lowerpenetration and failure to recognize familial cluster-ing. Genetic analysis must shift towards recognizingsuch low penetrant susceptibility genes and new par-adigms developed for detecting non-Mendelian clus-tering and the influence of modifier genes andenvironmental parameters. New methods of epi-demiological analysis will be required to accomplishthis. In addition, since the mapping information aris-ing from such approaches will inevitably be far lessprecise, more emphasis will have to be placed indefining candidate genes by function and not by mapposition.

MINISYMPOSIUM: MEN & VHL 417


References1 Weinberg RA. How cancer arises. Sci Am 1996; 275: 62–70.2 Harris H, Miller OJ, Klein G, Worst P, Tachibana T. Suppression

of malignancy by cell fusion. Nature 1969; 223: 363–8.3 Knudson Jr AG. Mutation and cancer: statistical study of

retinoblastoma. Proc Natl Acad Sci USA 1971; 68: 820–23.4 Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R,

Gallie BL et al. Expression of recessive alleles by chromosomalmechanisms in retinoblastoma. Nature 1983; 305: 779–84.

5 Weinberg RA. The action of oncogenes in the cytoplasm andnucleus. Science 1985; 230: 770–76.

6 Kinzler KW, Vogelstein B. Lessons from hereditary colorectalcancer. Cell 1996; 87: 159–70.

7 Kinzler KW, Vogelstein B. Cancer-susceptibility genes.Gatekeepers and caretakers. Nature 1997; 386: 761–3.

8 Lee WH, Bookstein R, Hong F, Young LJ, Shew JY, Lee EY.Human retinoblastoma susceptibility gene: cloning, identifica-tion, and sequence. Science 1987; 235: 1394–9.

9 Kinzler KW, Nilbert MC, Su LK, Vogelstein B, Bryan TM, LevyDB et al. Identification of FAP locus genes from chromosome5q21. Science 1991; 253: 661–5.

10 Groden J, Thliveris A, Samowitz W, Carlson M, Gelbert L,Albertsen H et al. Identification and characterization of thefamilial adenomatous polyposis coli gene. Cell 1991; 66:589–600.

11 Chandrasekharappa SC, Guru SC, Manickam P, Olufemi SE,Collins FS, Emmert-Buck MR et al. Positional cloning of thegene for multiple endocrine neoplasia type 1. Science 1997;276: 404–7.

12 Latif F, Tory K, Gnarra J, Yao M, Duh FM, Orcutt ML et al.Identification of the von Hippel–Lindau disease tumor sup-pressor gene. Science 1993; 260: 1317–20.

13 Mulligan LM, Kwok JBJ, Healey CS, Elsdon MJ, Eng C, GardnerE et al. Germ-line mutations of the RET proto-oncogene inmultiple endocrine neoplasia type 2A. Nature 1993; 363:458–60.

14 Schmidt L, Duh FM, Chen F, Kishida T, Glenn G, Choyke P et al.Germline and somatic mutations in the tyrosine kinasedomain of the MET proto-oncogene in papillary renal carcino-mas. Nat Genet 1997; 16: 68–73.

15 Marx J. Possible function found for breast cancer genes. Science1997; 276: 531–2.

Received 25 November 1997; accepted 22 January 1998.

Correspondence: Dr Peter L. Pearson, Department of HumanGenetics, Utrecht University, PO Box 80030, 3508 TA Utrecht, theNetherlands (fax: 130 253 8186; e-mail:[email protected]).

the genetic analysis of cancer

Documents