Predictive testing for multiple endocrine neoplasia type 1 using DNA polymorphisms

Predictive testing for multiple endocrine

neoplasia type 1 using DNA polymorphisms.

C Larsson, … , T Tokino, B Seizinger

J Clin Invest. 1992;89(4):1344-1349. https://doi.org/10.1172/JCI115720.

Multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominantly inherited predisposition to neoplastic lesions of the parathyroids, pancreas, and the pituitary. We have previously located the predisposing genetic defect to the long arm of chromosome 11 by genetic linkage. In this study, 124 members of six MEN1 families, including 59 affected individuals, were genotyped for restriction fragment length polymorphisms with different DNA probes, and the genetic linkage between these marker systems and MEN1 was determined. 13 marker systems (17 DNA probes) were found to be linked to MEN1. These markers are located within a region on chromosome 11 spanning 14% meiotic

recombinations, with the MEN1 locus in the middle. Four of the marker systems are on the centromeric side of MEN1, and four on the telomeric side, based on meiotic crossovers. The remaining five DNA probes are closely linked to MEN1, with no crossovers in our set of families. The 13 marker systems can be used for an accurate and reliable premorbid test for MEN1. In most clinical situations it is possible to identify a haplotype of this part of

chromosome 11 with the mutant MEN1 allele in the middle. The calculated predictive accuracy is greater than 99.5% if three such marker systems are informative. Therefore, genetic linkage testing can be used for informed genetic counseling in MEN1 families, […]

Research Article

Find the latest version:

Predictive Testing for Multiple Endocrine Neoplasia

Type 1 Using DNA Polymorphisms

Catharina Larsson,* JosephShepherd,*YusukeNakamura,'Carl Blomberg,* GuntherWeber,* Barbro Werelius,*Nick

Hayward,11 Bin Teh,t TakahashiTokino,' Bernd Seizinger,' Britt Skogseid,** KjellOberg,**** and MagnusNordenskjdld*

*Department ofClinicalGenetics,KarolinskaHospital,S-10401Stockholm,Sweden;tDepartment ofSurgery, UniversityofTasmania,

Hobart, Australia; ODivisionof Biochemistry, Cancer Institute, Tokyo,Japan; I1QueenslandInstituteofMedicalResearch, Brisbane,

Qld, Australia; 'Molecular Neuro-Oncology Laboratory, Massachusetts General Hospital, Boston,Massachusetts; **Departmentof Internal Medicine, UniversityHospital, Uppsala,Sweden; andt*LudwigInstituteforCancerResearch,

University Hospital, Uppsala, Sweden

Abstract

Multiple endocrineneoplasia type 1 (MEN1)is an autosomal

dominantly inherited predispositionto neoplasticlesions of the

parathyroids,pancreas, and thepituitary.Wehave previously located thepredisposinggenetic defect to the long arm of chro-mosome 11 bygenetic linkage.In this study, 124 members of

six MENi families, including 59 affected individuals, were

genotypedfor restriction fragmentlength polymorphisms with

differentDNA probes, andthegenetic linkagebetween these marker systems and MENi was determined. 13 marker sys-tems (17 DNA probes) were found to be linked to MEN1. Thesemarkers arelocated withinaregiononchromosome 11

spanning14%meiotic recombinations, withtheMENi locus in

themiddle. Four ofthemarkersystems are on thecentromeric side of MEN1, and fouronthetelomericside, basedonmeiotic

crossovers.Theremaining fiveDNAprobesarecloselylinked

to MEN1, with no crossovers in our setof families. The 13 marker systems canbeusedforan accurateand reliable

pre-morbidtestfor MEN1.In mostclinical situations it is possible

toidentifyahaplotype of thispartofchromosome 11withthe mutant

MENi

alleleinthemiddle.The calculatedpredictive

accuracyis>99.5% if threesuchmarkersystems are informa-tive.Therefore, genetic linkage testingcanbe usedforinformed genetic counselinginMENi families,andtoavoidunnecessary

biochemical screening programs. (J. Clin. Invest. 1992. 89:1344-1349.) Key words: cancer predisposition *

chromo-some 11*genetic linkage*restrictionfragmentlength polymor-phism* tumorsuppressor gene

Introduction

Multiple endocrine

neoplasia

type 1 (MEN

1)1

isa

genetic

dis-order characterized by neoplastic lesions in the parathyroid glands, theneuroendocrinepancreas,and the anterior lobe of

AddresscorrespondenceandreprintrequeststoDr.CatharinaLarsson,

Dept. of ClinicalGenetics, KarolinskaHospital,S- 104 01 Stockholm,

Sweden.

Receivedfor publication 01July1991 and in_revisedform30

Sep-tember1991.

1.Abbreviations usedinthis paper:MEN1,multipleendocrine

neopla-sia type 1;RFLP,restrictionfragmentlengthpolymorphism.

thepituitary gland (1, 2). Wepreviously mapped the MEN 1

locustochromosomalregion 1lql-q1 3 by genetic linkage to

restriction fragmentlengthpolymorphism (RFLP)markers in

familiessegregating the disease (3).Byidentifying meiotic cross-oversinMEN 1 families,wewereableto locate thediseasegene

closetothe PYGM locus andflankedby theDl1S288 locus

located7cM onthe centromeric side,and by INT2 7 cM on the

telomeric side ofthis marker (4) (Fig. 1).Furthermore, compari-sons of the genotypes in normal and tumortissue from the sameindividualshowed thatthe pathogenesis ofpancreatic, as well asparathyroid lesions, involve unmasking of a recessive

mutationattheMEN 1 locus (3, 5-8), in agreement with the

two-mutational modeloriginally postulatedby Knudson (9).

Deletion mapping oftwo MEN1-associated lesions suggested

that theMEN 1 locus istelomericto PYGM, and within the 5

cMintervalbetween PYGM and Dl15146(4, 7).

Based on the autosomal dominant mode of inheritance, offspring ofMEN lpatientshavearisk of closeto50% ofdevel-opingthedisease.Affected individualsmay presentwithanyof theMEN 1-associated lesions in earlyteenage, or escapeclinical

symptoms for several decades. Children ofMEN 1 patients usually undergo repeated biochemicalscreenings in orderto

identify lesions as early as possible (10). However, since

MEN 1-associated lesionsmaydevelop slowly,unaffected rela-tivesatriskcannotbe excludedaslateonsetgenecarriers until theageof35yr,despiteextensive biochemicaltesting.Hence, there isacallfor informedgenetic counselinginMEN1 fami-lies, and an avoidance ofunnecessary biochemical screening

tests. We have attemptedto develop areliable andclinically usefulpredictive testingforMEN1basedongenetic linkageto

DNAmarkers.

Suchtesting procedures can make useofthe detailed

ge-netic mapping data recently generated for the chromosomal region 1Iq12-q13. This involves isolation of new genetic markers whicharethenorderedbygenetic linkageinreference families, physically,on somaticcellhybrids,orbylong-range restriction mapping using pulsed-field gel

electrophoresis

(4,

11-14).Wehave testedgenetic linkagetothe disease

using

a number ofDNAprobes whichwehavemapped

physically

to

thechromosomeregion

harboring

theMEN1 gene. Extensive family material with 59MENI casesfrom five Swedish fami-lies and onelarge kindred from Tasmania was

analyzed.

In

total, 13markersystems(17probes)showed

significant

linkage

to MEN 1. Ofthese, 11 probeshadnot

previously

been tested

forlinkagetothedisease.

Five ofthese probesare located closeto thediseasegene

withnomeioticcrossoversdetected,whilefourmarkersystems flankMEN 1 onthecentromeric

side,

andfour marker systems

J.Clin. Invest.

©TheAmericanSocietyfor ClinicalInvestigation,Inc.

0021-9738/92/04/1344/06 $2.00

aretelomeric. The accuracy and usefulness of DNA-based

test-ing for MEN1 predispositionis illustrated ina setoffamilies suitableforsuchdiagnosis.

Methods

Clinical cases. Three of the five Swedish families with MEN 1(KA,EZ, LU) havebeen describedelsewhere(3). TheAXandRAfamilieswere

referred for clinical testing, and are shown inFig.2.Inadditiontothese fivefamilies, part of one large MEN 1 pedigreefirst reported by

Shep-herd in 1985 from Tasmaniawasstudied(15).Thisfamilyhas previ-ouslynotbeen studiedbylinkage. By the end of 1989,over150affected individuals from 2,300 family members spanning eight generations had been traced back to a womanwho cametoTasmania in the middle of the last century (16). Her description recorded by the ship's surgeon is consistent with the diagnosis of MEN1.The scoring for MEN 1was

conservative in ordertoavoidincorrect diagnosis. Unaffected individ-ualsatriskwerescoredas"normal" only if they had passed the age of 35 years and had undergone repeated and extensive biochemical screenings(10) withnosigns of the disease, while offsprings to MEN 1

caseswho didnotfulfill these criteriawerescoredas"unknown."

TableI. DNAMarkers LinkedtotheMEN] Locus

Locus Probe Constant Allele Allele Observed

name name Enzyme bands sizes freq. heterozygosity

Dl1S288 D11S149 p3C7 pTHH26 MspI 5.7 3.1 5.2 3.2 PvuII

Haplotype for markers Dl1S288and Dl 1S149

CD20 pB 1-21

Dl1S480 PGA PYGM Dl lS427 DllS750 Dl lS449 DllS97 D1lS146 cCI11-319§ pcAGP9 cCLPGA12§ pMCMPl§

cCL 15§

cCI11-4§ cCLGW4§ cCI11-219§ pMS51 pHBI59 cCL59§

Haplotype for markerDlIS 146

D 1S751 cCL-8§

clone 8

Haplotype for marker D1IS751

INT2 SS6 MspI PstI 2.8 several EcoRl PvuII Mspl EcoRI EcoRI Sad PvuII

4.4; 3.4; 3.0; 2.8; 2.1; 1.9; 1.4; 1.3;1.1

14;6.6; 5.4;

21; 16

11.3;6.4; 4.4;

1.6, 1.5; 1.3 several

HaeIII

TaqI

EcoRV

EcoRV

24;18; 7.4

20;5.2; 2.2

EcoRI

TaqI

BamHI

Haplotypeformarker INT2

9.0 6.0 3.0 1.4

19.0

16.6+ 15+ 3.5 16.6+ 15+ 12+ 3.5

11

6+ 5 2.2-2.5 5 alleles VNTR 19.0

17.5 9.0-9.6 4 alleles VNTR 3.6

2.4 3.7 3.0

1.3-4.3 9 alleles VNTR

1.2 0.8 7.0 6.4 6.6 5.8 7.4 6.7 4.1 2.3 8.4 5.6+2.8 0.72 0.28 0.14 0.84 0.47 0.53 0.50 0.50 0.06 0.37 0.54 0.37 0.63 58% 0.11 0.89 60% 0.80 0.20 0.33 0.67 89% 0.55 0.45 0.22 0.78 0.50 0.50 0.89 0.11 0.60 0.40 0.63 0.37

*_{Combined haplotype of two probes.}*_{Observed in the present study. Loci in bold indicate anchor markers previously mapped on reference}

families. § Indicates probes with repetitive elements for which human DNA was included in the hybridization.

DNAprobes, genotyping, and linkage analysis. The probesused and the RFLP's detectedwiththesemarkers are described in Table I.

Dl 1S288(17),Dl 1S149(17),PYGM/pMCMPI (17), andDl 5146/ pHBI59 (17) areavailable fromAmerican Type Culture Collection (Rockville,MD). CD20 (18) was obtained from T.F. Tedder, PGA/ pcAGP9 from B. Zelle,Dl S97 (18) from ICI, INT2 (18)from N.

Spurr,andclone8/Di 1S751(14, 19)fromT.Rabbits. The newly iso-lated cosmid clonesfor DlIS480,Dl 1S427,andDl 1S449have been

published(12). Table Iincludesonlythose 17 probes which showed

significantlinkageto the MEN1 locus(see below). For 10 ofthe probes

listed previouslypublished RFLPs existed(12,17, 18). Newallele sys-tems were identifiedthrough locus expansion in cosmids for PGA (cCLPGA12), PYGM (cCL15), Dl IS146 (cCL59), and Dl1IS751

(cCL8), and the newcosmid DllS750/cCLGW4 was isolated from a library constructedfromradiation-reduced somatic cell hybrids

con-tainingthe 1lqI I-q 13region( 14).

The gene ordergiven forthe markers in Table Ifrom the centro-mere (top)isaccordingtothe combined datafor RFLP mapping using

referencefamilies(4, 1 1),physicalmapping datafromhybrid cell

pan-els,andlong-rangerestrictionenzymemapping with pulsed-fieldgel

electrophoresis (13, 14). Seven "anchor"markersindicatedinboldare those probesmapped bylinkageinreference families(11)(Fig. 1). In thefollowing,the same gene order is used in tables andfigures.

Onlymarker systems that gave significant linkageto MEN1 are

listedin Table I. As shown, three marker systems(Dl lS146, Dl 1S751, andINT2) withtwo separate 2-allele RFLP's werecombinedto4-allele systemsbyhaplotyping.Forthe markersDl1 S288 and D 1 S 149

pre-vious linkagedata inreference familieshave shown a very closelinkage withnomeioticcrossoverbetween these two loci (4, 1 1).Similarly,we have not detected anyrecombinationsbetween these markers in our MEN1 families.Therefore,these two markers were alsocombinedby haplotyping, in order toincrease their relative information content.

Haplotyping ofthe two markersforthe PYGM locus and the combined

DllS97+Dl lS146markersisalsopossible, althoughwehave chosen toanalyzethemseparately sincetheinformationcontent washighfor each oneofthese markers.

Methods andconditions for DNApreparation,Southernblotting, hybridizationtoradioactivelylabeled DNAprobes,and

autoradiogra-phy were asdescribed ( 14).Several ofthe markers used contained

repet-itiveDNAelements (marked with§ in Table I), andweretherefore

hybridizedin the presenceof denatured,sonicated human DNA.

Familyand genotype data weretyped intothelinkagedata

manag-ing computer database "LINKSYS." Two-point linkage between

MEN1 anddifferent markerloci was calculated withthe "LIPED" computer program, assuming full penetrancefor MEN 1 under the strictdiagnostic criteria outlinedabove.

Results

Linkage to MEN]. A number of DNA markers mapped

within,orcloseto,chromosome bands 1 q 1 -q13weretested

for linkage to the disease in six MEN1 families. 13 marker

systems gavesignificantlodscores(Table II).Threepedigrees from Sweden (LU, KA, and EZ) previously outlined in the

originalmappingofMEN1 (3)werereanalyzed,sincesome of

themarkersystemsusedmay bemoreinformative ifthe experi-mentalconditionsareadjusted.Inaddition,two newSwedish

families (AX and RA)(Fig. 2) andpart ofone large MEN1

pedigreefromTasmaniawerestudied(15).Thetotal number

ofindividuals genotyped was 124, including34MENI cases

fromTasmania and 25 cases from the Swedish pedigrees. Thetwo-pointlinkage data ofMEN 1 to thedifferent

ge-netic markersaresummarized inTable II. 13differentmarker systems gavelod scores above the conventional cut-offlevelof

3.1. Themarker systems break down into three groups based

i

U

I

U

U

U

pter

/f

11 _'s

q ter

-DI1S149

D11S288 0%

CD20 3% 4%

D11S480 9 % 4%

-PGA 3 %

-PYGM

DllS427

DlIS750

5 % D11S449

-D1iS 97 -DI1S146

-INT2 0 %

0 % 0%

0%

0 %

5 % 4%

DllS751 3 %

2%

7%

Figure1.Map ofthe MEN Iregiononchromosome 11 determined by differentmappingstrategies.Seven anchor markers areindicatedto theleft,and thegenetic linkage distances inpercentagemeiotic re-combinations betweenthese markers, accordingtoreference 11,are

given.The new markers are placed in the gene order and in relation

tothe anchormarkers,asdetermined by physicalmapping (references 13 and 14).Thelinkagedistancesin percentage crossing over to MENIarefrom TableII.

onmeioticcrossovers.Four markersystemsarecentromericto

the MEN1 locus, DlIS288/Dl S149,CD20, DlIS480, and

PGA. Inourprevious linkage studies,wedid not findany

cross-oversbetween MEN1 andPGA,but this was now found in the Tasmanian family using the new cosmid probe for PGA

(cCLPGA 12).

Four marker systems are telomeric to MENl-DlIS97,

Dl1S146,DlIS751,and INT2.Betweenthese groups of

flank-ing marker systems there are five probes closely linked to

MEN1, for which we found no crossing over. The probe pMCMPl for the PYGM locus shows the highest lod score

(I10.80)at apeakrecombination distance of0=0,in agreement

with ourpreviousstudies (3, 4). Fournewly isolated cosmid probes,cCLl5 forPYGM,Dl 1S427,Dl1S449, andDlIS750, gave lod scoresin the range of 3-5. The lower lodscores for

these markersareinpartduetothefact thattheyareless infor-mative than the multiallele VNTR detected by the PYGM probe pMCMPl.

DNAtestingforMEN]. To illustrate how DNAtestingcan beusedinclinicalsituations, weselected sixSwedish nuclear families that hadundergoneextensive biochemicaltestingfor MEN 1.Thesefamiliesareoutlined inFig.2.Ineach of these families,oneof theparentsandatleastonechildareaffected with MEN 1. Basedonthegenotypes determinedfor different markersystemslinkedtoMEN1 intwoaffectedmembers in eachkindred,we canoutline themostlikely haplotypecarrying

the MEN1 allele in each affectedparent. For the KA-2 and

TableII.LodScoresfor Linkage ofMENJ toDifferentLoci within ChromosomeRegionllqIl-q13

Recombination fraction(0)

Marker system 0 0.001 0.01 0.05 0.10 0.15 0.20 0.30 0.40 0 Z

(DlIS288/Dl S149)* -0o 4.04 5.88 6.56 6.21 5.57 4.78 2.99 1.17 0.05 6.56

CD20 -0o 2.39 3.27 3.43 3.06 2.57 2.04 1.00 0.23 0.03 3.48

DIIS480 -00 0.58 2.59 3.82 3.96 3.71 3.27 2.04 0.66 0.09 3.98

PGA(cCLPGA12) -00 2.03 2.93 3.22 3.01 2.66 2.26 1.37 0.50 0.04 3.23 PYGM(pMCMPI) 10.80 10.78 10.63 9.87 8.82 7.69 6.50 3.99 1.52 0 10.80

PYGM(cCL15) 4.26 4.25 4.17 3.85 3.45 3.04 2.62 1.70 0.72 0 4.26

D1S427 4.78 4.77 4.68 4.28 3.78 3.26 2.75 1.72 0.70 0 4.78

Dl1S750 3.57 3.56 3.52 3.27 2.91 2.49 2.03 1.08 0.25 0 3.57

D1S449 3.39 3.38 3.28 2.84 2.33 1.86 1.43 0.71 0.19 0 3.39

Dl1S97 -00 4.99 7.77 8.86 8.46 7.64 6.64 4.34 1.90 0.05 8.86

Dl1S146 -00 7.63 9.43 9.92 9.32 8.40 7.33 4.90 2.25 0.04 9.96

Dl S751(cCL8/clone8) -00 4.94 5.79 5.79 5.21 4.49 3.72 2.15 0.75 0.03 5.92

INT2 -00 1.19 4.99 6.84 6.82 6.26 5.45 3.48 1.40 0.07 6.94

*_{Combinedhaplotype ofthetwomarkers.}

criteria. Hence, inthese kindred the DNA-based testinguses

the first affected childasreference case.

Foreachadditionalchild,thecarrierstatusforMEN 1 can nowbe inferred from its haplotype of these markersystems.

This is illustrated in theRAkindred, where the first daughter whoinherited MEN1 from her mother alsoinheritedthe 2-al-lele for the PYGM marker cCL1 5, and the 3-allele for

D 1IS427.Inaddition,the1-allele for PGA and the 3-allele for Dl1S97 were linkedtothe disease. Theyoungerbrotherhad

inherited the same haplotype ashis affected sister. The most

likely interpretation ofthese data is that he has inherited a

nonrecombinant stretch of this chromosomeregion, including allloci from PGAtoDl1S97. Hence, this would also include theMEN1 mutation. The predicted MEN1 diagnosis is

con-firmed for this boy, who underwentsurgeryforprimary hyper-parathyroidism, and treated for prolactinoma. Since four markerswereused,twoverycloseto MEN 1 andtwoflanking thediseaselocus, thisfamilyillustratesnotonlytheuseof the

veryclosely linked markers cCL15 and Dl 1S427 for genetic testing, but also that meioticcrossoverscloseto MEN 1 could beexcluded.

Ifacrossoverhad occurred in the second child, this would onlyescapedetection if the firstchildalso hadinheriteda

re-combinantchromosomewith thecrossoverpoint between the

same twomarkers.Inthiscasethen,anincorrect predictionof

thediseasepredispositionin the second child could onlyoccur

if bothchildren had inherited recombinantchromosomes.The probability forthis is<0.003, basedonthe linkagemap shown

inFig. 1.Furthermore,thiswouldonly result inincorrect

pre-dictionintheinstance that theoneof thesecrossover events

had occurred between thedisease locus and thetwomarker loci PYGM andD1I S427, for whichnorecombinanthas yet been

found.IntheRAfamily,thethird and still unaffected 24-yr-old child (marked with ?) is predicted to carry thedisease gene

basedonthegenotypedata (Fig. 2).

Ineachofthefollowingkindred outlined in Fig. 1, weshow

RA

PGApcAGP9 1 2 3 3

PYG -CCL15 2 1 2 2

D11S427 3 1 1 2 2

D11S97 3 4 5

2112 2g2 22

32 32 32

31|1 3j5 31

LU

l

I

O

PGA-pcAGP9 1 3 2 3

PYGM-pMCMfP 3 2 ] 1

INT2 A B C[ A

31

3 2 11 3 13

2 I 311 2 Ii

A C BlEA A IA

AX

PGA-pA4GP 2 1 2 2

PYGM-pAMCMP1 1 3 3 3

D11S97 4 1 2 3

242 111 112 242

13 3113313 113

42 11121 3I 412

KA-1

D11S28/2D11S149 B A B B

PYGM.CCL15 1 12

D11S427 1 2 1II 1

D11297

12

2I

D11S751 B C A B

BB A B B B A B

a_PYGMl512 1 2 21 2

D1B427 3i1 2 1 1 2 2 D211;SS7195B6 D A 2 *1 1 2. 1 42 1 1

21|G12 12 12 2 12 2

DI1427

II

11

l21111I

2

DlIS9273 1132 11 133

A2 2222 A 2 2 D

PGA-CCLPGA12 PYGM-PMCMP1

D11S427

D1S750

D11S97

Di 1S146

Figure 2.Pedigrees illustrating DNA basedpredictive testingfor MEN1.

Filledsymbolsindicate affected

familymembers. Numbers below eachindividualindicate the geno-typesfortheinformative marker systems,listednext to thesymbol of thecarrier parent. The haplotypes for eachindividualaredepicted

alonganillustratedchromosome segment, with theinferred mutant chromosomeindicated by a thin

line,andthewild type chromosome of the carrier parent hatched. The

arrows atkindred KA-2 indicate the

pointofmeiotic crossover in the fifth daughter. KA-1, KA-2, and KA-3 are subsetsof the KA pedigree

that theDNAtesting forMENI is based on at least one marker

whichshows nocrossover with the disease locus, and at least one marker flanking the disease on each side ofthe disease locus. Thus, we excluded crossovers close to MEN 1 in each of

these kindreds.Asillustrated inthe KA-2 kindred,a meiotic crossoverhad occurred in the fifthdaughter betweenDI S288 and PGA. Since the point ofcrossover could be located in this

girl, we could then use six alternative markers beyond this

point, PGA, PYGM, D11S427, D1I S750, D1I S97, and

Dl1 S146, and foundnoevidence forcrossingoverforanyof these markers.

Thesetof sixfamilies outlinedinFig. 2illustratehow DNA testing forMEN1 would bepossibleto useforpredictive

test-ing.In16offsprings (all childrenexcept onein eachfamily)we

know the outcome of 13 based on clinical and biochemical data. In all these cases the DNA test wastotally conclusive basedon a setofatleastoneclosely linked probe,(0=0%) and

twoflanking probesinallinstances(Fig. 2). Hence,thesetof markersdescribed here (Table Iand II)provide auseful and

accuratemeansforpredictive testing ofMEN1.Since theDNA test canbeappliedatany age,providingthefamilystructure

permits linkage analysis, itcan predate the biochemical and clinicalpresentation ofthedisease bytwo tothree decades with

very highaccuracy(>99.5%). Such predictionscanbemade forthethird child (age 24)in theRAkindred who is deducedto

carrytheMEN1predisposition. Similarlytheyoungest

daugh-ter(age 26) in theAXkindredispredicted unaffected,while her

youngerbrother(age 25), still unaffected, isagenecarrier.

Discussion

In this report, wehave presented an extensive family study using five RFLP markers closely linkedto MEN1, with no

meiotic recombinations with the disease locus. These five closely linked probesarethen flankedbyfour markersystems on eachside, all located within7 cMfrom the diseasegene,

basedon combining

genetic linkage

in reference families and

our own linkage data in MENI families (Table II). The 17 probes studiedinthisreportinclude7anchor markers which have been ordered by

multipoint linkage analysis (11).

All probes used here have also beenmapped

physically

within the

1lq 12-q 13 region (12-14).Thegeneorders,asdetermined

by

physical mapping (13, 14)and

genetic

linkagedata(4, 1

1)

are

infullagreementwiththe datapresentedhere.

Asillustrated inFig. 2,theDNAprobeslinkedto MEN1 providea setoflinkagemarkers that will enableaccurateand reliablepremorbid

diagnosis

ofthe disease

predisposition

al-readyatbirth,

providing

the

family

structure

permits linkage

analysis.Therearethree maincausesof incorrect results from DNA-basedtesting of inheriteddiseases: (a) meiotic

crossing

over between the marker and the disease

loci,

(b)

incorrect diagnosisofanindividual;and(c)

mixing

of

samples.

The latter

two can onlybe managed by correctclinical and laboratory procedures, while meiotic crossovers are a serious

problem

whenthe markersare notcloselylinkedtothedisease,orthe number of useful markers is restricted. The set of 17 DNA

probes linkedto MEN 1 providesauseful and reliable toolto

eliminate crossing over as cause of incorrect

prediction

of

MEN1.Thisincludes fiveclosely linkedprobesthatareflanked

by a similar set of markers on each side ofthedisease locus. Hence,we arenotonly able todeterminethe

linkage phase,

usingatleastone veryclose marker system (with no crossover with MEN 1), but also to detect if a meiotic crossover event has

occurredwithinthe chromosomalregionflanking the MEN 1

locus.Until the diseasegene has been cloned, predictive DNA

testingforMEN 1 should be supported by at least one closely

linked marker and one flanking marker on each side of the

diseaselocus, in ordertoavoid incorrect diagnosisdue to

mei-oticcrossingover. Inthis instance,the reliability of the

predic-tionwill exceed 99.5%, which isacommon level for

inaccura-cies duetosample mixingorincorrect diagnosisof patients. In conclusion, the marker systems in Table II that show

recombination fractions of- oo arelocatedoutsidethe MEN 1

regionbased on meiotic crossovers. Therefore, we would like to stressthatpredictive testing forMEN 1 should not be based on

linkagetothese markersalone.Instead, these markers are to be

usedasindicators of crossingovercloseto MEN 1,inorder to

identify matingswherepredictivetesting can not be considered

reliable.

Finally,oneshouldnotexclude the possibilitythatmeiotic crossingover may occurbetween MEN1 and anyof the five markers closesttothediseaselocus. In aclinicaltestsituation, thiswillnotpassundetected if thesetupillustrated in Fig. 2 is used.On the other hand,anysuchcrossoverwill beakey find-ing forfurtherlocating the MEN1genewithin this region, and willfacilitate the final cloning oftheMENI gene.

Acknowledgments

Wewouldlike toacknowledge MarilynWaltersfor establishingcell linesfromtheTasmanian pedigree.

The work was supported by grantsfromAxel and Margaret Ax:son Johnson'sFond,theSwedish MedicalResearchCouncil,theSwedish

CancerSociety,PaulMackayBoltonFoundation, TasmanianCancer

Committee, theAustralian National Health and Medical Research

Council,Torsten andRagnarSoderberg's,and LarsHierta'sand Mag-nusBergwall'sFond.

References

1. Wermer, P. 1963. Endocrine adenomatosisandpeptic ulcer in a large kindred.Am.J.Med.35:205.

2.Brandi,M.L., S. J. Marx, G. D. Aurbach, and L. A. Fitzpatrick. 1987.

Familialmultiple endocrine neoplasiatype 1; A new look at pathophysiology.

Endocr.Rev.8:391-405.

3. Larsson,C.,B.Skogseid,K.Oberg,Y.Nakamura, and M.Nordenskjold. 1988.Multipleendocrineneoplasiatype 1 gene maps to chromosome 11 and is lost ininsulinoma.Nature(Lond.)332:85-87.

4.Nakamura,Y.,C.Larsson,C.Julier, C. Bystrom,B.Skogseid, S. Wells,K.

Oberg,M.Carlson,T.Taggert, P.O'Connel,P.Leppert, J. M.Lalouel,M.

Nor-denskjold,and R.White. 1989.Localization ofthegeneticdefect inmultiple

endocrineneoplasiatype 1 withinasmallregion ofchromosome 11. Am.J.Hum.

Genet. 44:751-755.

5.Friedman,E., K.Sakaguchi, A. E.Bale, A.Falchetti,E.Streeten, M. B.

Zimering,L.S.Weinstein,W.0.McBride,Y.Nakamura, M. L.Brandi,J. A. Norton, G.D.Aurbach,A. M.Spiegel,and S. J. Marx. 1989. N. Engl. J.Med.

321:213-218.

6.Thakker,R.V.,P.Bouloux,C.Wooding,K.Chotai,P.M.Broad, N. K.

Spurr,G.M.Besser,andJ. L. H.O'Riordan. 1989.Association ofparathyroid

tumorsinmultipleendocrineneoplasiatype 1 with lossofallelesonchromosome 11. N.Engl. J.Med.321:218-224.

J. J.Shepherd. 1990. Clonal loss of Int-2 alleles in sporadic4ndfamilial pancre-atic endocrine tumours. Br. J.Cancer.62:253-254.

9. Knudson, A.G. 1978. Retinoblastoma:aprototypic hereditaryneoplasm. Semin.Oncol. 5:57-60.

10.Skogseid, B.,B.Eriksson, G. Lundqvist, L. E. Lorelius, J. Rastad,L.Wide, E.Wilander, G.Akerstrom,and K.Oberg. 1991.Multiple endocrineneoplasia

type 1-A ten yearprospectivescreeningstudy in four kindreds. J.Clin.

Endo-crinol. Metab.73:281-287.

11.Julier, C., Y. Nakamura, M. Lathrop, P. O'Connell, M. Leppert, M. Litt, T.Mohandas, J. M.Lalouel, and R.White. 1990.Detailedmapofthelong arm of chromosome 11. Genomics. 7:335-345.

12.Tokino, T., E.Takahashi,M. Mor, A.Tanigami,T.Glaser, J. W. Park, C. Jones, T.Hori,and Y. Nakamura.1991. Isolation andmapping of62 new RFLP markers on human chromosome11.Am.J.Hum.Genet.48:258-268.

13. Janson, M.,C. Larsson, B. Werelius, C. Jones, T. Glaser, Y. Nakamura, P. Jones,and M.Nordenskjold. 1991. Detailed physical map of human chromo-somalregionI 1q12-13 showshighmeiotic recombination rate around the MENI locus. Proc.Nail. Acad.Sci. USA.88:10609-10613.

14.Larsson, C.,G.Weber,E. Kvanta,K.Lewis,M.Janson, C.Jones, T.

Glaser,G.Evans,andM. Nordenskjold. 1991. Isolation andmapping of polymor-phic cosmid clones used for sublocalization of the MEN I locus. Hum. Genet. In press.

15.Shepherd, J. J. 1985. Latent familial multiple endocrine neoplasia in Tasmania. Med. J. Aust. 142:385-397.

16.Shepherd, J. J. 1991. The natural history of MEN I which may be highly unrecognised rather than highly uncommon. Arch. Surg. In press.

17.Kidd, K. K., A. M. Bowcock, J. Schmidtke, R. K. Track, F. Ricciuti, G.

Hutchings,A.Bale, P. Pearson, and H. Willard.1989. Human gene mapping 10.

Cytogenet.CellGenet. 51:622-947.

18.Williamson, R., A. Bowcock, K. Kidd, P. Pearson, J. Schmidtke, H. S. Chan, M. Chipperfield, D. N. Cooper, J. Hewitt, F. Lewitter, B. Maidak, M.

Quitt,F.Ricciuti,and R. K. Track. 1990. Human gene mapping 10.5.Cytogenet.

Cell Genet. 55:457-778.

19.Rabbitts, P. H., J. Douglas, P. Fisher, E. Nacheva, A. Karpas, D. Ca-tovsky, J. V. Melo, R. Baer, M. A. Stinson, and T. H. Rabbitts. 1988.