experience with direct molecular diagnosis of fragile x · ciated with fragile xhas been the...

3J Med Genet 1992; 29: 368-374

ORIGINAL ARTICLES

Experience with direct molecular diagnosis offragile X

J C Mulley, S Yu, A K Gedeon, A Donnelly, G Turner, D Loesch, C J Chapman,R J M Gardner, R I Richards, G R Sutherland

Department ofCytogenetics andMolecular Genetics,Adelaide Children'sHospital, NorthAdelaide, SouthAustralia 5006,Australia.J C MulleyS YuA K GedeonA DonnellyR I RichardsG R Sutherland

TumbatinDevelopmental Clinic,Prince of WalesChildren's Hospital,High Street,Randwick, NSW 2031,Australia.G Turner

Department ofPsychology, La TrobeUniversity, Bundoora,Victoria 3083,Australia.D Loesch

Department ofPaediatrics, School ofMedicine, Universityof Auckland,Auckland, NewZealand.C J Chapman

Medical Genetics,Department ofPaediatrics, DunedinHospital, Dunedin,New Zealand.R J M GardnerCorrespondence toDr Mulley.

Received 22 October 1991.Revised version accepted21 January 1992.

AbstractThe utility of the pfxa3 probe for directmolecular diagnosis of the fragile X(FRAXA) has been established. Thisprobe detects amplification of an un-

stable DNA element consisting of vari-able length CCG repeats. The size of theamplified fragment is correlated withphenotype and was determined usingPstd digested DNA in family members. In35 families with the fragile X, there was

correspondence in 183 cases between thepresence of an amplified unstable ele-ment and the presence of the fragile Xchromosome independently determinedby cytogenetics, position in the pedigree,or linked DNA markers flanking the fra-gile X. There was also correspondence in124 cases between the presence of thenormal 1 0 kb Psd fragment and absenceof the fragile X chromosome independ-ently determined by linked flankingmarkers. Six additional families con-

sidered to be isolated cases of 'fragile X'had been diagnosed before recognition ofFRAXD. The pfxa3 probe confirmed thecytogenetic diagnosis in three families,the other three being rediagnosed as

non-fragile X. A further two families hadconsistent expression of a different folatesensitive fragile site, FRAXE, close toFRAXA but not associated with fragile Xsyndrome and not detectable with thepfxa3 probe. Subsequent referrals were

received from additional family mem-

bers or from members of new familiesfor whom carrier status had not beenpredetermined by linked markers. Directpfxa3 diagnosis for the 135 females withinthese 222 additional cases was confirmedby dosage analysis with the control probepS8. Independent confirmation of theprimary pfax3 diagnosis was helpful forcorrect diagnosis of females because of asmall proportion who had an unstableelement expressed as a faint smear ofbands and because of the occasional pres-ence of high molecular weight bands fromincomplete Psd digestion. Diagnosis was

definitive for all males using pfxa3 alone inthe presence of only one X chromosome.

Fragile X syndrome is the most commonfamilial type of mental retardation in humans.The syndrome is associated with the folatesensitive fragile site at Xq27.3 (FRAXA).Families with fragile X syndrome had beenpreviously ascertained on the basis of cytoge-netic demonstration of the fragile site in amentally impaired family member.

Genetic counselling for family members hasbeen complicated by three factors. These areincomplete penetrance of cytogenetic expres-sion of the fragile site in a proportion of femalecarriers, the occurrence of transmitting maleswho have the fragile X genotype but do notshow mental impairment, and the presence of acommon fragile site, FRAXD, ar Xq27.21which can be expressed at low frequency underthe same conditions used to detect the fragile X.Transmitting males (defined as unaffected car-rier males who have reproduced or have thepotential to reproduce) have been detectedeither by risk analysis using closely linked poly-morphic markers flanking the fragile X2 or bytheir position in the pedigree. The occurrenceof these asymptomatic male carriers is a uniquephenomenon among X linked genetic disorders.However, the most puzzling phenomenon asso-ciated with fragile X has been the Shermanparadox,34 suggesting a progressive increase inseverity from generation to generation.Molecular characterisation of the fragile X

in three laboratories-7 has made possible adirect test for the fragile X genotype in affec-ted families. The existence of transmittingmales, and the Sherman paradox, are nowexplicable in terms of the increase in size fromgeneration to generation of the heritable un-stable DNA element which is the molecularbasis for this disorder.5689 This unstable ele-ment consists of variable lengths of a repeatingCCG trinucleotide sequence.810 The size of thePstI fragment containing this repeat is 1 0 kbin normal X chromosomes."'0 What initiatesamplification of the 1 0 kb fragment trans-forming it into a fragile X is not known;however, once the process has begun there is adirect correlation between the intellectualcomponent of the fragile X phenotype, cyto-genetic expression of the fragile site, andlength of the unstable element which charac-terises the fragile X genotype.9 Whether thecorrelations between phenotype and length ofthe unstable element, and between cytogenetic

368

on April 7, 2020 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/


expression and length of the unstable element,represent continuous relationships, or step-wise relationships with discontinuity at a PstIfragment size of approximately 1 6 kb, cannotbe resolved without additional data.9 Femaleswith larger amplifications (greater than 1-6 kb)were either normal or mentally affected, imply-ing that additional factors such as non-randomX inactivation may contribute to intellectualphenotype in females. The unstable elementusually increases in size from one generation tothe next when transmitted by females, butwhen transmitted by males remains the samesize or decreases. The larger sized elements aresomatically unstable and can present as multi-ple bands on Southern blots.5689

Materials and methodsPROBESThe pfxa3 probe is a 536 bp DNA fragmentexcised from pUC18 using the restrictionendonuclease PstI. The pfxa3 target sequenceis located immediately distal to the unstableelement which comprises the fragile X.5 Thepfxa3 probe detects a 1-0 kb fragment onSouthern blots of PstI digested DNA fromnon-fragile X chromosomes, while the un-stable fragile X element appears as a fragmentof higher molecular weight. The pS8 probe isan 800 bp fragment excised from pUC19 usingPstI. This sequence was derived from a YACcontaining the VK21 sequence and is used asan internal positive control probe in doublehybridisation with the pfxa3 probe on PstI

2 3 4 5 6 7 8 9

filters. It detects an 800 bp fragment on South-ern blots. An autoradiograph showing thepfxa3 and pS8 fragments is shown in fig 1. Thepfxa3 probe is available from Oncor, Inc, Box870, Gaithersburg, MD 20884, USA.

HYBRIDISATION CONDITIONSHybridisation was carried out at 42°C in5 x SSPE (pH 74), 1% SDS, 50% forma-mide, 10% dextran sulphate, and 100 jtg/mlsalmon sperm DNA. Filters were washed in0 1% SDS and 0.1 x SSC (0-015 mol/l sodiumchloride and 0 0015 mol/l sodium citrate) at70°C for at least 30 minutes. The high strin-gency wash at 70°C is crucial for removal of lanebackground, presumably related to the highGC content of the probe, in order to visualiseclearly the unstable element band or bands.

AC REPEAT MARKERS AT THE FRAGILE XTwo AC repeat markers detectable by thepolymerase chain reaction (PCR) have beendescribed: FRAXACl and FRAXAC2 within10 kb of the fragile X" and DXS548 150 kbfrom the fragile X.7 Of the two within 10 kb ofthe fragile X, FRAXAC2 is the more inform-ative (heterozygosity 71%). FRAXACI is instrong linkage disequilibrium with FRAXAC2and is therefore redundant for diagnosis.Primer sequences and reaction conditions weregiven by Richards et al."1

RISK ANALYSISCarrier risks were computed for unaffectedsubjects in 35 fragile X families using linkedDNA markers as previously described'2 butwith closer markers.213 Risk analysis was ap-plied to families where more than one subjecthad been clearly diagnosed by cytogeneticexpression of the fragile X and this providedconfirmation of pfxa3 diagnosis for unaffectedsubjects. Cytogenetic diagnosis had previouslybeen carried out and thus provided confirma-tion of pfxa3 diagnosis for affected subjects.The map of linked markers used in the riskanalysis was based on that previously estab-lished.213 This differs only very slightly fromthe map subsequently determined from theCEPH (Centre d'Etude du PolymorphismeHumain) families'4 using the microsatellitemarkers FRAXACI and FRAXAC2 torepresent the fragile site locus."

Q-8 - .- a*

0

*1

i...

Figure 1 Autoradiograph of PstI digested DNA probed with pfxa3 and pS8. (.Carrier female, (2) transmitting male, (3) non-carrier female, (4) affected malecarrier female, (6) affected male, (7) non-carrier female, (8) normalnon-transmitting male, (9) carrier female.

BASIS FOR DIAGNOSIS USING THE pfxa3 PROBE

pfxa3 PstI digests are the most instructive since theygive the necessary resolution on Southernanalysis to distinguish male transmitters from

pS8 normal males, to separate male transmittersfrom affected males, and to differentiate nor-mal females from carrier females9 (fig 1). Dia-gnosis is unambiguous in males. Hybridisationwith PstI digested chromosomal DNA gives a1-0 kb band in normal non-transmitting males,

1)(, a band of up to 1-6 kb in male transmitters of(5) the fragile X, and one or more bands of greater

than 1-6 kb in males with fragile X syndrome.

369



g.bmj.com

/J M


g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/

Mulley, Yu, Gedeon, Donnelly, Turner, Loesch, Chapman, Gardner, Richards, Sutherland

Where there is a smear of bands there may bebands within the 1 0 to 1 -6 kb range in additionto bands in excess of 1-6 kb. Male transmittersand affected males do not have a band ofnormal intensity in the 1-0 kb position andusually have no band in the 10kb positionwhen the CCG repeat has been amplified.

PstI digested DNA usually gives only a

1-0 kb band in the normal non-carrier females.Occasionally a doublet at 1 0 kb is detectable innon-carrier females who have two alleles at theextremes of the normal range of allele sizes forthe CCG polymorphism. In all affected carrierfemales and some of the unaffected carrierfemales the pfxa3 probe hybridises to a 1 0 kbfragment from the normal chromosome plus a

fragment or fragments of greater than 1 -6 kbfrom the fragile X chromosome. A faint smearof fragments greater than 1x6 kb in size may bedifficult to detect. Unaffected carrier femalesall have either a 1-0 kb band and a band of upto 1 6 kb or the same pattern as the affectedfemales; thus, the pfxa3 probe cannot predictthe intellectual phenotype of females when thePstI fragment exceeds 1-6 kb. Carrier femaleswith fragment size or sizes exceeding 1-6 kbusually express the fragile site cytogenetically,whereas females with a fragment size of 1 6 kbor less usually have only a single abnormalfragment which is clearly visible and they donot express the fragile site cytogenetically.9Unlike males with a CCG amplification,females always have a clearly visible 1-0 kbfragment from the normal X chromosome.

Results and discussionDETECTION OF FEMALE CARRIERS AND MALE

TRANSMITTERSAn example of diagnosis using the pfxa3 probefor one of the 35 families examined (whichcontained more than one carrier on the basis ofcytogenetic analysis) is shown in fig 2. Thispedigree shows those features routinelyencountered in direct fragile X diagnosis usingthe pfxa3 probe. The unstable element can beseen to increase in size from generation II(1-2 kb) to generation V (3-2 kb) with develop-ment of multiple bands (in IV.1, IV.19, andV.2) associated with the increase in size. Thefragile X was detected using pfxa3 for allaffected boys for whom DNA was available(IV.1, IV.10, V.1). The obligate male trans-mitter (II.4) had a 1 2kb fragment and theprobable male transmitter (99-8% risk) (III.6)had a 1-3 kb fragment. The women shown tohave the fragile X by cytogenetics were con-

firmed as carriers using the pfxa3 probe (IV.2,IV.8, IV. 19, V.2). The obligate female carriersfor whom DNA was available, who did notexpress the fragile X (II.2, III.3, III.8, III.9,III.13, IV.5), were all shown to have abnormalpfxa3 fragments. Two women (IV. 17 andIV.18) ascertained as probable carriers (99%)using linked markers were shown as definitecarriers using pfxa3. The remaining familymembers at low risk using linked markers(III.1, III.5, IV.3, IV.6, IV.9, IV.11, IV.13,IV.14, IV.15, IV.16, IV.20, IV.21) were con-

firmed as non-carriers using the pfxa3 probe.

Kl 8479

1-0 1-7 1-0 1 0 1.0 1 0 1-0, 1 0, 1 0, 1.0, 1-0, 1 .0,1-0 1-4 1-2 2 7 1.0 1-0

(0 9%) (0 9%) (0 2%) (0-2%) (0 2%) 2-2

(99%) (99%) (0 3%) (0 3%)

Figure 2 Fragile X pedigree showing typical behaviour of the unstable element detectable with pfxa3. The pfxa3 genotype is shown for subjectsfrom whom DNA was collected and consists of the normal 1 0 kb allele and the alleles of variable and abnormal size associated with the unstableelement. Corresponding risks determined by linkage analysis are shown in parentheses and were based on pairwise flanking markers from within theset DXS292, DXS297, DXS304, and DXS52. (Marker genotypes are not shown.)

1*0,1*0(02%)

11

III

IV

V

2-0 1-0, 1.0,2 2 2 2 1-0

(3%)

1 *0,1*0

(2%)

1 *0,2-2

3-2 1-0,3-02-52-0

El 0 Fragile X negativeE (0) Obligate carriers

O Fragile X positive* Fragile X positive, fragile X syndrome

370



g.bmj.com

/J M


g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/


Fragile X diagnoses carried out using the pfxa3 probe.

Males Females

Normals Transmitters Affecteds Non-carriers Carriers

Cytogenetically Cytogenetically-ve +ve

(A) Prediagnosed by linked DNA markers (n = 307)53 13 60 71 64 46

(B) Not prediagnosed by linked DNA markers (n = 222)47 7 33 61 47 27

Results for the remaining 34 families wherediagnosis had been predetermined by linkedmarkers were of the same pattern (data notshown). Diagnosis was carried out for a total of307 family members (table A). Detection bypfxa3 of a 1 0 kb PstI fragment from the 195 Xchromosomes of all non-carrier females andnormal males (independently assessed bylinked markers) in these 35 families excludedthe possibility of polymorphism at the PstIsites. Non-cleavage at either PstI site couldhave generated fragments of higher molecularweight indistinguishable from fragments con-taining CCG amplifications. In general, theunstable element responsible for fragile Xgenotype increased in size from generation togeneration when transmitted by females, butnot when transmitted by males (fig 2). Occa-sionally there was a small decrease in the sizeof the unstable fragments when transmittedfrom one generation to the next. In fig 1, forexample, the carrier mother in lane 1 has a1-3 kb fragment which reduces to 1-2 kb in hermale transmitter son, shown in lane 2. In oneunresolved case a normal pfxa3 fragment indouble dose was detected from a female withno abnormal fragment, who had a carrier hap-lotype determined by the AC repeat markersDXS297, DXS548 (150 kb proximal toFRAXA), and FRAXAC2 (within 10kb anddistal to FRAXA). This apparent contradic-tion might arise from a sample error causingmisinterpretation of linkage data or from con-traction of an amplified fragment during trans-mission from mother to daughter.

All subsequent referrals (222 cases fromconfirmed fragile X families), in which carrierstatus had not been predetermined by linkedmarkers, were assessed using the pfxa3 andpS8 probes hybridised to PstI digested DNA(table B). Frequently, subjects from fragile Xfamilies had heard of recent progress in thediagnosis of the fragile X and requested testingwhen their pregnancy had been confirmed (inthe case of females) and when pregnancy fortheir daughters had been confirmed (in thecase of potential male transmitters). The 135female genotype results determined by directdiagnosis alone agreed with dosage analysis. Inone case of male transmission a significantdecrease in the size of the unstable element wasobserved. A mildly affected male with a smearof PstI bands (1 -8 kb to 2-1 kb) had reproducedand passed on a fragment of 1-3 kb to onedaughter and 1-2 kb to his other daughter.

DOSAGE ANALYSISResults of pfxa3 hybridisation may be uncer-tain in females if the PstI digest gives a 'smear'of weakly hybridising bands. This may beresolved using EcoRI digests (where the nor-mal band is 5 kb). Using EcoRI, the dif-ferences in the lengths of the multiple bandsare not so great and the 'cluster' of amplifiedbands (not shown) is easier to detect.However, homozygosity or heterozygosity

for the normal 1-0 kb fragment assessed by theabsence or presence of an abnormal fragmentassociated with fragile X genotype can beeasily confirmed by dosage analysis on the PstIdigests. This involves comparison of signalintensity of pfxa3 compared with pS8 usingthe same PstI blot used to make the primarypfxa3 diagnosis. Such a comparison is advis-able for females in the absence of confirmatorylinkage studies, since abnormal bands in thehigher molecular weight range often occur asweakly hybridising multiple bands and can beeasily obscured by small amounts of back-ground hybridisation. For example, in fig 1 theintensity of the 1-0 kb fragment of pfxa3 is lessthan that ofpS8 in the carrier females in tracks1, 5, and 9 owing to the presence of only asingle dose ofDNA from only one X chromo-some. In the normal females in tracks 3 and 7the intensity of the 1x0 kb fragment of pfxa3 isclearly greater than that of pS8 owing to thepresence of two doses of DNA from two nor-mal X chromosomes.While relative intensity of the pfxa3 frag-

ment compared with the pS8 fragment mayvary from one hybridisation to the next,depending on variations in concentration andlabelling efficiency of each of the probes, theratio of the 1-0 kb pfxa3 fragment to the 0-8 kbpS8 fragment within lanes for any onehybridisation remains constant such that it ispossible to verify the primary diagnosis ofcarrier status. This dosage analysis isindependent of the amount of DNA in dif-ferent lanes, the relationship between the 1-0pfxa3 fragment and pS8 is constant for com-parison between normal males and normalfemales (fig 1, tracks 3, 7, and 8), and theprocedure was found to be reliable if the inten-sity of the 10kb pfxa3 and 0-8kb pS8hybridising fragments were approximatelyequivalent. Assessment of dosage agreed withobserved presence or absence of the pfxa3fragment for the 135 females tested (table B).Dosage analysis cannot be applied to the am-plified pfxa3 fragment because of the propen-

371



g.bmj.com

/J M


g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/


A

11

1-0 1.0 1-32-02 22-7

B

111 3

III

Figure 3 (A) Family withan isolated male with fragileX syndrome. (B) Familywith an isolated affectedfemale. Genotypes for pfxa3from PstI digested DNAare shown on the pedigrees.

EJ 0 Fragile X negative* * Fragile X positive, fragile X syndromeIJ Fragile X negative, non-fragile

I X mental impairment

sity to form multiple bands arising from soma-

tic instability.

FAMILIES WITH ISOLATED CASES OF FRAGILE XSYNDROMEIn addition to the familial cases describedabove, there were families with isolated cases

which were confirmed as fragile X by detectionof the unstable amplified fragment usingpfxa3. Two of these families are shown in fig 3.Fig 3A shows a family with an isolated male(II.3) affected with fragile X syndrome. Asmear of bands confirms the diagnosis for II.3.Observation of a 1-3 kb band for one of hisbrothers (II.4) was the only means of showingthat this brother was a male transmitter. Theobservation of a 1-2 kb band for their mother(I.2) was also the only means of showing thatshe was a carrier and that II.3 was not a new

mutation. A normal 1-0 kb band for I.2 con-

firms that his mental impairment is unrelatedto the fragile X, in agreement with existingcytogenetic analysis which did not detect thefragile X, and the clinical observation of a non-fragile X phenotype. Neither I.1 nor II.2carries the fragile X mutation.An isolated female with fragile X syndrome

is shown in fig 3B. Diagnosis of IV.6 as a

definite carrier could not be made beforeanalysis by pfxa3. Even with highly polymor-phic AC repeat markers at the fragile X, such a

diagnosis is unlikely to be made by linkageunless genotypes for I.1, I.2, II.1, II.2, II.4,and II.5 could all be inferred.The possibility of new mutation as a cause

for isolated cases of fragile X syndrome hasbeen excluded for all cases examined so far.9Recognition that 'isolated' cases of fragile Xare all familial has had considerable impact on

genetic counselling given to their relatives.Clarification of carrier status for relatives ofsuch isolated cases has only been possible sincethe availability of direct molecular diagnosis.

EXCLUSION OF FRAGILE X SYNDROME

At the time when analysis of the initial 35familial cases had been completed, six isolatedcases had at least 2% of cells positive bycytogenetics and had been diagnosed as fragileX syndrome. Three of these six cases were

reclassified as non-fragile X syndrome on thebasis of normal pfxa3 fragments from PstIdigested DNA. During the search for new

mutations,9 further isolated cases which hadbeen 'prediagnosed' by cytogenetics were

selectively sought from collaborators in orderto expand the number available for study. Ofthese, five isolated cases from a total of 11 were

reclassified as non-fragile X syndrome on thebasis of normal pfxa3 fragments from PstIdigested DNA. The proportion of such cases

within collections from various laboratorieswill vary depending upon the quality of thecytogenetic analysis and on how recently thisanalysis was carried out. The existence of a

common fragile site, FRAXD, at Xq27.2 isnow well established' and is one explanationfor misdiagnosis in cases completed beforerecognition of FRAXD. Isolated cases of fra-gile X should be retested using a probe whichdetects the unstable element. Thus, the pfxa3probe can be used to establish or exclude thediagnosis of fragile X syndrome for isolatedcases in addition to determination of carrierstatus of males and females in families wherethe primary diagnosis has been established byunequivocal cytogenetics and confirmed bymolecular analysis.The problem of false positive cytogenetic

diagnosis based on low rates of 'expression' isnot restricted to isolated cases. They have beenmade within fragile X families and identifiedby linked flanking markers.'5 Where false pos-itive cytogenetic diagnoses could only bedetected within families in the past by usinglinked markers, they are now easily ascertainedusing molecular diagnosis.

A NEW RARE FOLATE SENSITIVE FRAGILE SITE,

FRAXEAt the time when analysis of the first 35familial cases had established the reliability ofthe pfxa3 probe for diagnosis, the pfxa3 probefailed to detect an unstable element in the'fragile X' family reported by Romain andChapman.'6 This family was unusual in thatthe 'fragile X' carriers did not have fragile Xsyndrome despite consistent expression of a

folate sensitive fragile site at Xq27.3. Thelikely explanation is the presence in this familyof a distinct fragile site, FRAXE, which has no

IV

V

1*0

372



g.bmj.com

/J M


g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/


A 1 2 3 4 5 B 6 7 8 9

- 10

to be distinguished from abnormal pfxa3 frag-ments in order to avoid false positive molecu-lar diagnosis.

6 5

_ _ _

1.0 4' 0 ~08*0-8 .4-A

Figure 4 (A) A contaminating plasmid band of 6 5 kb is shown for a mal3. Track 5 shows incomplete digestion ofDNA from a carrier female. (B)and 9 show the pattern of incomplete digestion observed where digestion hasalmost to completion. (1) Carrier female, (2) carrier female, (3) normal nnormal male, (5) carrier female, (6) normal male, (7) carrier female, (8)male, (9) carrier female. In (A) relative intensity of pfxa3 and pS8 bandsapproximately equivalent and can be used for dosage analysis to confirm catoffemales. In (B) the pS8 hybridisation is inadequate for reliable dosage a

phenotypic effect but which mapsproximity to FRAXA. The family oand Chapman'6 was described beforeability of pfxa3 analysis and has featilar to two other families that hreported which have high fragile X ewithout mental retardation.6' 17 Thetested so far6' 16 did not show thepfxa3 fragment. In another recentment, FRAXE was shown to mapFRAXA by in situ hybridisation usixnear FRAXA.'8 This confirms the exat least one other rare folate sensitisite locus, FRAXE, located veryFRAXA. The FRAXE site is not i

fragile X syndrome (associated with iand is not the common fragile site (.

which is proximal to FRAXA.

ARTEFACTSOne pedigree showed a 6-5 kb frajthree generations which mimicked ninheritance. This fragment was unrthe pfxa3 detectable fragment, as inddosage of the 1 0 kb pfxa3 allele, st'transmission' from generation to geand presence of the 1-0 kb fragment inwho also had the 6-5 kb fragmentconfirmed as plasmid contaminationing earlier autoradiographs used tclinked markers from the same DN.and by reprobing the pfxa3 filter wit}sequence. Fig 4A shows the contaplasmid band and the pattern aris:what we interpret as incomplete digecarrier female. Fig 4B shows clearerof incomplete digestion. These artef;

5- 3 PRENATAL DIAGNOSISThe first prenatal diagnosis using the pfxa3

42 probe has been described fully elsewhere.'93.7 Fragile X genotype was detected in a male

fetus and confirmed by cytogenetics and theinformative flanking markers DXS296 and

2 5 DXS297. No evidence for somatic variationwas found in fetal tissues suggesting that thesize of the unstable element should be useful forpredicting the phenotype prenatally as well aspostnatally. Prediction of phenotype at prenataldiagnosis should be made with extreme caution

* pfxa3 when fragment sizes are close to 1-6 kb sincethis basis for separation of male transmittersfrom affected males has as yet been determined

pS8 from only crude measurement of PstI fragmentpS8 size on Southern blots.9 Prenatal diagnosis hasalso been reported using the probe Ox1.9.20

If prenatal diagnosis is to be carried outmore rapidly than is possible by Southern

'ein track analysis, it needs to be based on PCR tech-Tracks 8 nology using highly polymorphic markers at or

nale, (4) very close to the fragile X. When by Southernnormal analysis the signal associated with abnormal

stis pfxa3 fragments in some carrier females isrrner status weak (fig 1), extended exposure may be neces-

sary to exclude the possibility of a false nega-in close tive result. This is not a problem if the fetus is

f Romain male where diagnosis can be carried out just as

the avail- readily by observed absence of a normal 1 0 kb:ures sim- PstI fragment as by observation of abnormalave been size of the pfxa3 fragment, when the pS8 bandxpression is clearly present as a positive control (fig 1).families PCR based diagnosis by linked AC repeat

amplified markers would provide rapid exclusion; how-ascertain- ever, a positive PCR diagnosis by linkagedistal to would need the additional pfxa3 result. Carrierng probes males would be predicted to be mentally unaf-:istence of fected transmitters if the PstI fragment sizeve fragile was 1-6 kb or less and mentally affected if itclose to exceeded 1-6 kb. Carrier females would be

related to predicted to be mentally unaffected if the PstIFRAXA) fragment size was 1 6 kb or less. However,FRAXD) unlike males, there are at present insufficient

data to enable phenotypic prediction for afemale fetus who has a specific PstI fragmentsize which is greater than 1 6 kb. Having elim-inated the possibility of mental impairment for

gment in carrier females with PstI fragment sizes of 1 6 kbnendelian or less, the remaining carrier females with PstIrelated to fragment sizes exceeding 1 6 kb must as a grouplicated by be at greater risk of mental impairment than theability in risks given for the unsubdivided group.4 These-neration, risks were 32% when the fragile X was passedone man through an unimpaired mother and 55% whenIt was passed through an impaired mother.

by check-) analyseA source Conclusionsi plasmid The pfxa3 probe is a reliable diagnostic toolminating for detecting fragile X genotype and predictinging from fragile X phenotype when applied to PstIstion in a digested DNA. Used in conjunction with thepatterns pS8 probe, results may be checked for consist-

acts need ency by dosage analysis. Alternatively, used in

373



g.bmj.com

/J M


g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/


conjunction with closely linked and highlypolymorphic AC repeat markers, results maybe confirmed by family analysis without theneed for laborious determination of RFLPmarkers by Southern analysis. There is noneed for computerised risk analysis whenusing markers virtually in absolute linkagewith FRAXA. These confirmatory proceduresguard against the possibility of false positivemolecular diagnosis caused by plasmid con-tamination or incomplete digestion. They alsoguard against false negative molecular dia-gnosis in those females who have not beencytogenetically tested and who have a lowintensity smear of abnormal pfxa3 bands whichcould be easily masked by low level back-ground. Such females are usually cytogen-etically positive and therefore are unlikely tobe misdiagnosed when parallel cytogeneticexamination has been carried out. Carrierfemales who are cytogenetically negative aredetectable without difficulty using the pfxa3probe because they usually have only a singlestrongly hybridising fragment of 1-6 kb or lessapart from the normal 1-0 kb fragment. Confir-matory family studies involving the AC repeatmarkers FRAXAC2 or DXS548 could be con-sidered for inclusion in routine diagnostic pro-tocols, not only to confirm the pfxa3 diagnosisand to carry out rapid PCR based prenataldiagnosis of genotype, but also to check forsample misidentification which can cause bothfalse positive and false negative results.The advent of molecular diagnosis raises the

question of the role of cytogenetic analysis.Cytogenetic and molecular procedures will con-tinue to be carried out in parallel until labora-tories gain experience in molecular diagnosis.The initial ascertainment of fragile X familieswith amplification of the CCG repeat is likely tocontinue to be made by the cytogenetics labora-tory from referrals for delayed development, ormental retardation, resulting from many causes.Molecular and cytogenetic diagnosis of fragileX syndrome might be complicated by a smallnumber of cases which do not have an unstableelement. These could result from mutationwithin the associated gene other than at theCCG repeat, such as a point mutation, micro-deletion, or microduplication. One deletion hasbeen detected and the extent of this deletion isbeing characterised in an affected boy where thepfxa3 probe failed to hybridise to PstI or EcoRIdigested DNA (unpublished data).PCR amplification of the polymorphic CCG

repeat from normal chromosomes and thosewith small CCG amplifications now enablesdirect sizing of alleles at the fragile X locus.8This analysis has established the allele distri-butions of this polymorphism in normal malesand in male transmitters and confirmed anupper limit for amplification of 600 bp intransmitting males. Accurate delineation ofthese two groups by direct PCR genotypingwould be preferable to separation based onestimation of PstI fragment size by Southernanalysis. The PCR results, however, must beinterpreted with caution given the extent ofsomatic instability in affected males, which maygive rise to one or more additional bands in the

normal to transmitter range. This inherent dif-ficulty precludes PCR analysis of CCG amplifi-cation from prenatal diagnosis. Conversely,direct PCR genotyping will exclude carrierstate in females with two normal alleles as longas each corresponds to the unaffected allele of aparent or sib. Direct PCR genotyping has notbeen demonstrated for the large amplificationsin excess of 600 bp. It is most likely that,together with PCR based genotyping of flank-ing AC repeats,1' the PCR analysis of the fragileX CCG repeat will supplement, rather thanreplace, direct detection by Southern blots.

Note added in proofExperience with direct molecular diagnosis offragile X has also been reported using theprobe StB12.3 by Rousseau et al (N Engl JMed 1991;325:1673-81).

This work was supported by grants from theNational Health and Medical Research Coun-cil of Australia and the Adelaide Children'sHospital Research Trust.

1 Sutherland GR, Baker E. The common fragile site in bandq27 of the human X chromosome is not coincident withthe fragile X. Clin Genet 1990;37:167-72.

2 Suthers GK, Mulley JC, Voelckel MA, et al. Geneticmapping of new DNA probes at Xq27 defines a strategyfor DNA studies in the fragile X syndrome. Am J HumGenet 199 1;48:460-7.

3 Sherman SL, Morton NE, Jacobs PA, Turner G. Themarker (X) syndrome: a cytogenetic and genetic analysis.Am J Hum Genet 1984;48:21-37.

4 Sherman SL, Jacobs PA, Morton NE, et al. Further segre-gation analysis of the fragile X syndrome with specialreference to transmitting males. Hum Genet 1985;69:289-99.

5 Yu S, Pritchard M, Kremer E, et al. Fragile X genotypecharacterised by an unstable region of DNA. Science1991;252:1 179-81.

6 Oberle I, Rousseau F, Heitz D, et al. Instability of a 550-base pair DNA segment and abnormal methylation infragile X syndrome. Science 1991;252:1097-102.

7 Verkerk AJMH, Pieretti M, Sutcliffe JS, et al. Identifica-tion of a gene (FMR-1) containing a CGG repeat coincid-ent with a breakpoint cluster region exhibiting lengthvariation in fragile X syndrome. Cell 1991;65:905-14.

8 Fu Y-H, Kuhl DPA, Pizzuti A, et al. Variation of the CGGrepeat at the fragile X site results in genetic instability:resolution of the Sherman paradox. Cell 1991;67:1047-58.

9 Yu S, Mulley J, Loesch D, et al. Fragile X syndrome:unique genetics of the heritable unstable element. Am JfHum Genet 1992;50.

10 Kremer EJ, Pritchard M, Lynch M, et al. Mapping of DNAinstability at the fragile X to a trinucleotide repeat se-quence p(CCG)n. Science 1991;252:1711-14.

11 Richards RI, Holman K, Kozman H, et al. Fragile Xsyndrome: genetic localisation by linkage mapping of twomicrosatellite repeats FRAXACI and FRAXAC2 whichimmediately flank the fragile site. Jf Med Genet1991;28:818-23.

12 Mulley JC, Gedeon AK, Thom KA, Bates LJ, SutherlandGR. Linkage and genetic counselling for the fragile Xusing DNA probes 52A, F9, DX13 and St14. Am J MedGenet 1987;27:435-48.

13 Richards RI, Shen Y, Holman K, et al. Fragile X syndrome:diagnosis using highly polymorphic microsatellitemarkers. Am J Hum Genet 1991;48:1051-7.

14 Dausset J, Cann H, Cohen D, Lathrop GM, Lalouel JM,White R. Centre d'ttude du Polymorphisme Humain(CEPH): collaborative genetic mapping of the humangenome. Genomics 1990;6:575-7.

15 Mulley JC, Turner G, Bain S, Sutherland GR. Linkagebetween the fragile X and F9, DXS52 (St14), DXS98(4D-8) and DXS105 (cX55.7). Am J Med Genet1988;30:567-80.

16 Romain DR, Chapman CJ. Fragile site Xq27.3 in a familywithout mental retardation. Clin Genet (in press).

17 Voelckel MA, Philip N, Piquet C, et al. Study of a familywith a fragile site of the X chromosome at Xq27-28without mental retardation. Hum Genet 1989;81:353-7.

18 Sutherland GR, Baker E. Characterization of a new rarefragile site easily confused with the fragile X. Hum MotGenet (in press).

19 Sutherland GR, Gedeon A, Kornman L, et al. Prenataldiagnosis of fragile X syndrome by direct detection of thecharacteristic unstable DNA element. N Engl J Med1991;325:1720-2.

20 Hirst M, Knight S, Davies K, et al. Prenatal diagnosis offragile X syndrome. Lancet 1991;338:956-7.

374



g.bmj.com

/J M


g.29.6.368 on 1 June 1992. Dow

nloaded from

http://jmg.bmj.com/

experience with direct molecular diagnosis of fragile x · ciated with fragile xhas been the...

Documents