doi:10.1136/jmg.38.11.777 2001;38;777-783 J. Med. Genet. Jean-Claude Chomel, Kris De Boeck, Harry Cuppens and João Lavinha Philippe Birembaut, Dominique Hubert, Thierry Bienvenu, Michèle Adoun, Vieira, Claude Féréc, Mireille Claustres, Marie des Georges, Christine Clavel, Cécile Cazeneuve, Michel Goossens, Martine Blayau, Claudine Verlingue, Isabel Doudounakis, Thilo Dörk, Burkhard Tümmler, Emmanuelle Girodon-Boulandet, Rodrigo Cabanas, Xavier Estivill, Maria Tzetis, Emmanuel Kanavakis, Stavros Dapena, Silvia Gartner, Carlos Vásquez, Javier Pérez-Frías, Casilda Olveira, Penque, Paulo Nogueira, Celeste Barreto, Beatriz Lopes, Teresa Casals, Javier Margarida D Amaral, Paula Pacheco, Sebastian Beck, Carlos M Farinha, Deborah homozygotes: a large European study splicing mutation have milder disease than F508del Cystic fibrosis patients with the 3272-26A>G http://jmg.bmjjournals.com/cgi/content/full/38/11/777 Updated information and services can be found at: These include: Rapid responses http://jmg.bmjjournals.com/cgi/eletter-submit/38/11/777 You can respond to this article at: service Email alerting top right corner of the article Receive free email alerts when new articles cite this article - sign up in the box at the Topic collections (189 articles) Cystic fibrosis (3321 articles) Genetics Articles on similar topics can be found in the following collections Notes http://www.bmjjournals.com/cgi/reprintform To order reprints of this article go to: http://www.bmjjournals.com/subscriptions/ go to: Journal of Medical Genetics To subscribe to on 16 May 2005 jmg.bmjjournals.com Downloaded from
Transcript
doi:10.1136/jmg.38.11.777 2001;38;777-783 J. Med. Genet.
Jean-Claude Chomel, Kris De Boeck, Harry Cuppens and João Lavinha Philippe Birembaut, Dominique Hubert, Thierry Bienvenu, Michèle Adoun, Vieira, Claude Féréc, Mireille Claustres, Marie des Georges, Christine Clavel,Cécile Cazeneuve, Michel Goossens, Martine Blayau, Claudine Verlingue, Isabel Doudounakis, Thilo Dörk, Burkhard Tümmler, Emmanuelle Girodon-Boulandet,Rodrigo Cabanas, Xavier Estivill, Maria Tzetis, Emmanuel Kanavakis, Stavros Dapena, Silvia Gartner, Carlos Vásquez, Javier Pérez-Frías, Casilda Olveira,Penque, Paulo Nogueira, Celeste Barreto, Beatriz Lopes, Teresa Casals, Javier Margarida D Amaral, Paula Pacheco, Sebastian Beck, Carlos M Farinha, Deborah
homozygotes: a large European studysplicing mutation have milder disease than F508del Cystic fibrosis patients with the 3272-26A>G
http://jmg.bmjjournals.com/cgi/content/full/38/11/777Updated information and services can be found at:
Anderson-Fabry disease: clinical manifestationsand impact of disease in a cohort of 60 obligatecarrier females
K D MacDermot, A Holmes, A H Miners
EDITOR—Anderson-Fabry disease (AFD) is asphingolipid storage disorder resulting fromthe deficiency of the lysosomal enzymeá-galactosidase. Unlike most other lysosomaldiseases, the inheritance is X linked. Diseasemanifestations in female heterozygotes havebeen reported, but are considered to be rareand usually mild.1 Asymptomatic cornealdystrophy (cornea verticillata and posteriorlenticular cataract) is present in about 70% andis useful for heterozygote detection. About30% of women have minimal angiokeratomasand <10% have infrequent attacks of neuro-pathic pains.1 However, female heterozygoteswith severe and early cerebrovascular disease,strokes, and renal failure have been docu-mented, but these serious manifestations wereestimated to occur in only 1%.1 Because this isan X linked disorder, these severe manifesta-tions in females were explained by skewed Xinactivation.2
Case reports are open to selection bias andthere are no reported data on the frequencyand severity of AFD manifestations in a largecohort of carriers. Women are not usuallyevaluated in the clinic unless they present withserious AFD complications, but they do attendfor genetic counselling when their son or malerelative is diagnosed with the disease.
During regular reviews of AFD families inthe genetic clinic, we have observed multipleand more frequent AFD manifestations infemale carriers than expected. Owing to thelack of clinical data and present health carepolicy, we were unable to perform clinicalinvestigations. Therefore, a large cohort studyusing an AFD specific questionnaire wasperformed to establish the mortality and thefrequency of all AFD manifestations in obligatecarrier females.
MethodsAll known (n=80) obligate carrier females, whowere over the age of 18 and had demographicdata recorded in the AFD register, were askedto participate in this study. Sixty (75%) agreedand provided data; non-responders were re-minded on one occasion. Some participantsdid not complete all measures, so the exact
number is reported for each result. Local ethi-cal committee approval was obtained at thestart of the study.
ASCERTAINMENT
We have studied women who are obligatecarriers for AFD by family history. Thisapproach eliminates selection bias, as completeclinical, biochemical, and genetic mutationdata were available mainly for young women ofchildbearing age for the purpose of prenataldiagnosis. Obligate carrier females were ascer-tained from the UK AFD register, which con-tains data on 67 families and 80 currently livingaVected males. This genetic and clinical regis-ter has been maintained for 15 years and con-tains demographic, clinical, biochemical, andDNA data on families from the UK and Eire.Data were obtained from continuous assess-ment in the genetic clinic, which are sharedwith general practitioners and supplementedby results from individual patients’ hospitalspecialists. á-galactosidase gene mutationswere identified in all families (except one),were diVerent in each family, and includednonsense, missense, stop, and exon skippingmutations.3
MORTALITY DATA
Obligate carriers included were those who haveat least one first degree relative known to us,and who were able to give information on thediagnosis and cause of death or provide a deathcertificate.
AFD SPECIFIC QUESTIONNAIRE
Carriers for AFD were assessed in the geneticclinic, following the diagnosis of AFD in theirfamily, or following incidental findings ofcornea verticillata or angiokeratoma in afemale without a previous family history of thedisease. The AFD specific questionnaire wasthus created, based on in depth interviews withpatients and medical practitioners. (The text ofthis questionnaire is available electronically onrequest.) In the questionnaire, the obligate car-rier status was first confirmed by a statementthat the woman had/has an aVected father orbrother or has an aVected son. The question-naire contained questions on general healthand the presence or absence of manifestations
J Med Genet 2001;38:769–807 769
J Med Genet2001;38:769–775
Department ofMedicine,Addenbrooke’sHospital, Hills Road,Cambridge CB2 2QQ,UKK D MacDermotA Holmes
Primary Care andPopulation Sciences,University CollegeMedical School, RoyalFree Campus,Rowland Hill Street,London NW3 2QG, UKA H Miners
of the disease as is seen in aVected males. Thecharacter and nature of neuropathic pains wereclearly explained and direct responses re-quested as to the onset of pain, medication, andits eVect on everyday life. The McGill PainQuestionnaire (short form)4 was furtheradapted for carriers by selecting the following:the intensity of the pain on average during anattack on a scale of 0 = no pain to 10 = pain asbad as you can imagine, the presence andintensity of background pain on average, itslocation, and eVect on attendance at school,work, social activities, and participation atsports (scale 0-10). Further, women wereasked direct questions: have you ever suVered astroke and have you ever experienced aweakness (in your arm, leg, or face) or doublevision lasting only minutes/hours? We have fol-lowed all reported TIA and CVA by a phonecall to these women and to their general practi-tioners to obtain a detailed history. Informationon renal function was obtained by askingwhether any blood and urine test was done inthe last year, whether it was normal, is thewoman under assessment by a renal unit, is onrenal dialysis, or had renal transplant. For heartfunction, a direct question was asked aboutchest pain/palpitations and results of tests per-formed in the last year. Next, women wereasked to respond yes/no to specific questionson the presence of manifestations of AFD,which included “Fabry rash” (angiokeratoma),sweating, tinnitus, exhaustion, gastrointestinalsymptoms, and others. X linked inheritanceand the understanding of AFD transmissionwas assessed simply by asking whether anaVected father can pass on the geneticmutation to his son or to his daughter and,similarly, what were the chances of transmis-sion of the disease by a carrier mother.
Lastly, the ratings of medical care and thisquestionnaire were assessed, followed by aseries of questions about the respondents’reproductive choices. The eVect of the diseaseon emotional health was measured on a 10item table from the McGill questionnaire(short form).4
STATISTICAL METHODS
The cumulative probability of death for AFDfemale carriers was calculated using theKaplan-Meier survival method and currentdata on the UK female general population wasprovided by the UK Government ActuariesDepartment using 1997-1999 life tables.
ResultsMORTALITY
The survival function in fig 1 shows the mediancumulative survival to be 70 years, whichrepresents an approximate reduction of 15years from the general population. A gradualdecline of the AFD female carrier curve is evi-dent from the age of 35. Six of these 32 womenhad renal failure, of whom half have had a renaltransplant and two have died of lymphoma fol-lowing immunosuppression for renal trans-plantation. Results of renal biopsy were notavailable, as all these women presented withend stage renal failure (ESRF). A total of
15.6% (5/32) had mitral valve replacement andcardiomyopathy, 28% (9/32) died of cerebro-vascular accidents (CVAs), often multiple,aged between 56 and 83 years, and four died ofcancer.
PREVALENCE
The prevalence of AFD carriers is 1 in 339 000females in the UK. The AFD register containsdata on 67 families with AFD of whom 47 hada family history of the disease and 20 aVectedmales did not. In half of these, carrier status intheir mother was confirmed by clinicalexamination, leucocyte, and hair rootá-galactosidase estimations, and subsequentlyby mutation analysis. In the remaining males,the origin of the new mutation is at presentunknown.
The register contains data on 89 obligate andmanifesting carriers and 40 female first degreerelatives, whose carrier status is at presentunknown. The cohort examined in this studyconsisted of 60 obligate carrier females, meanage 44.9 years (95% CI 41.01, 48.79). Pastmedical history included the following: bron-chitis in 31.7%, hypertension in 31.7%, andheart attack, angina, or heart failure in 16.7%;a further 3.3% were treated for breast cancerand 5% were given a diagnosis of rheumaticfever in the past. None had diabetes.
SERIOUS AND DEBILITATING AFD
MANIFESTATIONS
Multiple serious or debilitating complicationsof the disease were documented in aVectedmales, but were not expected in carrier women.In 30% (20 women), multiple and seriousmanifestations were present (table 1). Theseare illustrated by the following case histories. A37 year old carrier presented with ESRF andhas attacks of diplopia and severe peripheralneuropathic pains. Unusually severe AFDsymptoms, even for a hemizygous male, were
Figure 1 Cumulative probability of death for AFD femalecarriers calculated using the Kaplan-Meier survivalmethod. The survival function shows the mediancumulative survival to be 70 years, which represents anapproximate reduction of 15 years from the generalpopulation. A gradual decline of the AFD female carriercurve is evident from the age of 35.
100
80
60
40
20
0
Age (years)
UK femalegeneral population
Survival time: Kaplan–Meier curveMedian: 70 years (25th–75th centile, 57–78)
observed in a young girl aged 17. She was con-fined to bed for three months during her lastyear at school with severe incapacitatingperipheral neuropathic and muscle pains andextreme exhaustion. Following home tuition,she was able to complete a university courseand obtain a job. At her present age of 26, herpain control has been obtained with high dosesof Gabapentine, Amitryptiline, and Solpa-deine. However, she is unable to travel to workor carry out her desk job because of extremefatigue, myalgia, and acute attacks of pain.Another 36 year old carrier appears to suVerfrom a personality disorder and has severeattacks of pain requiring treatment withmorphine. In childhood, the nature of her painswere not recognised and she was under psychi-atric care for many years, until her fatherdeveloped renal failure and the diagnosis wasmade.
NEUROPATHIC PAIN
Seventy percent (n=60) of carriers haveexperienced some form of neuropathic pain(n=42) and 30% have not (n=18). Of the totalnumber of AFD carriers who reported experi-encing pain, 42 (80%) said that the pain hadnever stopped since onset. However, in 10 car-riers (20%) the pain did stop, at a mean age of28.8 (95% CI 17.3, 40.3). Of the 42 carrierswith pain, 41 women had pain attacks duringpyrexial illness or in hot weather, of whom 19also had continuous background pain inbetween attacks. Only one woman experiencedcontinuous pain alone. The mean pain scorefor pain intensity during an attack was 7.4(scale 0-10) and for background pain was 6.8.
A total of 17.1% were taking anticonvulsantsand/or opiates. Over half of the remainder tooknon-steroidal anti-inflammatory drugs regu-larly. The commonest pain location in womencarriers for AFD were the hands and feet,followed by the large joints of the upper andlower limbs.
A total of 27.5% (n=29) stated that the painaVected their attendance at school (mean scoreof 8, McGill questionnaire), in 55.5% (n=28) itaVected participation in sports (mean score8.6), it aVected family life in 27.5% (meanscore 6.6), and social activities in 46.6% (meanscore 6.5). These results suggest that neuro-pathic pain is a significant manifestation ofAFD in a high proportion of carrier females.
Table 2 shows the onset and table 3 showsthe frequency of AFD disease specific manifes-tations in carriers >18 years. The response tothe questionnaire was complete in this cohort;
however, data on renal and cardiac function areapproximate for frequency in carriers as only athird of the women were investigated.
TIA AND CVA
Twenty-three women stated that they hadexperienced at least one TIA. Table 4 showsthe signs and symptoms in 17 carriers in whomTIA was considered likely; a further threewomen were excluded and three remainedunknown. Four women with CVA have alsorecorded a previous history of TIAs.
RENAL FUNCTION
Renal function was assessed in only 31.5% ofcarriers in the last year. Two women are receiv-ing renal dialysis at a mean age 36, of whomone has presented with ESRF. None of therespondents has had renal transplant.
CARDIAC FUNCTION
Cardiac function was assessed in only 35.6% inthe last year, but 52.5% have reported chestpains and/or palpitations and ankle swelling;16.9% of women had abnormal test results(table 3).
GASTROINTESTINAL (GI) SYMPTOMS
GI symptoms were commonly recorded; overhalf of the respondents complained of bloated-ness, indigestion, and abdominal cramps. Diar-rhoea and constipation were present in over40% and vomiting in 14%. No additional diag-nosis of bowel disease was reported by thesewomen. The symptoms are suggestive of being
Table 1 Frequency of serious and debilitatingmanifestations in AFD carriers (n=60)
The cohort examined in this study consisted of 60 obligate car-rier females, mean age 44.9 years (95% CI 41.01, 48.79). Dataon disease manifestations obtained from the questionnaire werecomplete; however, as only one-third of these women wereinvestigated, the frequencies of cardiac and renal abnormalitiesshould be regarded as approximate.
the manifestation of AFD, but the results of GIinvestigations are required to confirm thesefindings.
The frequency of angiokeratoma is mostlikely an underestimate, as on clinical examina-tion several carriers in this cohort had sparselesions, in the form of diVuse tortuous capillar-ies on the upper chest or patches of angio-keratoma in dermatome distribution on theupper or lower limbs.
In this cohort, hypohidrosis was reported by32.8% as an inability to sweat on exertion, dur-ing sport sessions, and when pyrexial. The triadof hypohidrosis, overheating, and mild periph-eral pains was used by the women as diagnosticof their carrier status. Tinnitus was reported by25%, in the form of unpleasant attacks, usuallylasting seconds or minutes, often associatedwith dizziness.
Disabling joint and muscle pains, diagnosedas fibromyalgia were present in five carriers(7%) aged between 45 and 55 years. They haveseverely reduced mobility, experience constantpain, and are unable to work. One woman pre-sented with polyarteritis nodosa-like symptomsand with proteinuria. One carrier presented
with muscle pains and localised proximal mus-cle atrophy in the lower limb and neurologicalinvestigations failed to identify the cause.
MEDICAL CARE RATINGS
The majority of carriers (66%, n=59) rated theAFD information given at diagnosis as excel-lent or satisfactory and 44% rated it as poor.Genetic counselling was evenly divided be-tween poor and satisfactory. Routine medicalassessment is not being carried out in two-thirds of women.
REPRODUCTIVE DECISION MAKING
Responses are shown in fig 2.
EMOTIONAL HEALTH
About a third of the sample reported beinghappy, full of life, and energetic. Another thirdadmitted to being depressed, anxious, tired,and frustrated some of the time; 15.8% weretired all the time, 5.7% were frustrated, and5.4% were suicidal most/all of the time.
Personal needs summary listed the follow-ing: education of medical practitioners aboutclinical manifestations of AFD, improvementin genetic counselling, access to ERT trialsinformation with implications for treatment incarriers and young males aVected with AFD,regular monitoring for heart and kidney prob-lems, and medical management with moreunderstanding and sympathy.
DiscussionThis is the first report of a large cross sectionalcohort survey in adult obligate carriers forAFD. Included were all obligate carriers knownto the UK AFD register with confirmedpedigree structure and AFD diagnosis. Thoseunder the age of 18 were excluded as publishedreports and clinical observation suggestedmainly adult onset of AFD complications. Thedata are based on an AFD specific question-naire survey for self reported signs andsymptoms, and on clinical findings in sympto-matic AFD carriers seeking medical attention.Asymptomatic women were not investigated inthis study, as the present health care policydoes not recognise them as patients aVectedwith AFD. The results of abnormal renal andcardiac function reported here are thereforeunderestimates.
MORTALITY
The median cumulative survival in 32 obligatecarriers was 70 years, which represents anapproximate reduction of 15 years from thegeneral population. A gradual decline of theAFD female carrier curve is evident from theage of 35. These data suggest milder expressionof the disease in AFD carriers than in aVectedmales.
SERIOUS MANIFESTATIONS
The frequency of serious AFD manifestationsin carriers is unknown. A number of carrierswith cerebrovascular complications or withrenal failure were reported.1
Table 4 Transient ischaemic attacks in AFD carriers
Case No Age (years) Signs and symptoms
1 54 Arm weakness, migraine headaches2 68 Arm weakness3 23 Diplopia lasting minutes, ∼6/year4 55 Arm and face weakness lasting minutes5 66 Face weakness, loss of speech6 33 Diplopia, arm weakness × 2 lasting minutes7 59 Diplopia frequent, blindness for 45 minutes, ∼4/year,
ipsilateral numbness UL10 73 Diplopia11 69 Arm weakness lasting minutes12 70 Diplopia at 50 y × 2, blurred vision, dizziness, blindness
lasting ∼1 minute13 52 Arm weakness, blindness lasting ∼1 minute14 50 Diplopia15 38 Blurred vision, migraine headaches16 45 Diplopia, migraine headaches17 39 Unilateral weakness UL+LL lasting hoursExcluded18 55 Weak and stiV joints19 45 Weak, hot, and swollen leg and shoulders20 56 Fibromyalgia, heavy legs often unable to bear weightUnknown21 30 No description given22 42 No description given23 44 No description given
Signs and symptoms of TIAs reported by AFD carriers. UL, upper limbs; LL, lower limbs.
Figure 2 Reproductive decisions taken by AFD carrier women who had children after1980. TOP=termination of pregnancy.
A review of clinical manifestations in heterozy-gotes for AFD estimated the frequency of neu-ropathic pain as <10%.1 The accumulation ofsphingolipid was determined at necropsy in thespinal and sympathetic ganglia in a carrier withsevere neuropathic pains and cardiomyopathy.A significant increase of 34- and 48-foldrespectively above normal was found,5 thusexplaining the likely aetiology of severe neuro-pathic pains.
In this cohort, a substantially higher fre-quency of pain was evident and, as in AFDmales, the mean onset of neuropathic pains incarriers was 15 years and in 20% the pain hadstopped at the age of 28. The pain scores forboth attacks and background pain in carrierswere surprisingly high, >5, and suggest thatthese women require improved pain manage-ment. Only 17.1% of carriers were receivinganticonvulsants or opiates. High scores wereobtained for the eVects of pain on attendance atschool, participation in sports, and the impacton family and social life. A third to half ofrespondents scored the impact of pain on theirlife as being high. Similarly to aVected males,the duration of the neuropathic pain appears tobe life long and stops in a small percentage ofwomen.
TIA AND CVA
The frequency of cerebrovascular complica-tions in carriers was estimated as <1%, occur-ring in rare female variants with 0-5%á-galactosidase activity.1 Neurological manifes-tations such as TIAs, CVA, tinnitus, severehypostatic hypotension, and sensorineuralhearing loss have been documented as casereports.6 A cohort study of 10 female heterozy-gotes with cerebrovascular complications re-ported the mean age of onset as 40.3 years andthe signs and symptoms, in descending order offrequency, as: memory loss, dizziness, ataxia,hemiparesis, loss of consciousness, and he-misensory symptoms.7 In another study, theincidence of thrombotic accidents was 10%(n=15), where two carriers had CVA aged 43and 49 respectively and one had central retinalartery thrombosis at 31.8 The incidence of TIAin carriers is unknown.
In this cohort, a direct question was askedabout past history of TIA and a high frequencywas found (table 4). Several carriers in thisgroup were in their 60s and 70s, whencerebrovascular events are common and notthe result of AFD. However, the mean age forTIA at 52 years indicates that in the majority ofthese women the onset of TIA was early andmost likely the result of AFD pathology.Results from this cohort suggest a higherfrequency of cerebrovascular complicationsthan previously reported.
Two unrelated carriers in this cohort hadmild idiopathic MR, and both have developedESRF, aged 35 and 37; further investigationsare ongoing (one of these was not included inthis study because of failure to return the ques-tionnaire).
RENAL FUNCTION
The incidence of renal failure in carriers wasestimated as <1%, occurring in rare femalevariants with 0-5% á-galactosidase activity.1
Abnormal renal function, proteinuria, andESRF were reported in several female carriers.1
Characteristic renal changes of AFD were alsodescribed in asymptomatic carriers with nor-mal renal function, when investigated tobecome kidney donors for their male relatives.9
Two heterozygotes had a mosaic pattern offocal patchy AFD inclusions/degenerativechanges among normal appearing glomeruli onelectron microscopy.9 The Gb3 accumulationin the kidney was age related and Gb3 contentin urine sediment of heterozygotes was in-creased up to 20-fold compared with normal.1
From these reports to date, we can concludethat probably all carriers accumulate Gb3 inthe kidneys and usually at a much slower ratethan AFD males. The rate of Gb3 depositionappears variable.
In this cohort, two women (3.3%) developedESRF at ages 35 and 37 respectively, and so farneither has had a renal transplant. In these twowomen, the age at onset was comparable toAFD males, but the frequency of ESRF wasmuch lower in carriers than the 26% found inhemizygous males.10 The frequency of abnor-mal renal function in this cohort could not bedetermined because of lack of data.
CARDIAC FUNCTION
The incidence of cardiac involvement in carri-ers was estimated as <1%, occurring in rarefemale variants with 0-5% á-galactosidaseactivity.1 Cardiomyopathy and MV prolapsewere documented in two studies of 13 and 12carriers, respectively. Sakuraba et al11 describedMV prolapse, dilated aortic root, concentricLVH, and ECG abnormalities in 5/13 womenaged 36-64. Only those over 50 years of age hadsymptoms of palpitations and dyspnoea. Gold-man et al12 found no LVH on echocardiogramin 12 carriers, who were however younger, agedbetween 25 and 37 years. In both studies, thefrequency of MV prolapse was over 50%, withadditional patients having redundant MV.Renovascular hypertension was present in2/13. These cohort data suggest that the onsetof cardiac lipid storage is evident around theage of 35 years in carriers and is present inaround 50%. Cardiac histopathology in onecarrier showed focal granular staining patternwith anti-CTH monoclonal antibody.13 Thisfemale developed proteinuria and LVH at 22years, subsequently had CVA and renal andcardiac failure, and died of ventricular fibrilla-tion aged 60. This report documented a mosaicpattern of myocardial involvement and subse-quent multiorgan progression of AFD over 40years. Further documentation of cardiomyopa-thy was shown by increased thallium-201uptake by scintigraphy in three carriers14 and a53 year old female carrier presented with endstage restrictive cardiomyopathy and under-went heart transplantation.15
In this cohort, chest pain and palpitationswere reported by 52.5%, but only a third were
investigated. Four carriers had mitral valvereplacement aged between 55 and 65 years.
GASTROINTESTINAL MANIFESTATIONS
Gastrointestinal structure and function wereinvestigated in 13 aVected males and in 17 het-erozygotes.16 Twenty-nine percent of heterozy-gotes were symptomatic and jejunal biopsyshowed characteristic electron dense sphingoli-pid deposits situated in the neuronal andvascular tissue in the bowel submucosa. Theselesions were not visible by light microscopy.The bowel involvement in carriers was less fre-quent than in males and less pronounced. Theonset of symptoms was later than in males, atbetween 15-35 years, and several girls also hadmild growth delay. In this cohort, 58.3% ofcarriers reported having GI symptoms, whichincluded abdominal pain, diarrhoea, or consti-pation and vomiting. Most women had exten-sive GI and gynaecological investigations.None of the carriers in this cohort experiencedGI complications in the form of bowel perfora-tion or infarction and the severity of symptomsappeared to progress with age. In none of thesewomen was the cause of GI symptomsidentified.
OTHER MANIFESTATIONS OF AFD
Fatigue was the single most common manifes-tation reported by 66% of carrier women,which interfered with their family and sociallife. The second commonest manifestation wasangiokeratoma and was reported by 35%. Thisobservation is in agreement with the frequencyreported previously,1 but discrete patches ofangiokeratoma observed in this study becameevident only on careful clinical examination.The findings of angiokeratoma in dermatomedistribution supports the concept of mosaic Xinactivation, reported in X linked disorders,such as incontinentia pigmenti.17 The inci-dence of hypohidrosis in carriers was estimatedas <1%, occurring in rare female variants with0-5% á-galactosidase activity.1 In this cohort,subjective hypohidrosis was reported by 32.8%of women documenting higher frequency thanexpected. Recurrent polyarthralgia in thehands and feet, often with pyrexia, occursfrequently at an early age and the diagnosis ofrheumatic fever, chronic inflammatory syn-drome, or polyarteritis nodosa has beenmade.18 Perivascular inflammation and polyar-teritis nodosa-like necrotising vasculitis patho-logical changes were found in several tissuesincluding the bowel in an AFD carrier.19
In this cohort, severe disabling fibromyalgiawas diagnosed in five women aged between 45and 55 years and one had proximal muscleatrophy. A similar presentation was describedin an AFD male, who had characteristic sphin-golipid deposits seen on muscle biopsy.20 Largejoint pains were a frequent complaint of AFDcarriers over the age of 50, but these were notformally assessed.
REPRODUCTIVE DECISIONS
The majority of women understood X linkedinheritance and their carrier status was knownbefore pregnancy. Nearly half of women of
childbearing age opted against having morechildren because of AFD. Prenatal diagnosiswas used by only half of the women in thiscohort, most likely reflecting the dilemma oftermination of pregnancy for a disease with lateonset of complications.
In this study, obligate female carriers re-ported multiple AFD manifestations, with highfrequency, which was not recognised previ-ously. However, similar findings were recentlyreported, in cohorts of 12 and 21 female carri-ers.21 22 The progression of the disease is slowerin carrier females, as would be expected insubjects with partial enzyme deficiency. Thepatchy distribution of cells with lipid deposi-tion in cardiac, renal, and skin tissues confirmsrandom X inactivation even in carriers withsevere manifestations of AFD.
We are grateful to the AFD patient support group, to the fami-lies for their cooperation and to Transkaryotic Therapies Inc,Boston, USA for financial support of AH. Professor B Winches-ter at the Enzyme and Molecular Genetics Laboratory, Instituteof Child Health, London analysed most of the samples from thiscohort.
1 Desnick RJ, Ionnou Y, Eng CM. Fabry disease: alpha galac-tosidase A deficiency. In: Scriver CH, Beaudet AL, Sly WS,Valle D, eds. The metabolic and molecular bases of inheriteddisease. New York: McGraw Hill, 1995:2741-84.
2 Ropers HH, Wienker TF, Grimm T, Schroetter K, BenderK. Evidence for preferential X-chromosome inactivation ina family with Fabry disease. Am J Hum Genet 1977;29:361-70.
3 Davies JP, Eng CM, Hill JA, Malcolm S, MacDermot K,Winchester B, Desnick RJ. Fabry disease: fourteenalpha-galactosidase A mutations in unrelated families fromthe United Kingdom and other European countries. Eur JHum Genet 1996;4:219-24.
4 Melzack R. The short-form McGill Pain Questionnaire.Pain 1987;30:191-7.
5 Hozumi I, Nishizawa M, Ariga T, Miyatake T. Biochemicaland clinical analysis of accumulated glycolipids in sympto-matic heterozygotes of angiokeratoma corporis diVusum(Fabry’s disease) in comparison with hemizygotes. J LipidRes 1990;31:335-40.
6 Bird TD, LagunoV D. Neurological manifestations of Fabrydisease in female carriers. Ann Neurol 1978;4:537-40.
8 Utsumi K, Yamamoto N, Kase R, Takata T, Okumiya T,Saito H, Suzuki, T, Uyama E, Sakuraba H. High incidenceof thrombosis in Fabry’s disease. Intern Med 1997;36:327-9.
9 Gubler MC, Lenoir G, Grunfeld JP, Ulmann A, Droz D,Habib R. Early renal changes in hemizygous andheterozygous patients with Fabry’s disease. Kidney Int1978;13:223-35.
10 MacDermot KD, Holmes H, Miners AH. Anderson-Fabrydisease: clinical manifestations and impact of disease in acohort of 98 hemizygous males. J Med Genet 2001;38:750-60.
11 Sakuraba H, Yanagawa Y, Igarashi T, Suzuki Y, Suzuki T,Watanabe K, Ieki K, Shimoda K, Yamanaka T. Cardiovas-cular manifestations in Fabry’s disease. A high incidence ofmitral valve prolapse in hemizygotes and heterozygotes.Clin Genet 1986;29:276-83.
12 Goldman ME, Cantor R, Schwartz MF, Baker M, DesnickRJ. Echocardiographic abnormalities and disease severityin Fabry’s disease. J Am Coll Cardiol 1986;7:1157-61.
13 Itoh K, Takenaka T, Nakao S, Setoguchi M, Tanaka H,Suzuki T, Sakuraba H. Immunofluorescence analysis of tri-hexosylceramide accumulated in the hearts of varianthemizygotes and heterozygotes with Fabry disease. Am JCardiol 1996;78:116-17.
14 Tsuda T, Yokoyama A, Masani F, Kodera K, Yamamoto T,Watanabe K, Izumi T, Shibata A, Kimura M. Myocardialinvolvement in female Fabry’s disease: evaluation bythallium-201 myocardial scintigraphy (in Japanese). J Car-diol 1988;18:135-44.
15 Cantor WJ, Daly P, Iwanochko M, Clarke JT, Cusimano RJ,Butany J. Cardiac transplantation for Fabry’s disease. CanJ Cardiol 1998;14:81-4.
16 Sheth KJ, Werlin SL, Freeman ME, Hodach AE. Gastro-intestinal structure and function in Fabry’s disease. Am JGastroenterol 1981;76:246-51.
17 Happle R. The lines of Blaschko: a developmental patternvisualizing functional X-chromosome mosaicism. CurrProbl Dermatol 1987;17:5-18.
18 Dubost JJ, Sauvezie B, Galtier B, Tixeron J, Rampon S.Fabry’s disease. Rare etiology of a long-term inflammatorysyndrome. Apropos of a case (in French). Rev Rhum MalOsteo-Articulaires 1986;53:525-28.
20 Uchino M, Uyama E, Kawano H, Hokamaki J, KugiyamaK, Murakami Y, Yasue H, Ando M. A histochemical andelectron microscopic study of skeletal and cardiac musclefrom a Fabry disease patient and carrier. Acta Neuropathol1995;90:334-8.
21 Whybra C, Bunge S, Beck M. Clinical manifestation infemale Fabry disease patients and molecular geneticscarrier diagnosis (abstract). 5th International Symposium onMucopolysaccharide and Related Diseases, Vienna, 1999.
22 Whybra C, Beck M. Neurological complications in femaleFabry patients (abstract). 6th International Symposium onMucopolysaccharide and Related Diseases, Minneapolis,2000.
Thrombophilic polymorphisms in pre-eclampsia:altered frequency of the functional 98C>Tpolymorphism of glycoprotein IIIa
Kevin M O’Shaughnessy, Beiyuan Fu, Sarah Downing, Nicholas H Morris
EDITOR—Pre-eclampsia is a heritable endothe-lial disorder, which is unique to pregnancy1 andcharacteristically associated with haemostaticand thrombophilic abnormalities. This associ-ation has led to the identification of a numberof gene variants that might confer throm-bophilic risk in pre-eclampsia. An increasedcarrier rate for two of these polymorphisms,factor V Leiden2 and the thermolabile variantof methylenetetrahydrofolate reductase(MTHFR),3 has been reported in some womenwith pre-eclampsia. However, we have beenunable to replicate these findings in our ownEast Anglian population.4
We have now looked at two further candidatethrombophilic polymorphisms in our pre-eclampsia cohort that are involved in the regu-lation of vascular thrombosis. The first is the20210G>A polymorphism in the 3' UTRregion of the prothrombin (PT) gene thatcauses a modest rise in plasma prothrombinlevels,5 and is reportedly associated with severepre-eclampsia in an Israeli cohort.6 The secondis a coding variant (98C>T) in exon 2 of theGPIIIa gene that provides the common betasubunit for several â3-integrins including theplatelet fibrinogen receptor. This polymor-phism causes a 33Leu>Pro substitution andthe existence of two antigenically distinct formsof the mature GPIIb/IIIa antigen on platelets(the Pl(A) antigens 1 and 2). Loss of functionalGPIIIa is associated with a rare bleeding disor-der (Glanzmann’s thrombasthenia) and the33Pro variant itself has been associated withrisk of premature acute coronary syndromesand stroke in young white women.7
Materials and methodsThe method of recruitment and the definitionof pre-eclampsia used in our cohort have beenpublished previously.4 In brief, they all hadproteinuric pregnancy related hypertension asdefined by the criteria of Redman andJeVeries.8 The local ethics committee approvedthe study and all patients gave written in-formed consent before taking part. At the timeof this study, we had genomic DNA availablefrom 356 women with pre-eclampsia. All thewomen were white northern Europeans. The200 normotensive control samples were those
reported in our earlier study4 and had nohistory of pregnancy induced hypertension orpre-eclampsia.
Genomic DNA was obtained from venousblood using standard methods and genotypedfor the two polymorphisms using PCR/RFLPas follows.
GPIIIa C1565T POLYMORPHISM
Exon 2 of the GPIIIa gene was amplified usingforward (5' TCT GAT TGC TGG ACT TCTCTT) and reverse (5' TCT CTC CCC ATGGCA AAG AGT) intronic primers (Genbankaccession No AH00311). The 266 bp productwas digested with 10 U of NciI (New EnglandBiolabs) overnight and fragments resolved on a3% Metaphor® gel (FMS Bioproducts, Rock-port, USA).
GENOTYPING OF THE PT G20210A POLYMORPHISM
The 3' UTR region of the PT gene was ampli-fied using previously reported forward (5' TCTAGA AAC AGT TGC CTG GC) and reverse(5' CAA CGC CTG GTA TCA AAT GG)primers.5 The 345 bp PCR product wasdigested overnight with 10 U of HindIII (NewEngland Biolabs). Fragments were again re-solved using a 3% Metaphor® gel.
The genotype and allele frequencies werecompared by ÷2 analysis. Hardy-Weinbergequilibrium was tested using the HWE pro-gram of the LINKUTIL package (downloadedfrom http://linkage.rockefeller.edu/ott/linkutil.htm). Odds ratios and confidence intervalswere calculated by the method of Wolf.9 A pvalue of <0.025 (0.05/2) was consideredsignificant to reflect a necessary Bonferronicorrection.
ResultsThe genotyping results for the two polymor-phisms are shown in table 1. Both polymor-phisms were in Hardy-Weinberg equilibrium inboth groups of subjects. The carrier rate for theprothrombin A20210 variant was not signifi-cantly diVerent between our control andpre-eclamptic groups and no subjects in eithergroup were homozygous for the A20210variant. There was no evidence that the
J Med Genet2001;38:775–777
Clinical PharmacologyUnit, Department ofMedicine, University ofCambridge ClinicalSchool, Cambridge, UKK M O’ShaughnessyB FuS Downing
Academic Departmentof Obstetrics andGynaecology, Chelseaand WestminsterHospital and NorthLondon Hospitals Trust,London, UKN H Morris
Correspondence to: DrO’Shaughnessy, ClinicalPharmacology Unit, Level 6ACCI, Addenbrooke’sHospital, CambridgeCB2 2QQ, UK,[email protected]
G20210A polymorphism was a severity markerwithin the pre-eclampsia group, 5/149 (3.4%)delivering before 37 weeks of gestation and2/42 (4.8%) with severe disease (defined ashaving HELLP syndrome or an eclamptic sei-zure) were A20210 carriers. These were notsignificantly diVerent frequencies from thepre-eclampsia group as a whole.
For the glycoprotein IIIa polymorphism,there was a significant excess of C98 homozy-gotes in the pre-eclampsia group (18/356) ver-sus controls (1/200, p<0.01) with a corre-sponding excess of the C98 allele (p<0.02).Expressing the relative risk as odds ratios: forC98 versus T98 homozygosity it was 11.3(95% CI 1.5-85.5), and for carriage of at leastone C98 allele it was 1.4 (95% CI 1.0-2.1). Inthe pre-eclampsia group, 9/149 (6.0%) deliver-ing before 37 weeks and 3/42 with severedisease (see above) were C98 homozygotes(7.1%). Again, these were not significantly dif-ferent frequencies from the pre-eclampsiagroup as a whole.
DiscussionIn our population, the carrier frequency for theA20210 prothrombin gene variant was withinthe range previously reported for white popula-tions of 0.7-4%.10 However, our results con-trast sharply with previous reports that carriageof the A20210 variant is commoner in womenwith pre-eclampsia.6 7 These studies found athree to four-fold increase in the frequency ofthis thrombophilic variant, although the num-bers of subjects studied were small, consistingof just 34 subjects with pre-eclampsia com-pared to over 350 in our current study group.6
DiVerences in the severity of pre-eclampsiabetween studies is unlikely to be importantsince the carrier rate for A20210 within ourown sample was not aVected by severity,expressed either as premature delivery, or thepresence of HELLP syndrome, or eclampticseizure as used in the study of Kupferminc etal.6 Interestingly, the most recent report fromKupferminc’s group, using an increased sam-ple size, has failed to replicate the original find-ings in patients with severe pre-eclampsia.12
This work is in keeping with our current nega-tive results and suggests that the original publi-cation may have represented a type I error.Even supposing the G20210A is a susceptibil-ity gene variant for pre-eclampsia, its rolewould be restricted to certain ethnic groups; itis virtually absent, for example, in non-whitepopulations.10
In contrast, we did see an excess of homozy-gotes for the T98 allele of GPIIIa (17 in thepre-eclamptics versus 1 in the controls, table 1)
and overall the frequency of the T98 allele wassignificantly increased in our pre-eclampticsubjects (20.1% versus 14%, p<0.01). Thisraises the intriguing possibility that the GPIIIaC98 polymorphism may predispose women topre-eclampsia. Previous association studieswith this polymorphism have modelled itseVects as a dominant allele, largely reflectingthe bias of early studies that could type on thebasis of the presence or absence of the Pl(A)1and Pl(A) 2 surface antigens. Expressing ourdata in this fashion gives an odds ratio forwomen carrying at least one T98 allele of 1.4.Although this appears a rather modest eVect, itis similar to that previously published for thefactor V Leiden and MTHFR variants inwomen with pre-eclampsia.2 3
The GPIIIa gene codes for the â3 subunit ofthe â3 integrin subfamily; this consist of áIIbâ3(glycoprotein IIb/IIIa) and ávâ3. The glyco-protein IIb/IIIa receptor is confined to plateletsand megakaryocytes and has a key role inplatelet aggregation by binding fibrinogen andvon Willebrand factor. In contrast, the ávâ3integrin is a widely expressed receptor forligands such as vitronectin and fibronectin. Itappears to be an important player in tumourangiogenesis, but is also expressed by invadingtrophoblast suggesting a role in placentation.13
Indeed, mice knockouts in which the â3integrin is deficient have normal implantationbut develop a characteristic placental defect.14
The â3 integrin has also been implicated in thefailure of the cytotrophoblast to adopt a vascu-lar phenotype in pre-eclampsia.15
Previous epidemiological work on the C98TGPIIIa polymorphism has focused on its roleas a risk allele for arterial thrombosis, specifi-cally for acute coronary syndromes. Weiss et al16
initially reported that carriage of the T98 vari-ant of GPIIIa was a risk factor for myocardialinfarction in young patients. This has beenconfirmed in a large cohort of sib pairs withpremature heart disease and probably explainswhy some studies have failed to replicate theoriginal work of Weiss et al16 because they usedelderly patient cohorts.17 Clear evidence nowexists that the T98 GPIIIa allele that carries thethrombotic risk is functionally distinct from theC98 wild type receptor. This is apparent bothin terms of the level of IIb/IIIa expression onactivated platelets as well as in the adhesionproperties of cells expressing ávâ3.18 19 Giventhe importance of platelet and trophoblastfunction in the pathogenesis of pre-eclampsia,the GPIIIa gene is a very attractive biologicalcandidate gene for pre-eclampsia and the func-tional C98T polymorphism in particular. Ourfindings here that the thrombotic risk variant,T98, is commoner in pre-eclamptic womensuggests that we may have identified a new riskallele for the disease.
In summary, we have not been able to showany association of the prothrombin polymor-phism, G20210, with pre-eclampsia. There issignificant excess, however, of the thromboticvariant of the â3 integrin protein GPIIIa in thesame population. This result needs confirma-tion in other populations, preferably usingrobust family based association studies, but it
Table 1 Comparison of the frequencies of the genotypes and alleles for the TP 20210 andGPIIIa C98T polymorphisms in subjects with pre-eclampsia versus controls
No (% frequencies) of individual genotypes No (% freqencies) of alleles
suggests that integrin polymorphisms might bean important new focus for studies on geneticrisk factors for pre-eclampsia.
This work was funded by the British Heart Foundation.
1 Morris N, Eaton BM, Dekker G. Nitric oxide, the endothe-lium, pregnancy and pre-eclampsia. Br J Obstet Gynaecol1996;104:4-15.
2 Dizon TD, Nelson LM, Easton K, Ward K. The factor VLeiden mutation may predispose women to severepreeclampsia. Am J Obstet Gynecol 1996;175:902-5.
3 Grandone E, Margaglione M, Colaizzo D, Cappucci G, Pal-adini D, Martinelli P, Montanaro S, Pavone G, Di MinnoG. Factor V Leiden, C>T MTHFR polymorphism andgenetic susceptibility to preeclampsia. Thromb Haemost1997;77:1052-4.
4 O’Shaughnessy KM, Fu B, Ferraro F, Lewis I, Downing S,Morris NH. The factor V Leiden and thermolabile methyl-enetetrahydrofolate reductase gene variants in an EastAnglian pre-eclampsia cohort. Hypertension 1999;33:1338-41.
5 Poort SR, Rosendaal FR, Reitsma PH, Bertina RM. A com-mon genetic variation in the 3'-untranslated region of theprothrombin gene is associated with elevated plasmaprothrombin levels and an increase in venous thrombosis.Blood 1996;88:3698-703.
6 Kupferminc MJ, Eldor A, Steinman N, Many A, Bar Am A,JaVa A, Fait G, Lessing JB. Increased frequency of geneticthrombophilia in women with complications of pregnancy.N Engl J Med 1999;340:9-13.
7 Carter AM, Catto AJ, Bamford JM, Grant PJ. Platelet GPIIIa PlA and GP Ib variable number tandem repeatpolymorphisms and markers of platelet activation in acutestroke. Arterioscler Thromb Vasc Biol 1998;18:1124-31.
8 Redman CW, JeVeries M. Revised definition of pre-eclampsia. Lancet 1988;i:809-12.
9 Gardiner MJ, Altman DG. Statistics with confidence. London:BMJ Publishing Group, 1989:51. ISBN 0-7279-0222-9.
10 Rosendaal FR, Doggen CJ, Zivelin A, Arruda VR, Aiach M,Siscovick DS, Hillarp A, Watzke HH, Bernardi F,Cumming AM, Preston FE, Reitsma PH. Geographic dis-tribution of the 20210 G to A prothrombin variant. ThrombHaemost 1998;79:706-8.
11 Kupferminc MJ, Fait G, Many A, Gordon D, Eldor A, Less-ing JB. Severe preeclampsia and high frequency of geneticthrombophilic mutations. Obstet Gynecol 2000;96:45-9.
12 Kupferminc MJ, Peri H, Zwang E, Yaron Y, Wolman I,Eldor A. High prevalence of the prothrombin genemutation in women with intrauterine growth retardation,abruptio placentae and second trimester loss. Acta ObstetGynecol Scand 2000;79:963-7.
13 Thirkill TL, Douglas GC. The vitronectin receptor plays arole in the adhesion of human cytotrophoblast cells toendothelial cells. Endothelium 1999;6:277-90.
14 Hynes RO, Hodivala-Dilke KM. Insights and questionsarising from studies of a mouse model of Glanzmannthrombasthenia. Thromb Haemost 1999;82:481-5.
15 Zhou Y, Damsky CH, Fisher SJ. Preeclampsia is associatedwith failure of human cytotrophoblasts to mimic a vascularadhesion phenotype. One cause of defective endovascularinvasion in this syndrome? J Clin Invest 1997;99:2152-64.
16 Weiss EJ, Bray PF, Tayback M, Schulman SP, Kickler TS,Becker LC, Weiss JL, Gerstenblith G, Goldschmidt-Clermont PJ. A polymorphism of a platelet glycoproteinreceptor as an inherited risk factor for coronary thrombo-sis. N Engl J Med 1996;334:1090-4.
17 Goldschmidt-Clermont PJ, Coleman LD, Pham YM,Cooke GE, Shear WS, Weiss EJ, Kral BG, Moy TF, YookRM, Blumenthal RS, Becker DM, Becker LC, Bray PF.Higher prevalence of GPIIIa PlA2 polymorphism insiblings of patients with premature coronary heart disease.Arch Pathol Lab Med 1999;123:1223-9.
Cystic fibrosis patients with the 3272-26A>Gsplicing mutation have milder disease thanF508del homozygotes: a large European study
Margarida D Amaral, Paula Pacheco, Sebastian Beck, Carlos M Farinha, Deborah Penque,Paulo Nogueira, Celeste Barreto, Beatriz Lopes, Teresa Casals, Javier Dapena, Silvia Gartner,Carlos Vásquez, Javier Pérez-Frías, Casilda Olveira, Rodrigo Cabanas, Xavier Estivill,Maria Tzetis, Emmanuel Kanavakis, Stavros Doudounakis, Thilo Dörk, Burkhard Tümmler,Emmanuelle Girodon-Boulandet, Cécile Cazeneuve, Michel Goossens, Martine Blayau,Claudine Verlingue, Isabel Vieira, Claude Féréc, Mireille Claustres, Marie des Georges,Christine Clavel, Philippe Birembaut, Dominique Hubert, Thierry Bienvenu,Michèle Adoun, Jean-Claude Chomel, Kris De Boeck, Harry Cuppens, Joao Lavinha
EDITOR—Cystic fibrosis (CF, MIM 219700) is acommon, severe, autosomal recessive diseasecaused by mutations in the CF transmembraneconductance regulator (CFTR) gene cloned in1989.1–3 The disease, characterised by chronic
lung disease which is the main cause of morbid-ity and mortality, pancreatic dysfunction, raisedelectrolyte levels in sweat, and male infertility, iscaused by altered chloride (Cl−) secretion acrossthe apical membrane of epithelial cells.4 There
+ This study investigated the role of twothrombophilic gene polymorphisms, pro-thrombin 20210G>A and the 98C>Tvariation in the â3 integrin glycoproteinIIIa (GPIIIa), as risk factors for pre-eclampsia in an East Anglian cohort of356 aVected women.
+ In contrast to an earlier report, we foundno excess of carriers for the prothrombin20210A variant in our cohort. However,for the GPIIIa 98C>T polymorphismthere was an excess of 98T homozygotesin our pre-eclampsia group (18/356) ver-sus controls (1/200, p<0.01). The oddsratio for 98T homozygotes was 11.3(95% CI 1.5-85.5) and for carriage of atleast one T98 allele, 1.4 (95% CI1.0-2.1).
+ These data suggest that carriage of the98T polymorphism of GPIIIa may be apreviously unidentified risk factor forpre-eclampsia.
is, however, substantial variability in the clinicalmanifestations aVecting the various organs.4 5
One single mutation, F508del, generallyassociated with severe disease, accounts forabout 70% of CF chromosomes world wide,although with a heterogeneous geographicaldistribution.5 Patients homozygous for theF508del mutation have the classical severeform of the disease which includes chronicmucous obstruction of the lung and conduct-ing airways, followed by recurrent infectionsmostly by Pseudomonas aeruginosa (Pa) andStaphylococcus aureus (Sa), exocrine pancreaticinsuYciency (PI), resulting in failure to gainweight and height, and raised levels of Cl−,sodium, and potassium in exocrine sweat.5
However, almost 1000 genetic alterations havebeen detected in the CFTR gene (CFTR Muta-tion Database), most presumed to be diseasecausing mutations. About half of these areamino acid substitutions (missense mutations)and about 20% are splicing mutations. Theremainder are nonsense, frameshift (includingsmall deletions and insertions), and a smallproportion of promoter mutations.
The relationship between genotype, that is,the mutations in the CFTR gene, and the clini-cal phenotype of CF patients has been diYcultto establish, in particular for lung disease.
It was previously shown that the 3272-26A>G mutation leads to the creation of analternative acceptor splice site competing withthe normal one during RNA processing andresulting in the occurrence of an alternativelyspliced mRNA with 25 extra nucleotides fromintron 17a and a premature stop codon soonthereafter.6 Previously it was reported thatthree patients carrying the 3272-26A>G muta-tion in one of their CFTR alleles had mild CF.7
Another patient with the 3272-26A>G/F508del genotype was reported to have severeCF.8 Here, we report the clinical phenotypes of60 CF patients from several European centres,with the 3272-26A>G mutation on one CFTRallele, and mostly another severe mutation(73% F508del) on the other allele. Wecompare them, by statistical methods, with theclinical phenotypes of F508del homozygotes(n=89) from the same centres, matched for ageand sex as exactly as possible.
Subjects and methodsPATIENTS
Patients were clinically and genetically charac-terised in the various CF centres involved inthis study (table 1). For the control group, one
(two, when possible) F508del homozygouspatient for each patient with the 3272-26A>Gmutation was selected from the same centres,matching for age and sex with the 3272-26A>G patients as far as possible. However,CF patients with 3272-26A>G were generallyamong the oldest CF patients in each centre.Therefore, for some cases it was not possible tohave two control F508del homozygotes exactlymatched. In such cases only one was includedand this was chosen as the oldest F508delhomozygote from the same centre. The meancurrent age (SD) of patients in the groups of3272-26A>G (n=60) and F508del homo-zygous (n=89) patients were 20.5 (SD 17.5)and 17.0 (SD 11.5) years, respectively. It wasfound, however, that this diVerence was notsignificant (data not shown).
DNA ANALYSIS
Genomic DNA was isolated from peripheralblood lymphocytes according to standard pro-tocols. The 3272-26A>G mutation (as well asnon-F508del mutations on the other allele)were detected either by single stranded confor-mation analysis (SSCA)9 or by denaturinggradient gel electrophoresis (DGGE)10 afterPCR amplification of genomic DNA in theregion of the corresponding CFTR exon.Amplicons with abnormal patterns were se-quenced either by the ABI PRISM™ Dye Ter-minator Cycle Sequencing System (Perkin-Elmer, Norwalk, CA) or by the dideoxy manualmethod with a [35S] nucleotide.
Detection of the F508del mutation waseither by dot blotting, by amplification refrac-tory mutation system (ARMS),11 heteroduplexanalysis on polyacrylamide gel electrophoresis(HA-PAGE),12 or oligonucleotide ligation assay(OLA).13
Microsatellites IVS8(CA)n were investigatedas described previously.14 The IVS8 (TG)nTm
polymorphic tract was also analysed as previ-ously described15 and sequenced with thefollowing primer: 5'GAAATTACTGAAGAA-GAGGC3'. The GATT tetranucleotide inintron 6a IVS6a-(GATT)n was analysed byPCR amplification and electrophoresis of theproducts in a 12% (w/v) polyacrylamide gel.16
The diallelic markers XV-2c/TaqI, KM.19/PstI,MP6-D9/MspI, J44/XbaI, M470V/HphI(1540A>G), and T854/AvaII (2694G/T) wereanalysed by PCR amplification and digestionwith appropriate restriction enzymes as de-scribed previously.17
Table 1 Distribution of CF patients in this study according to their genotype and country
Country/No patients
3272-26A>G/anyControl group(F508del/F508del)F508del Other mutation
Centro de GenéticaHumana, InstitutoNacional de Saúde DrRicardo Jorge, Av PadreCruz, 1649-016 Lisboa,PortugalM D AmaralP PachecoS BeckC M FarinhaD PenqueJ Lavinha
Departmento de Químicae Bioquímica, Faculdadede Ciências daUniversidade de Lisboa,Lisboa, PortugalM D AmaralC M Farinha
Observatório Nacional deSaúde, Instituto NacionalSaúde Dr Ricardo Jorge,Lisboa, PortugalP Nogueira
Unidade de FibroseQuística, Serviço dePediatria, Hospital de StªMaria, Lisboa, PortugalC Barreto
Hospital D Estefânia,Lisboa, PortugalB Lopes
Molecular GeneticsDepartment-IRO,Hospital Duran I Reynals,Barcelona, SpainT CasalsX Estivill
Cystic Fibrosis Unit,Hospital UniversitarioVirgen del Rocio, Sevilla,SpainJ Dapena
The clinical data included in this study wereage at diagnosis, sweat test values, pulmonarystatus assessed by forced expiratory volume inone second (FEV1) % predicted, and forcedvital capacity (FVC) % predicted (which arepredicted values for the non-CF population byKnudson et al18), lung colonisation with bacte-rial pathogens, pancreatic status, history ofmeconium ileus, weight and height centiles (forpatients under 18) or body mass index (BMI,for patients over 18), Chrispin-Norman (CN)chest radiological score (from 0 to 38, with 0being the best score, as defined by Conway andLittlewood19), Shwachman-Kulczycki (SK)general status score (100 is the best score, alsoas defined by Conway and Littlewood19), nasalpolyposis, and other clinical complications orabnormalities.
STATISTICAL ANALYSIS
Data for all clinical parameters are presented(table 2) as the mean (SD) as well as medianand interquartile deviation (QD). For distribu-tions of quantitative measurements, the hy-pothesis of equal variances between the twogroups of CF patients under study was testedusing Levene’s test and there was no evidenceto reject it (p>0.05), except for current age, ageat diagnosis, and weight centile (table 3). Thus,for these two distributions, the Mann-WhitneyU test (Wilcoxon) for two independent sam-ples20 was applied. For the other distributionswith equal variances, statistical significancecomparisons between the two groups of CFpatients were performed using the parametricStudent’s t test for two unpaired samples.20 Forqualitative distributions, the hypothesis of
association was tested with each of the twogroups of CF patients, considered as two inde-pendent samples in 2 × 2 crosstabs (1 degree offreedom), using both the Pearson chi-squarecalculation with continuity correction andFisher’s exact test.20 CoeYcients with a p valueless than 0.05 were considered to be statisti-cally significant. The SPSS® for Windows soft-ware (SPSS Inc, Chicago, IL) was used for allstatistical calculations.
ResultsWe have compared the clinical phenotypes of60 CF patients with the CFTR genotype 3272-26A>G/any mutation with those of 89 patientswho were F508del homozygotes, that is,presenting classical CF disease. Patients withthe 3272-26A>G mutation were from sixdiVerent countries, namely France, Spain,Greece, Germany, Portugal, and Belgium(table 1). Among these, 44 (73%) patients wereF508del compound heterozygotes (one ofthem died at the age of 59 years and anotherone received a lung transplant at the age of 34),four patients were heterozygotes for R1162X(two sib pairs) and two (also sibs) for E822X(table 1). Nine patients carried the followingmutations in the other CFTR allele (one ofeach): W1282X, 2869insG, L206W, N1303K,1717-1 G>A, G542X, 4218insT, W846X,P99L. Except for the P99L and L206W substi-tutions, all the mutations are known torepresent severe CF alleles. For one of thepatients carrying the 3272-26A>G mutationthe second mutation was not identified (table1). However, the clinical phenotype was clearlyCF, although mild, so the patient was includedin the study.
Most of the patients analysed (80%) had thehaplotype D (2,2) at the XV-2c/KM19 lociassociated with the 3272-26A>G mutation.For these, an extended haplotype analysis,when done, showed the same variants, beingthus consistent with the presence of the samemutant allele in all these patients. The3272-26A>G mutation was also found in alle-les with haplotypes B (1,2) and C (2,1) atXV-2c/KM19 in one French patient and inanother previously described Belgian patient,21
respectively. The other markers in the extendedhaplotype for the Belgian patient were all thesame as for D.
Four other patients, representing about 13%of patients analysed (one in Greece, one inGermany, and two in France) had the haplo-type A (1,1) at XV-2c/KM19 in linkagedisequilibrium with the 3272-26A>G muta-tion. An extended haplotype with other mark-ers was determined for the German and theGreek patients, diVering substantially fromthose present in the D allele.
CLINICAL PHENOTYPES
Table 2 shows the average values (mean andmedian) for the incidence of several clinicalparameters of the CF patients in this study. Dataare shown for a total of 60 patients with theCFTR genotype 3272-26A>G/any mutationand for 89 F508del/F508del patients used as thecontrol group. DiVerences between these clini-cal parameters in the two groups of patients weretested for significance (see Methods). Resultsfrom the statistical analysis, and reference to thetest applied in each case, are shown in table 3.Significant diVerences were found for thefollowing parameters: age at diagnosis (higher
for the 3272-26A>G group), FEV1 and FVC(also higher), incidence of lung colonisation withPa (lower), occurrence of pancreatic insuY-ciency (lower), weight centile or BMI (higher),and nasal polyposis (higher). Altogether theseresults indicate milder CF disease in patientscarrying the 3272-26A>G mutation on oneCFTR allele. For none of the other parameters(electrolyte concentration in sweat, colonisationwith pathogens other than Pa, occurrence ofmeconium ileus, height centile, and SK and CNscores) were diVerences found to be significant.
DiscussionPatients who have milder symptoms or atypicalCF often have one severe and one so called classV mutation.22 This class includes mutations thatleave residual levels of normal CFTR transcriptsand protein.23 Mild CF disease was generallyreported for the following class V (splicing)mutations: 3849+10 kb C>T,24 IVS8-5T,9 25 and2789+5G>A.26 However, the severity of the dis-ease resulting from these mutations was neverassessed for significance in comparison totypical CF, that is, presented by most F508delhomozygotes. Yet this information is importantfor genetic counselling, in particular for prenataldiagnosis. Here, we report that patients with the3272-26A>G mutation (and another CFTRmutation, mostly a severe one) exhibit signifi-cant diVerences for various major clinicalparameters in comparison to typical CF disease.
TWO INDEPENDENT ORIGINS FOR 3272-26A>G IN
EUROPE
Among the 29 patients included in this studywhose haplotypes at the XV-2c/KM19 loci weredetermined and from data on one patient
Table 3 Significance tests for comparisons of clinical features between cystic fibrosis patients with 3272-26A>G/anymutation and F508del/F508del genotypes
Parameter*Total Nostudied
Levene’s testp value
Equalvariances Test applied p value Significance
Age at diagnosis 143 7.2 × 10−9 No Mann-Whitney 7.8 × 10−6 †Sweat test 105 4.6 × 10−1 Yes Student’s t 5.9 × 10−2 NSFEV1, % predicted 110 5.2 × 10−1 Yes Student’s t 3.8 × 10−3 †FVC, % predicted 108 1.5 × 10−1 Yes Student’s t 1.8 × 10−3 †Lung colonisation with bacterial pathogens
published elsewhere,21 80% have the D haplo-type (2,2) in association with the 3272-26A>Gmutation.
However, haplotype data also suggest that the3272-26A>G change must have occurredthrough a second mutational event. Indeed, fourother patients (two French, one German, andone Greek) carry the 3272-26A>G mutation inassociation with the A haplotype at the XV-2c/KM19 loci. According to data available from theGerman patient (and also partially from theGreek patient) for other markers, this is a totallydiVerent allele, thus strongly suggesting a secondorigin for the 3272-26A>G mutation. Possiblythis happened in populations living in moreeastern regions of Europe, considering theorigin of two of these patients (one Greek andanother from former East Germany). Owing tothe lower frequency of allele A (13%) incomparison to D (80%) in association to the3272-26A>G mutation, the former may haveoccurred later. Haplotype data suggesting thefurther change/recombination of the D alleleafter occurrence of the splicing mutation alsosupport this hypothesis.
The A haplotype is associated with the TG11
repeat (v TG10 for the D haplotype). Since thelength of the TG repeat in intron 8 seems tocorrelate with the extent of exon 9 alternativesplicing, that is, the higher the number of TGrepeats, the lower the amount of exon 9+ tran-scripts produced,27 it might be expected thatthe patients with the A haplotype would have alower amount of exon 9+ transcripts and hencea more severe phenotype. Although only four ofthe 3272-26A>G patients analysed here hadthe A haplotype, no significant diVerence inclinical phenotype was found for these patientsin relation to those with haplotype D.
3272-26A>G CAUSES MILD CF
Patients carrying the 3272-26A>G mutationon one allele (and mostly a severe CFTR muta-tion on the other) were shown here to be diag-nosed later, to have better lung function, lowerincidence of lung colonisation with Pa, andmore often to have normal pancreatic functionand higher weight centiles or BMI thanpatients homozygous for F508del. Altogether,such results indicate milder CF disease in thesepatients than typical CF. Unexpectedly, nasalpolyposis occurs more frequently in thesepatients (about 37% as opposed to around10% in F508del homozygotes). It is plausiblethat nasal polyposis may be underdiagnosed inthis study (particularly in the group of F508delhomozygotes), since it has been reported tooccur generally in about 37% of CF patients28
and, in particular, in about 40% of F508delhomozygotes.29 The 5T allele, generally associ-ated with mild CF, was also detected withincreased frequency in subjects with sinopul-monary disease of ill defined aetiology.30 Thus,the incidence of nasal polyposis, which is gen-erally considered as a minor complication, doesnot disprove the fact that patients with3272-26A>G on one allele have milder CFthan F508del homozygotes.
This mutation has been previously shown6 tocreate an alternative acceptor splicing site in
intron 17a that competes with the normal onebut still allows some normal CFTR mRNA tobe produced. We postulate that the remainingnormal CFTR mRNA still existing in thesepatients lessens the severity of CF disease. Themolecular basis of this significantly milder CFphenotype must thus lie in the existence ofremaining CFTR mRNA that is still normallyprocessed, thus giving rise to functionalprotein.6 This, however, is not enough to com-pletely avoid lung disease. There are reports ofother mutations, both in the CFTR gene(reviewed by Kerem and Kerem31) and in othergenes such as â globin,32 indicating that reduc-tion in the normal protein levels generallycauses milder disease than mutations leadingto total absence of or non-functional protein.
The variability and organ involvement inpatients carrying this mutation may criticallydepend on the levels of CFTR mRNA (andprotein) still present. DiVerences between thenormal and the alternative splicing processescan result from diVerential expression ofsplicing factors which will thus act as modify-ing factors of CFTR expression.33 Indeed, somevariability was observed among patients in-cluded in this study. Some patients (withF508del on the other allele) were diagnosed atan early age, have Pa colonisation, reducedlung function, and are pancreatic insuYcient.Others have very diVerent clinical records. Onepatient (with F508del on the other allele) wasonly diagnosed at the age of 32 because of per-sistent cough during her first pregnancy.Another (also with F508del on the other allele)at the age of 37 had still not developed lungdisease, nor pancreatic insuYciency, and wasonly genotyped because of congenital bilateralabsence of the vas deferens (CBAVD).
Owing to the relatively high incidence of thismutation in Europe, it might have beenexpected that one or more homozygotes for thismutation would be found, as described forother mutations of comparable incidence.24 34
One possible explanation for our failure toidentify any homozygotes is that they may notbe CF patients. In light of the fact that just one3272-26A>G allele, in compound heterozy-gosity with a severe mutation (like F508del),causes a significant reduction in phenotypicseverity, it is plausible that two 3272-26A>Galleles would avoid CF totally.
Further characterisation of CFTR splicing inthese patients, namely by quantifying normalmessenger still present, is our current research,and it will be useful to estimate the minimumlevels of CFTR necessary to avoid CF lung dis-ease. It is hoped that these insights may alsoprovide clues for novel therapies, through themodulation of factors that enhance the normalversus the alternative splicing.
This work was supported by PRAXIS XXI P/SAU/55/96(Portugal), FIS 99/0654 and Instituto Catala de la Salut(Spain), Deutsche Forschungsgemeinschaft (Germany), andAssociation Française de Lutte contre la Mucoviscidose AFLM(France) research grants. SB and CMF were recipients of BPD/17059/98 and BD/11094/97 fellowships, respectively.
1 Kerem B, Rommens JM, Buchanan JA, Markiewicz D, CoxTK, Chakravarti A, Buchwald M, Tsui LC. Identificationof the cystic fibrosis gene: genetic analysis. Science1989;245:1073-80.
2 Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R,Grzelczak Z, Zielenski J, Lok S, Plavsic N, Chou JL. Iden-tification of the cystic fibrosis gene: cloning and characteri-zation of complementary DNA. Science 1989;245:1066-73.
3 Rommens JM, Iannuzzi MC, Kerem B, Drumm ML,Melmer G, Dean M, Rozmahel R, Cole JL, Kennedy D,Hidaka N, Zsiga M, Buchwald M, Riordan JR, Tsui LC,Collins FS. Identification of the cystic fibrosis gene: chromo-some walking and jumping. Science 1989;245:1059-65.
5 Welsh M, Tsui LC, Boat TF, Beaudet AL. Cystic fibrosis.In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. Themetabolic and molecular basis of inherited disease. 7th ed. NewYork: McGraw-Hill, 1995:3799-876.
6 Beck S, Penque D, Garcia S, Gomes A, Farinha C, Mata L,Gulbenkian S, Gil-Ferreira K, Duarte, Pacheco P, BarretoC, Lopes B, Cavaco J, Lavinha J, Amaral MD. Cystic fibro-sis patients with the 3272-26A→G mutation have mild dis-ease, leaky alternative mRNA splicing, and CFTR proteinat the cell membrane. Hum Mutat 1999;14:133-44.
7 Kanavakis E, Tzetis M, Antoniadi T, Trager-Synodinos J,Kattamis C, Doudounakis S, Adam G. Mild cystic fibrosisphenotype in patients with the 3272-26A>G mutation. JMed Genet 1995;32:406-7.
8 Bienvenu T, Beldjord C, Kaplan JC, Hubert D, Dusser D.Severe cystic fibrosis phenotype in a F508del/3272-26A>Gcompound heterozygote. J Med Genet 1995;32:919.
9 Chillon M, Casals T, Mercier B, Bassas L, Lissens W, SilberS, Romey MC, Ruiz-Romero J, Verlingue C, Claustres M.Mutations in the cystic fibrosis gene in patients withcongenital absence of the vas deferens. N Engl J Med 1995;332:1475-80.
10 Fanen P, Ghanem N, Vidaud M, Besmond C, Martin J,Costes B, Plassa F, Goossens M. Molecular characteriza-tion of cystic fibrosis: 16 novel mutations identified by
analysis of the whole cystic fibrosis conductance trans-membrane regulator (CFTR) coding regions and splice sitejunctions. Genomics 1992;13:770-6.
11 Ferrie RM, Schwarz MJ, Robertson NH, Vaudin S, SuperM, Malone G, Little S. Development, multiplexing, andapplication of ARMS tests for common mutations in theCFTR gene. Am J Hum Genet 1992;51:251-62.
12 Rommens J, Kerem BS, Greer W, Chang P, Tsui LC, Ray P.Rapid nonradioactive detection of the major cystic fibrosismutation. Am J Hum Genet 1990;46:395-6.
13 Eggerding FA, Iovannisci DM, Brinson E, Grossman P,Winn-Deen ES. Fluorescence-based oligonucleotide ligationassay for analysis of cystic fibrosis transmembrane conduct-ance regulator gene mutations. Hum Mutat 1995;5:153-65.
14 Morral N, Nunes V, Casals T, Chillon M, Gimenez J,Bertranpetit J, Estivill X. Microsatellite haplotypes forcystic fibrosis: mutation frameworks and evolutionary trac-ers. Hum Mol Genet 1993;2:1015-22.
15 Costes B, Girodon E, Ghanem N, Flori E, Jardin A, SoufirJC, Goossens M. Frequent occurrence of the CFTR intron8 (TG)n 5T allele in men with congenital bilateral absenceof the vas deferens. Eur J Hum Genet 1995;3:285-93.
16 Dörk T, Neumann T, Wulbrand U, Wulf B, Kalin N, MaassG, Krawczak M, Guillermit H, Ferec C, Horn G. Intra-and extragenic marker haplotypes of CFTR mutations incystic fibrosis families. Hum Genet 1992;88:417-25.
17 Kerem BS, Zielenski J, Markiewicz D, Bozon D, Gazit E,Yahav J, Kennedy D, Riordan JR, Collins FS, Rommens JM.Identification of mutations in regions corresponding to thetwo putative nucleotide (ATP)-binding folds of the cysticfibrosis gene. Proc Natl Acad Sci USA 1990;87:8447-51.
18 Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B.Changes in the normal maximal expiratory flow-volumecurve with growth and aging. Am Rev Respir Dis 1983;127:725-34.
19 Conway SP, Littlewood JM. Cystic fibrosis clinical scoringsystems. In: Dodge JA, Brock DJ, Widdicombe JH, eds.Cystic fibrosis. Current topics. New York: Wiley, 1996:339-58.
20 Sokal R, Rohlf F. Biometry. The principles and practice of sta-tistics in biological research. 2nd ed. New York: Freeman &Co, 1981.
21 Cuppens H, Teng H, Raeymaekers P, De Boeck C,Cassiman JJ. CFTR haplotype backgrounds on normal andmutant CFTR genes. Hum Mol Genet 1994;3:607-14.
22 Kerem B, Kerem E. The molecular basis for diseasevariability in cystic fibrosis. Eur J Hum Genet 1996;4:65-73.
23 Wilschanski M, Zielenski J, Markiewicz D, Tsui LC, CoreyM, Levison H, Durie PR. Correlation of sweat chlorideconcentration with classes of the cystic fibrosis transmem-brane conductance regulator gene mutations. J Pediatr1995;127:705-10.
24 Highsmith WE, Burch LH, Zhou Z, Olsen JC, Boat TE,Spock A, Gorvoy JD, Quittel L, Friedman KJ, SilvermanLM. A novel mutation in the cystic fibrosis gene in patientswith pulmonary disease but normal sweat chloride concen-trations. N Engl J Med 1994;331:974-80.
25 Jarvi K, Zielenski J, Wilschanski M, Durie P, Buckspan M,Tullis E, Markiewicz D, Tsui LC. Cystic fibrosistransmembrane conductance regulator and obstructiveazoospermia. Lancet 1995;345:1578.
26 Highsmith WE Jr, Burch LH, Zhou Z, Olsen JC, Strong TV,Smith T, Friedman KJ, Silverman LM, Boucher RC, Col-lins FS, Knowles MR. Identification of a splice sitemutation (2789 +5 G>A) associated with small amounts ofnormal CFTR mRNA and mild cystic fibrosis. Hum Mutat1997;9:332-8.
27 Cuppens H, Lin W, Jaspers M, Costes B, Teng H, Vankeer-berghen A, Jorissen M, Droogmans G, Reynaert I,Goossens M, Nilius B, Cassiman JJ. Polyvariant mutantcystic fibrosis transmembrane conductance regulatorgenes. The polymorphic (TG)m locus explains the partialpenetrance of the T5 polymorphism as a disease mutation.J Clin Invest 1998;101:487-96.
28 Hadfield PJ, Rowe-Jones JM, Mackay IS. The prevalence ofnasal polyps in adults with cystic fibrosis. Clin Otolaryngol2000;25:19-22.
29 Coste A, Gilain L, Roger G, Sebbagh G, Lenoir G, ManachY, Peynegre R. Endoscopic and CT-scan evaluation of rhi-nosinusitis in cystic fibrosis. Rhinology 1995;33:152-6.
30 Friedman KJ, Heim RA, Knowles MR, Silverman LM.Rapid characterization of the variable length polythymidinetract in the cystic fibrosis (CFTR) gene: association of the5T allele with selected CFTR mutations and its incidencein atypical sinopulmonary disease. Hum Mutat 1997;10:108-15.
31 Kerem E, Kerem B. Genotype-phenotype correlations incystic fibrosis. Pediatr Pulmonol 1996;22:387-95.
32 Faustino P, Lavinha J, Marini MG, Moi P. Beta-thalassemiamutation at -90C→T impairs the interaction of the proxi-mal CACCC box with both erythroid and nonerythroidfactors. Blood 1996;88:3248-9.
33 Pagani F, Buratti E, Stuani C, Romano M, Zuccato E, Nik-sic M, Giglio L, Faraguna D, Baralle FE. Splicing factorsinduce cystic fibrosis transmembrane regulator exon 9skipping through a nonevolutionary conserved intronic ele-ment. J Biol Chem 2000;275:21041-7.
34 Casals T, Pacheco P, Barreto C, Gimenez J, Ramos MD,Pereira S, Pinheiro JA, Cobos N, Curvelo A, Vazquez C,Rocha H, Seculi JL, Perez E, Dapena J, Carrilho E, DuarteA, Palacio AM, Nunes V, Lavinha J, Estivill X. Missensemutation R1066C in the second transmembrane domain ofCFTR causes a severe cystic fibrosis phenotype: study of 19heterozygous and 2 homozygous patients. Hum Mutat1997;10:387-92.
+ We report here clinical phenotypes of 60cystic fibrosis (CF) patients from sixEuropean countries with the 3272-26A>G mutation on one allele of thecystic fibrosis transmembrane conduct-ance regulator (CFTR) gene, and anothermutation (mostly F508del) on the otherallele. These were compared with theclinical phenotypes of F508del homo-zygous patients (n=89) from the samecentres and matched for age and sex asexactly as possible.
+ Clinical phenotypes of CF patients with3272-26A>G were found to be signifi-cantly milder (p<0.05) than those ofF508del homozygotes, namely older ageat diagnosis, better pulmonary function,lower incidence of colonisation withPseudomonas aeruginosa, lower occur-rence of pancreatic insuYciency, andhigher weight centile or body mass index.Altogether, this information is importantfor genetic counselling, in particular forprenatal diagnosis.
+ Unexpectedly, incidence of nasal polypo-sis was found to be significantly higher in3272-26A>G patients.
+ The heterogeneity of CFTR haplotypesindicates that this mutation, which isspread all over Europe, must have evolvedfrom more than one mutational event. Itis shown for the first time that CFpatients with one mutation causing alter-native splicing (and mostly another se-vere mutation) have significantly milderdisease than patients with two severemutations. We postulate that the remain-ing normal CFTR mRNA still existing inthese patients alleviates the severity of CFdisease.
Maternal gene eVect in neurofibromatosis 2: factor artefact?
Michael E Baser, J M Friedman, D Gareth R Evans
EDITOR—Neurofibromatosis 2 (NF2) is a rareautosomal dominant disease that is character-ised by benign nervous system tumours, skinlesions, and ocular abnormalities.1–3 Two stud-ies have found that NF2 patients with a familyhistory of the disease and with maternal inher-itance have more severe disease than inheritedcases with paternal inheritance. Kanter et al4
noted that patients with maternal inheritancehad an earlier age at onset and Evans et al5
found that patients with maternal inheritancehad both an earlier age at onset and an earlierage at death. In both studies, the mean age atonset was 18 years with maternal inheritanceand 24 years with paternal inheritance.
These results require confirmation for sev-eral reasons. First, Parry et al6 found identicalmean ages at onset (22.8 years) in symptomaticNF2 patients with paternal or maternal inher-itance. Second, these studies were based onrelatively small numbers of patients. Kanter etal4 studied 38 inherited cases, Evans et al5 stud-ied 56 inherited cases, and Parry et al6 studied36 inherited cases; 66% of the patients inKanter et al4 and 64% in Evans et al,5 but only39% in Parry et al,6 had maternal inheritance.Third, in examining the eVect of maternalinheritance on disease severity, these studiesreported only age at onset (and age at death inEvans et al5) as indices of disease severity. Inaddition to age at onset or age at diagnosis,NF2 disease severity is defined by the numberof non-vestibular schwannoma cerebral tu-mours and of spinal tumours.6 Fourth, none ofthe studies examined potential confoundingfactors (such as type of treatment centre andconstitutional NF2 mutation type) that canaVect age at onset, age at diagnosis, ormortality. NF2 patients who are treated atnon-specialty centres have higher odds of deaththan those who are treated at specialty centres,and NF2 patients with missense mutationshave lower odds of death than those with non-sense or frameshift mutations.7 Genotype-phenotype correlation studies have found thatNF2 patients with constitutional NF2 missense
mutations or large deletions generally havemild disease, those with splice site mutationshave variable disease severity, and those withnonsense or frameshift mutations have severedisease.8–11
We reassessed the question of maternal geneeVect in NF2 with a larger number of patientsand with consideration of potential confound-ers. We used data from the United KingdomNF2 Registry, based in the Department ofMedical Genetics, St Mary’s Hospital, Man-chester. Patients are ascertained by contactingneurosurgeons, neurologists, otolaryngolo-gists, paediatricians, dermatologists, and ge-neticists throughout the United Kingdom,augmented in the North West Region by theRegional Cancer Registry. As of 15 September2000, the registry had data on 140 inheritedcases (85 with maternal inheritance and 55with paternal inheritance, including the 56inherited cases previously reported by Evans etal5). All patients met the Manchester clinicaldiagnostic criteria for NF25 or had identifiedconstitutional NF2 mutations.
For univariate analyses, the two tailed t testwas used for age variables, the ÷2 test fordiscrete variables (for example, distribution ofmutation types), and life tables for mortalityanalysis. Multivariate analyses were used toexamine the independent eVects of covariates(linear regression for age variables and the Coxproportional hazards model for mortalityanalysis). Because age at onset and age at diag-nosis are highly correlated (in the presentstudy, r2 = 0.78, p<0.001), separate multiplelinear regressions were done with each agevariable as the outcome. p values <0.05 wereconsidered to be statistically significant.
In univariate comparisons of patients withmaternal and paternal inheritance, patientswith maternal inheritance were treated slightlymore frequently at non-specialty centres (75%versus 62%, p=0.10) and had a slightly lowermean age at onset (23.1 years versus 25.5years, p=0.30). The distribution of constitu-tional NF2 mutation types varied significantly
J Med Genet2001;38:783–784
2257 Fox Hills Drive,Los Angeles, CA 90064,USAM E Baser
Department ofMedical Genetics,University of BritishColumbia, Vancouver,CanadaJ M Friedman
Department ofMedical Genetics,St Mary’s Hospital,Manchester, UKD G R Evans
by mode of inheritance (p=0.03); in particular,patients with maternal inheritance had a lowerproportion of missense mutations (7% versus24%). Patients with maternal inheritance hadsignificantly higher mortality (p=0.02).
In the multiple linear regression analyses,age at onset and age at diagnosis did not varysignificantly with mode of inheritance (table1). Patients with missense mutations and largedeletions had significantly higher ages at onsetand diagnosis than patients with nonsense orframeshift mutations. Patients treated at spe-cialty centres had lower age at onset andsignificantly lower age at diagnosis thanpatients treated at non-specialty centres. In themultivariate Cox proportional hazards model,the odds of death did not vary significantly withmode of inheritance or type of constitutionalNF2 mutation (table 2). The odds of deathvaried significantly with type of treatment cen-tre, age at diagnosis, and number of intracra-nial meningiomas.
The results of this study, the largest to date,are consistent with previous genotype-phenotype correlation studies in NF2,8–11 butdo not support maternal inheritance as anindependent correlate of disease severity inNF2. The eVects of constitutional NF2 muta-tion type and treatment in specialty centrescould not be examined in the earlier, more lim-ited data of Kanter et al,4 Evans et al,5 and Parry
et al.6 Potential confounders should be rou-tinely examined when outcomes that can beaVected by multiple factors are being studied.
2 Trofatter JA, MacCollin MM, Rutter JL, Murrell JR, DuyaoMP, Parry DM, Eldridge R, Kley N, Menon AG, PulaskiK, Haase VH, Ambrose CM, Munroe D, Bove C, HainesJL, Martuza RL, MacDonald ME, Seizinger BR, ShortMP, Buckler AJ, Gusella JF. A novel moesin-, ezrin-,radixin-like gene is a candidate for the neurofibromatosis 2tumor suppressor. Cell 1993;72:791-800.
3 Rouleau GA, Merel P, Lutchman M, Sanson M, Zucman J,Marineau C, Hoang-Xuan K, Demczuk S, Desmaze C,Plougastel B, Pulst SM, Lenoir G, Bijlsma E, Fashold R,Dumanski J, de Jong P, Parry D, Eldridge R, Aurias A,Delattre O, Thomas G. Alteration in a new gene encodinga putative membrane-organizing protein causes neuro-fibromatosis type 2. Nature 1993;363:515-21.
4 Kanter WR, Eldridge R, Fabricant R, Allen JC, Koerber T.Central neurofibromatosis with bilateral acoustic neuroma:genetic, clinical and biochemical distinctions from peri-pheral neurofibromatosis. Neurology 1980;30:851-9.
5 Evans DGR, Huson SM, Donnai D, Neary W, Blair V, TeareD, Newton V, Strachan T, Ramsden R, Harris R. A geneticstudy of type 2 neurofibromatosis in the United Kingdom.I. Prevalence, mutation rate, fitness, and confirmation ofmaternal transmission eVect on severity. J Med Genet 1992;29:841-6.
6 Parry DM, Eldridge R, Kaiser-Kupfer MI, Bouzas EA,Pikus A, Patronas N. Neurofibromatosis 2 (NF2): clinicalcharacteristics of 63 aVected individuals and clinicalevidence for heterogeneity. Am J Med Genet 1994;52:450-61.
7 Baser ME, Friedman JM, Evans DGR. Predictors ofmortality in neurofibromatosis 2. Am J Hum Genet1999;65(suppl 4):A61.
8 Parry DM, MacCollin MM, Kaiser-Kupfer MI, Pulaski K,Nicholson HS, Bolesta M, Eldridge R, Gusella JF.Germ-line mutations in the neurofibromatosis 2 gene: cor-relations with disease severity and retinal abnormalities.Am J Hum Genet 1996;59:529-39.
9 Ruttledge MH, Andermann AA, Phelan CM, Claudio JO,Han FY, Chretien N, Rangaratnam S, MacCollin M, ShortP, Parry D, Michels V, Riccardi VM, Weksberg R, KitamuraK, Bradburn JM, Hall BD, Propping P, Rouleau GA. Typeof mutation in the neurofibromatosis type 2 gene (NF2)frequently determines severity of disease. Am J Hum Genet1996;59:331-42.
10 Kluwe L, Bayer S, Baser ME, et al. Identification of NF2germ-line mutations and comparison with neurofibromato-sis 2 phenotypes. Hum Genet 1996;98:534-8.
11 Evans DGR, Trueman L, Wallace A, Collins S, Strachan T.Genotype/phenotype correlations in type 2 neurofibroma-tosis (NF2): evidence for more severe disease associatedwith truncating mutations. J Med Genet 1998;35:450-5.
Familial medullary thyroid carcinoma andprominent corneal nerves associated with thegermline V804M and V778I mutations on thesame allele of RET
Lidia Kasprzak, Serge Nolet, Louis Gaboury, Carles Pavia, Carles Villabona,Francisca Rivera-Fillat, Josep Oriola, William D Foulkes
EDITOR—Germline mutations in the RETproto-oncogene are associated with multipleendocrine neoplasia (MEN) type 2A, 2B, andfamilial medullary thyroid carcinoma(FMTC). MEN 2A is characterised primarilyby medullary thyroid carcinoma (MTC),parathyroid hyperplasia, and phaeochromocy-toma. MEN 2B has additional stigmata includ-ing a marfanoid habitus, mucosal neuromas,and corneal nerve thickening. In FMTC, the
only lesion present is MTC. Although charac-teristic of MEN 2B, prominent corneal nerveshave been noted in subjects with MEN 2A.
Case reportsWe have studied a family presenting with MTCand corneal nerve thickening (CNT), but noevidence of MEN 2 related disease outside thethyroid gland (fig 1). The proband (III.1) wasfound to have prominent corneal nerves on
Table 2 Results of multivariate Cox proportional hazards model
Covariate Odds of death (95% CI) p
Type of treatment centre 0.05 (0.00, 0.45) 0.008(specialty compared to non-specialty)Age at diagnosis (per year) 0.84 (0.78, 0.90) <0.001
routine ophthalmological examination in hislate 30s. No further action was taken and hepresented with a palpable thyroid mass at theage of 53.
MEN 2B was suspected and he was referredfor laboratory screening tests for MTC andphaeochromocytoma. Serum calcitonin wasraised at 4090 ng/l (normal range <60 ng/l).Serum calcium and phosphorus were normal.Twenty four hour urinary total catecholamine,metanephrine, and vanillylmandelic acid levelswere within normal limits. A thoracic andabdominal CT scan performed before surgerywas normal and scanning of the neck confirmedthe presence of the right thyroid mass. Thepatient underwent a total thyroidectomy andpathological examination indicated bilateralMTC with extension beyond the capsule. Hisendocrinologist referred him for a genetic evalu-ation with a diagnosis of MEN 2B. However,other than CNT, he and his relatives did nothave any of the other stigmata of this subtype.Family history was notable for thyroid carci-noma in the maternal cousin (III.4) and her son(IV.2). Interestingly, his mother (II.1), whoappeared to be an obligate carrier, remainedunaVected and died at 93 years of age relatedcauses. The proband was encouraged to informhis family in Spain about the option of geneticcounselling through the centre in Barcelona.Subsequently, IV.2 and all of his children wereassessed and oVered genetic testing. Patient IV.2first noticed a right cervical enlargement at theage of 35 and later was diagnosed with multiplenodules and subjected to total thyroidectomy.The histological examination showed a multi-centric medullary thyroid carcinoma with bilat-eral thyroid involvement. Lymph node and bone
metastases were identified. Catecholamines andmetanephrines were within normal limits. Hewas also found to have prominent cornealnerves. His children were screened for serumcalcitonin. One of the three oVspring (V.2) hadhad raised basal and pentagastrin stimulatedlevels (CT 15 pg/ml, basal and 23 pg/ml, fiveminutes).
The proband’s genomic DNA was screenedfor mutations in exons 10, 11, and 13-16 of theRET gene. Direct sequencing identified twoheterozygous missense mutations. Analysis ofexon 13 showed a single base pair substitution ofG to A at codon 778 resulting in a change fromvaline to isoleucine in the predicted amino acidsequence. Sequence analysis of exon 14 showeda heterozygous single base pair substitution of Gto A at codon 804 resulting in a change fromvaline to methionine. Only exons 13 and 14 wereanalysed in the proband’s mother, sister, andniece. Both mutations were subsequently identi-fied in II.1 (an obligate carrier) and neither wasfound in III.2 or IV.1. Testing in Spain was per-formed concurrently on the aVected subjectIV.2, who was found to be positive for bothV778I and V804M. All of his children weretested and one had the same genotype as herfather while two were negative for both exon 13and 14 missense mutations. Of note, thepreviously undescribed variant in exon 13 waspresent only in subjects carrying the V804Mmutation. No mutation was found in RETcodon 778 in four unaVected (V804M negative)family members, or in 73 and 20 controls fromMontreal and Barcelona, respectively. Further-more, we have amplified the proband’s DNA byPCR, encompassing exon 13/intron 13/exon 14of RET, by using an upstream primer localised at
Figure 1 A truncated pedigree of the family is shown. Filled symbols indicate aVected subjects. The letters below thesymbols refer to the site of cancer, the numbers following the abbreviations indicate age at diagnosis. MTC, medullarythyroid cancer; PSU, primary site unknown; d, age at death; CNT, corneal nerve thickening.
I
II
III
IV
V
1 2PSUd 56
2PSUd 50s
1MTC 53
CNT
53 4MTC 26d 46
2
2MTC 35CNT
1
1 2 3
1d 93
3 42
J Med Genet2001;38:000–000
Program in CancerGenetics, Departmentsof Oncology and HumanGenetics, McGillUniversity, Montreal,Quebec, CanadaL KasprzakW D Foulkes
Département dePathologie, CentreHospitalier del’Université deMontréal, Montréal,Quebec, CanadaS NoletL Gaboury
Secció deEndocrinologia,Hospital Universitari deSant Joan de Deu,Barcelona, SpainC Villabona
Serveid’Endocrinologia, CSUBellvitge, HospitalPrincipes de España,Barcelona, SpainF Rivera-FillatJ Oriola
Department ofMedicine, ResearchInstitute of the McGillUniversity HealthCentre, MontrealGeneral Hospital, RoomL10-116, 1650 CedarAvenue, Montreal,Quebec H3G 1A4,CanadaW D Foulkes
the end of intron 12 and a downstream primerlocalised at the beginning of intron 14. The PCRproduct is 1497 bp long. Since the G to A atposition 778 destroys a single MaeII restrictionsite (ACGT) on the PCR product, only the nor-mal homozygous allele is cut by the MaeIIenzyme. After an overnight incubation withMaeII to assure complete digestion, we per-formed a second round PCR on the MaeIIdigestion product. Only an allele with the A(rather than the G) nucleotide in codon 778 canbe amplified since this amplified PCR product isleft intact by MaeII. Another MaeII digestion onthis second PCR product confirmed the absenceof a MaeII site. Restriction analysis with NlaIII(CATG), which cuts the mutated allele (G to A)at codon 804, indicated that the 778I variantand the 804M mutation are on the same allele.Moreover, we performed direct sequencing ofthe second PCR amplification product and wewere able to confirm the presence of the 778Ivariant on the same 804M mutated allele.
DiscussionThe point mutations in exon 14 leading to asubstitution of valine 804 by either leucine ormethionine have been reported in several fami-lies with FMTC.1–3 The prevalence of this par-ticular alteration in codon 804 seems to bequite low in FMTC families (frequency of 0.03according to Hansford and Mulligan4) andthere have been only three reports of the germ-line V804M in apparently sporadic MTCcases.5–7 Nilsson et al8 showed that the V804Lmutation is not exclusively associated withfamilial MTC as it has been identified in theMEN 2A pedigree. It appears that certaingermline sequence variants identified in exons13 and 14 are seen at a higher frequency inpatients with sporadic MTC as compared withthat of the general population. Two of thereported polymorphic variants (L769L in exon13 and S836S in exon 14) have beenimplicated as having a potentially promotingeVect in the aetiology of MTC.9 10 The restric-tion of the V778I mutation to V804M carrierswas noted and we have confirmed that V778I isin cis with the disease allele in the proband. Thenovel V778I germline alteration may be a lowpenetrance allele contributing to the develop-ment of the CNT in this pedigree. Notwith-standing, codon 804 mutations have beeninitially associated with a later onset and lessaggressive disease phenotype suggesting lowerpenetrance7; however, there are several reportsindicating a quite variable clinical pictureincluding anticipation.3 11 The specific V778Isequence variant may be responsible forsignificantly earlier age of onset of MTC andmore aggressive disease course in this family.Even though the replacement of valine byisoleucine in the predicted amino acid se-quence results in a rather benign change(hydrophobic, branched chain to hydrophobic,branched chain amino acid) at the level of pri-mary protein structure, it is not clear how itmight alter the tertiary conformation. In addi-tion, it should be noted that both exon 13 and14 mutations aVect the intracellular domain ofthe receptor tyrosine kinase encoded by the
RET proto-oncogene. Miyauchi et al12 reporteda patient with a MEN 2B phenotype owing toa germline V804M mutation and a novelY806C mutation occurring on the same alleleof the RET gene. The authors speculated thatthe presence of the apparently non-pathogenicY806C mutation in the cis position might haveresulted in a structural change of the RET pro-tein potentiating the significance of V804Malone. This particular case presented at the ageof 23 with bilateral MTC with multiple fociand lymph node metastasis. Alternatively,Y806C could represent a low penetrance vari-ant which aVects the phenotype because of anadditive eVect in the presence of the secondmutation. Subsequent comparison of thebiological activity of the V804M/Y806C com-bination with that of the MEN 2B exclusivemutations M918T or A883F suggests similarbiological properties of these three mutantproteins.13 Furthermore, the transforming ac-tivity of RET with the V804/Y806C compoundwas apparently eight to 13 times that of RETwith each mutation analysed separately. Fur-ther studies of the RET alleles in this and otherkindreds may identify a unique eVect of theknown disease causing mutations in thepresence of rare variants on the same chromo-some 10, resulting in a diVerent phenotypethan observed in the presence of the mutationalone. Prominent corneal nerves which can bevisualised by slit lamp examination are presentin a large proportion of the MEN 2B patients14
and are less frequently observed in those withMEN 2A.15 As there are few such cases, it is notclear whether CNT in association with MEN2A can be correlated with specific mutations inthe RET gene. The significance of CNT wasoverlooked in our proband whose aVected sta-tus might have been established earlier ifgenetic consultation had been considered bythe ophthalmologist.
We thank Dr W Gregory for referring the proband to theHereditary Cancer Clinic at McGill University and the familyfor their continued cooperation.
1 Bolino A, SchuVenecker I, Luo Y, Seri M, Silengo M, ToccoT, Chabrier G, Houdent C, Murat A, Schlumberger M,Tourniaire J, Lenoir GM, Romeo G. RET mutations inexons 13 and 14 of FMTC patients. Oncogene1995;10:2415-19.
2 Fink M, Weinhusel A, Niederle B, Haas OA. Distinctionbetween sporadic and hereditary medullary thyroid carci-noma (MTC) by mutation analysis of the RET proto-oncogene. Study Group Multiple Endocrine NeoplasiaAustria (SMENA). Int J Cancer 1996;69:312-16.
3 Fattoruso O, Quadro L, Libroia A, Verga U, Lupoli G, Cas-cone E, Colantuoni V. A GTG to ATG novel point muta-tion at codon 804 in exon 14 of the RET proto-oncogene intwo families aVected by familial medullary thyroidcarcinoma. Hum Mutat 1998;suppl 1:S167-71.
4 Hansford JR, Mulligan LM. Multiple endocrine neoplasiatype 2 and RET: from neoplasia to neurogenesis. J MedGenet 2000;37:817-27.
+ A family with MTC and CNT wasassessed through a 53 year old aVectedsubject.
+ RET exon 13 and 14 mutations in cis(V778I and V804M) were identified infour family members.
+ A novel variant V778I was found, poten-tially aVecting the phenotype through anadditive eVect in the presence of V804M.
5 Scurini C, Quadro L, Fattoruso O, Verga U, Libroia A,Lupoli G, Cascone E, Marzano L, Paracchi S, Busnardo B,Girelli ME, Bellastella A, Colantuoni V. Germline andsomatic mutations of the RET proto-oncogene in appar-ently sporadic medullary thyroid carcinomas. Mol CellEndocrinol 1998;137:51-7.
6 Uchino S, Noguchi S, Adachi M, Sato M, Yamashita H,Watanabe S, Murakami T, Toda M, Murakami N, Yamas-hita H. Novel point mutations and allele loss at the RETlocus in sporadic medullary thyroid carcinomas. Jpn J Can-cer Res 1998;89:411-18.
7 Shannon KE, Gimm O, Hinze R, Dralle H, Eng C.Germline V804M mutation in the RET proto-oncogene intwo apparently sporadic cases of MTC presenting in theseventh decade of life. J Endocr Genet 1999;1:39-46.
8 Nilsson O, Tisell LE, Jansson S, Ahlman H, Gimm O, EngC. Adrenal and extra-adrenal pheochromocytomas in afamily with germline RET V804L mutation. JAMA1999;281:1587-8.
9 Wiench M, Wygoda Z, Gubala E, Wloch J, Lisowska K,Krassowski J, Scieglinska D, Fiszer-Kierzkowska A, LangeD, Kula D, Zeman M, Roskosz J, Kukulska A, Krawczyk Z,Jarzab B. Estimation of risk of inherited medullary thyroidcarcinoma in apparent sporadic patients. J Clin Oncol2001;19:1374-80.
10 Gimm O, Neuberg DS, Marsh DJ, Dahia PL, Hoang-Vu C,Raue F, Hinze R, Dralle H, Eng C. Over-representation of
a germline RET sequence variant in patients with sporadicmedullary thyroid carcinoma and somatic RET codon 918mutation. Oncogene 1999;18:1369-1373.
11 Feldman GL, Edmonds MW, Ainsworth PJ, SchuVeneckerI, Lenoir GM, Saxe AW, Talpos GB, Roberson J, PetrucelliN, Jackson CE. Variable expressivity of familial medullarythyroid carcinoma (FMTC) due to a RET V804M(GTG→ATG) mutation. Surgery 2000;128:93-8.
12 Miyauchi A, Futami H, Hai N, Yokozawa T, Kuma K, AokiN, Kosugi S, Sugano K, Yamaguchi K. Two germline mis-sense mutations at codons 804 and 806 of the RET proto-oncogene in the same allele in a patient with multipleendocrine neoplasia type 2B without codon 918 mutation.Jpn J Cancer Res 1999;90:1-5.
13 Iwashita T, Murakami H, Kurokawa K, Kawai K, MiyauchiA, Futami H, Qiao S, Ichihara M, Takahashi M. A two-hitmodel for development of multiple endocrine neoplasiatype 2B by RET mutations. Biochem Biophys Res Commun2000;268:804-8.
14 Vasen HF, van der FM, Raue F, Kruseman AN, Koppe-schaar HP, Pieters G, Seif FJ, Blum WF, Lips CJ. Thenatural course of multiple endocrine neoplasia type IIb. Astudy of 18 cases. Arch Intern Med 1992;152:1250-2.
15 Kinoshita S, Tanaka F, Ohashi Y, Ikeda M, Takai S. Incidenceof prominent corneal nerves in multiple endocrine neoplasiatype 2A. Am J Ophthalmol 1991;111:307-11.
Comprehension of cancer risk one and 12 monthsafter predictive genetic testing for hereditarynon-polyposis colorectal cancer
Katja Aktan-Collan, Ari Haukkala, Jukka-Pekka Mecklin, Antti Uutela, Helena Kääriäinen
EDITOR—The main purpose of oVering predic-tive genetic testing for hereditary cancer is toreduce unnecessary worry among those with alow risk of cancer (mutation negative) and torecognise those with a high risk (mutationpositive), so as to promote preventive meas-ures.1 2 Ideally, those shown to be at high riskwould understand this, would learn to live withthe knowledge, and, most importantly, wouldattend cancer surveillance programmes regu-larly. Those at low risk would feel relieved andno physically and emotionally uncomfortablesurveillance would be needed. This resultwould be seriously hampered if those tested didnot fully understand the meaning of the testresults, which could lead to unnecessary worryor failure to adhere to surveillance. At present,most predictive tests are performed in carefullyorganised settings, which include comprehen-sive pre- and post-test counselling that wouldbe expected to minimise the risk of misunder-standing the test result. However, commercialtests predicting cancer are already available,3
and this has raised concerns about predictivetesting with minimal counselling or even with-out personal contact with a health care profes-sional.1 4
The impact of genetic counselling on riskperception and impact of risk perception ongenetic testing intentions has been studied pre-viously.5 However, this is the first report on thecomprehension of test results (perception ofcancer risk) after predictive genetic testing forcancer, in this case hereditary non-polyposiscolorectal cancer (HNPCC), which is the mostcommon form of hereditary colon cancer.
HNPCC is an autosomal dominant disease ofadulthood with an 80-90% lifetime risk forcolorectal cancer and a lesser risk of extraco-lonic cancers, the most common of which areendometrial and gastric cancer.6 Hereditarycolorectal cancer diVers essentially from otherhereditary cancers, of which hereditary breastcancer is a good example. Firstly, in HNPCC,the life long risk of developing colorectal canceramong mutation positive subjects is uniformlyvery high (80-90%),7 8 while in breast cancerthe corresponding risk varies between 40 and80%.9–11 Secondly, for HNPCC, in contrast tobreast cancer, clinical surveillance among thoseat high risk has been shown to reduce mortalityfrom colorectal cancer substantially.12 13 Predic-tive genetic testing is now possible in all thefamilies with HNPCC in which the predispos-ing germline mutation is known. Previousstudies among first degree relatives of patientswith colorectal cancer have suggested a greatinterest in possible testing and intention tolearn the results.14 15 Reports on the actualoVering of predictive genetic tests to HNPCCfamilies have shown that the acceptance rateshave varied greatly (14-81%).16–19 This studydescribes how the members of HNPCCfamilies comprehended their predictive testresults in terms of their risk of developing colo-rectal cancer and discusses what may haveinfluenced this.
MethodsDuring 1995-1997, we oVered counsellingabout predictive genetic testing to adults at50% risk in 36 HNPCC families in whichmutations in the MLH1 gene had previously
J Med Genet2001;38:787–792
Department of MedicalGenetics, HaartmanInstitute, University ofHelsinki, Helsinki,FinlandK Aktan-Collan
Department of MedicalGenetics, FamilyFederation of Finland,POB 849, FIN-00101Helsinki, FinlandK Aktan-CollanH Kääriäinen
Department of SocialPsychology, Universityof Helsinki, Helsinki,FinlandA Haukkala
Department of Surgery,Jyväskylä CentralHospital, Jyväskylä,FinlandJ-P Mecklin
Department ofEpidemiology andHealth Promotion,National Public HealthInstitute, Helsinki,FinlandA Uutela
been characterised.20 21 The counselling andtesting procedure has been described in detailelsewhere.17 22 Briefly, all known eligible (aged18 or older and without a diagnosis of cancer)members of these HNPCC families wereinformed about the study by letter. Those whoconsented to participate in the study wereinvited to an individual pre-test counsellingsession, which followed a uniform scheme,comprising taking the family history and givinginformation about HNPCC, its mode of inher-itance, the gene defect, the nature and the riskof colon cancer, the risk of other cancers, andthe methods available for early detection oftumours. Early in 1995, when we started thecounselling, no data on the risk of developingcolorectal cancer were available for mutationpositive HNPCC family members. However,most (32/36) of the families were high riskfamilies fulfilling the Amsterdam criteria,including verified colorectal cancer in at leastthree relatives (one of whom was a first degreerelative of the other two) in at least two succes-sive generations, and at least one of the caseshad been diagnosed before the age of 50years.23 Therefore, the risk of colorectal cancerwas estimated to be very high, close to 100%.This was communicated to the counsellees atthe pre-test session. Furthermore, the benefitsand disadvantages of a predictive gene test werethoroughly discussed.
After a two week period for reflection, coun-sellees were contacted by telephone and askedif they wanted the test. Those who chose totake the test signed a consent form anddonated a blood sample. Those who declinedthe test and remained at 50% risk wereencouraged to adhere to the clinical surveil-lance, comprising colonoscopy every threeyears and gynaecological examinations yearlyfor females over 35 years old.
Those tested were invited, preferably with anaccompanying person, to a post-test counsel-ling session at which the test result and itsimplications were discussed. For those whohad the mutation, the high risk of colorectalcancer (close to 100%) was reiterated andclinical surveillance was organised. Subjectswho did not have the mutation were remindedof the general risk of cancer, to prevent anyfalse reassurance. The result of the mutationanalysis and, accordingly, surveillance recom-mendations were also given in written form.
Of the eligible subjects (n=446), 90%(n=401) consented to participate in the study
and 85% (n=381) returned a baseline ques-tionnaire I.17 The educational counsellingsession was attended by 347 subjects, of whom333 (96%, 75% of the total population) optedfor a predictive genetic test. Seven subjectsrefused to fill in any further questionnaires. Afollow up questionnaire was sent at one monthto 326 subjects, of whom 299 (92%) replied.Another follow up questionnaire was sent atone year to these 299 subjects, of whom 271(91%) filled in this final form. Thus, the studysample consisted of those 271 subjects whoattended both counselling sessions, and com-pleted the pre- and post-test questionnaires. Ofthe subjects, 68% attended the pre-test sessionconducted by a nurse specifically trained forpre-test counselling and 32% by a physicianspecialising in medical genetics (KA-C). Fiftyseven percent had a post-test counsellingsession conducted by the physician (KA-C)and the rest were counselled by a gastroentero-logical surgeon (J-PM). Both pre- and post-testsessions were standardised as far as possible, inthat all counsellors followed a similar struc-tured protocol including similar, previouslyagreed, risk counselling.
Of the study subjects (n=271), 57% werewomen, 72% lived with a spouse or partner,73% had children, 76% were employed, and62% had an education higher than primarylevel.22 The participants were aged 19-77 years(mean 43 years) and 68% had a previoushistory of clinical cancer surveillance becauseof their high risk status. Thirty one percent(n=84) were mutation positive. The diVerencesbetween the groups defined by mutation statusare presented in table 1. The subjects lost tofollow up (n=62) who did not completequestionnaires after the post-test counsellingsession did not diVer significantly from thestudy subjects in any of the variables describedhere.
The study is based on questionnaires, whichwere filled in three times during the procedure:before the first counselling session (baselinemeasurement) and one month and one yearafter the test disclosure session. Exceptionally,the anxiety scale was filled in at the test disclo-sure session soon after the test result had beencommunicated. Understanding of the testresult was assessed in both follow up question-naires by two questions. (1) “What was yourtest result?” (1 = I was found to have the muta-tion predisposing to colorectal cancer, 2 = Iwas found not to have the mutation). (2)“What does your risk of developing colorectalcancer look like after testing? In this connec-tion, the risk refers to what the cancer riskwould be without regular cancer surveillanceaimed at prevention of cancer.” (1 = the risk ishigh, close to 100%, 2 = the risk is approxi-mately 50%, 3 = the risk is quite low,corresponding to that of the general popula-tion.) The alternatives chosen were based onthe information given in the pre-and post-testcounselling. Accordingly, the correct optionswere 3 for the mutation negative subjects and 1for the mutation positive subjects. In the analy-sis, those who chose the correct option werelabelled “understanding” the result and those
Table 1 Associations of baseline demographic variables with groups defined by mutationstatus
Mutationnegative(n=187)
Mutationpositive(n=84) ÷2 t test
Mean age in years (SD) 45.6 (12.9) 37.8 (11.5) p<0.001Female 110 (59%) 45 (54%) NSHaving children 145 (78%) 54 (64%)Married or cohabiting 139 (74%) 57 (68%) NSEmployed 132 (71%) 70 (84%) p<0.05*Previous history of cancer surveillance 126 (67%) 58 (69%) NSEducation in years: mean (SD) 11.05 (3.49) 12.06 (3.22) p<0.05*Pre-test risk perception: mean (SD) 1.98 (0.51) 2.11 (0.56) NS
NS = non-significant.*After adjustment for age, the diVerence disappeared.
choosing the incorrect option “misunderstand-ing” the result. Sociodemographic informationobtained from the questionnaires included age,gender, having children (1 = yes, 2 = no),marital status (1 = married or cohabiting, 2 =single, divorced, or widowed), employmentstatus (1 = employed, 2 = unemployed orretired), education in years, previous history ofcolorectal cancer surveillance (1 =colonoscopy/colonoscopies performed previ-ously, 2 = no colonoscopy performed) pre-testrisk perception (risk of having HNPCC: 1 =low, 2 = medium, 3 = high). General anxietywas measured by the state measure of theState-Trait Anxiety Inventory (STAI) which isa 20 item scale.24 Response categories for theitems range from 1 (not at all) to 4 (very muchso). Scores range from 20 to 80, the higherscores indicating greater state anxiety. In thisstudy, the measurements were performed atbaseline and at the test disclosure session. Inboth measurements, Cronbach alpha was 0.9,indicating high internal consistency. Worryabout the risk of developing colorectal cancerbased on the test results was assessed at the oneyear follow up by a question with multiplechoice answers. “Are you worried about yourcurrent risk of developing colorectal cancer?”(1 = not at all worried, 2 = worried to someextent, 3 = very worried, 4 = can’t say).
All data analyses were done with theprogram SPSS for Windows version 9.0. Westudied bivariate associations between groupsdefined by understanding the results andexplanatory variables, including socio-demographic information, anxiety at baselineand at the test disclosure session, and associa-tions between understanding the risk and per-ceived worry about it. DiVerences between cat-egorical variables were assessed with ÷2 (df)tests and McNemar tests (table 2), and diVer-ences between continuous variables with inde-pendent sample t tests. The variables showingstatistical significance in bivariate analysis(p<0.05) were subjected to binary logistic
regression analysis to predict misunderstand-ing of the cancer risk among mutation positiveand mutation negative subjects. Among themutation negative subjects, those who incor-rectly claimed that their risk was 50% or 100%were compared with those who correctlyreported that their risk was low. Among themutation positive subjects, those who claimedthat their risk was low or 50% were comparedwith those who correctly stated their risk to benear 100%. All predictors were analysed ascontinuous variables and were entered simulta-neously.
ResultsNearly all the respondents (268/268 at the onemonth follow up and 266/268 at the one yearfollow up) correctly recalled whether or notthey had inherited the mutation predisposingto cancer. However, the mutation negativesubjects understood their post-test risk ofdeveloping colorectal cancer significantly moreoften than those who were mutation positive(92% v 48%, ÷2=68.17 (1), p<0.0001), and atthe one year follow up, the diVerence was evengreater (90% v 36%, ÷2=86.19 (1), p<0.0001)(table 2). The answers were similar, irrespec-tive of the counsellor at the pre-test or the testdisclosure session. Regarding the mutationpositive subjects, misunderstanding at the onemonth follow up was more common among theolder (t=−2.09, p<0.05), the less educated(t=3.19 p<0.001), and those who had per-ceived the pre-test risk as lower than the others(t=3.17, p<0.01). Among mutation negativesubjects, those few who misunderstood theirrisk had perceived their (pre-test) risk to behigh (t=−3.0, p<0.01) and had high scores onanxiety immediately after the test disclosuresession (t=−3.01, p<0.01 ) as compared withthose who understood their risk. With regard toother demographic data, the groups did notdiVer.
According to the logistic regression modelpresented in table 3, the only predictor of mis-understanding the result was initially lowerpre-test risk perception among the mutationpositive group. Among the mutation negativegroup, those who perceived their pre-test riskto be higher and were anxious immediatelyafter the test disclosure were more likely tohave misunderstood the result. Because misun-derstanding had increased among the mutationpositive group at the one year follow up, wecarried out similar regression analysis; the sig-nificant predictor for misunderstanding con-tinued to be a lower pre-test perception of therisk (OR=0.27 (0.10-0.74)).
At the one year follow up of the mutationpositive subjects (n=83), 8% reported that theywere very worried about their risk of develop-ing colorectal cancer, 69% that they were wor-ried to some extent, 2% could not say whetherthey were worried or not, and 21% stated thatthey were not at all worried. The correspond-ing percentages for the mutation negative sub-jects (n=182) were 2%, 25%, 11%, and 62%,respectively (÷2=59.75 (3), p<0.0001). Fig 1illustrates an analysis that compares worryabout the risk with a correct or incorrect
Table 2 Perception of the post-test risk of developing colorectal cancer, assuming that noclinical surveillance existed at the one month or the one year follow up
understanding of the test results. In theillustration, worry was analysed as a continu-ous variable (0 = not worried, 1 = worried tosome extent, 2 = very worried; the option“can’t say” (n=22) was excluded). The muta-tion positive subjects who understood theresult were significantly more worried aboutthe risk of developing colorectal cancer thanthose with the mutation who did not under-stand the result correctly (mean scores 2.03 v1.78, t=2.04, p<0.05). By contrast, the muta-tion negative subjects who misunderstood thetest result were more worried about the riskthan their counterparts (mean scores 1.86 v1.27, t=−4.37, p<0.001).
DiscussionThe purpose of predictive genetic testing, asstated in the introduction, was partly met in ourstudy since nearly all the respondents recalledtheir test results correctly and most of those whowere mutation negative correctly interpretedtheir test result. However, a majority of themutation positive subjects underestimated theircancer risk as being only 50% or below, insteadof the correct very high risk of which they hadbeen informed in the counselling. One explana-tion may be the diYculty of expressing a highbut not inevitable risk in percentages and, there-fore, simplifying the risk, either it will happen orit will not. The misunderstanding may alsoreflect protective coping mechanisms such asdenial, for the misunderstanding was morecommon among those who actually had highrisk. A similar phenomenon has been describedin studies of testing of cystic fibrosis carriers.25 26
According to multivariate analysis, the only sig-nificant predictor of this phenomenon was aninitially low perception of the risk, which couldreflect incorrect information or, again, protec-tive coping mechanisms. Denial may be, on theone hand, an important coping mechanismwhich enables the mutation positive subjects toface the future. On the other hand, if it decreasesadherence to clinical surveillance, it could haveserious consequences.
Few data are available on cancer surveillancebehaviour after genetic testing. Lerman et al27
reported that many women (32%) with
BRCA1/2 mutation did not follow mammogra-phy surveillance recommendations one yearafter testing. However, the impact of riskperception was not studied and only young age(<40 years) predicted non-adherence, whichmay reflect the absence of data on the eYcacyof mammography among younger women.
In previous studies, the amount of worryabout cancer has been found to be associatedwith cancer screening behaviour in a compli-cated manner. Lerman et al27 found, in a popu-lation based study, that women who wereexcessively worried about breast cancer wereless likely to attend mammography.28 Moderateworry, however, may serve as an importantpromoter of cancer screening behaviour.29 Pre-test worry about cancer has been shown bothto be positively associated with high perceivedrisk of developing cancer30 and with intentionsand uptake of genetic testing.5 In our study,most of the mutation positive subjects wereworried, at least to some extent, about their riskof developing colorectal cancer and misunder-standing the result was associated with lessworry. There is a danger that misunderstand-ing connected with less worry may disturbadherence to surveillance. There is a pressingneed of studies to clarify this issue.
Misunderstanding of the results was muchless common (only 8-10%) among the muta-tion negative subjects. It was predicted by ahigher estimate of the pre-test risk and highscores on anxiety immediately after hearing thetest results. Consistently, these subjects weremore worried about their risk than those whoanswered correctly. This could reflect thepreviously described adverse feelings abouttesting in the mutation negative subjects, suchas trouble in finding a new life perspective, sur-vivor guilt, or worry about the mutationpositive family members.31 32
A recent review of risk communication ingenetic testing for cancer suggests that percep-tions of personal risks of cancer are resistant tostandard pre-test education and counselling.5
This reflects the challenges faced by health careprofessionals with regard to the patients’understanding of their risk. Should thoseunderestimating their risk be reinformed aboutthe actual risk after testing? This might even beimpossible in a normal clinical setting withoutpost-test questionnaires. However, as riskperception seems to be a complex issue, furtherresearch about the impact of diVerent post-testcounselling approaches on comprehension ofcancer risk is needed. In any case, it isnecessary to ensure that clinical surveillance isreadily available and that adherence to the sur-veillance is actively supported.
In contrast, if the health professionals wereaware of some who overestimated their post-test risk and were worried about it, should thesubjects concerned be oVered further counsel-ling? In the present study, the subjects involvedwere very few in number, which suggests thatfurther counselling could be performed with-out placing too great a strain on the health caresystem. However, published reports suggestthat improving risk comprehension may not be
Figure 1 The association between comprehending the testresult (cancer risk) correctly and extent of worry about itone year after receiving the test result.
successful among those with high levels of can-cer related distress. Thus, it is possible thatcounselling with the emphasis on the psycho-logical issues may be more beneficial than tra-ditional risk counselling.33
Our study population consisted of those whohad completed the whole testing procedure,including the questionnaires. Although thoselost to follow up did not diVer from the partici-pants in any of the background variables used inthis study, we know nothing about the post-testrisk perception, worry, or anxiety of thedropouts. The risk perception based on the testresult was assessed with one categorical variable.Alternatives of the question were based oninformation given during counselling (very high,close to 100%, for the mutation positive andvery low for the mutation negative). Before thetest, all the participants had been informed thatthey ran a 50% risk of cancer. The categoricalalternatives make the distinction between cor-rect and incorrect options easier to assess.Assessment using risk perception as a continu-ous variable (0-100 scale) would have beenproblematical because we had not providedexact numerical risk figures during counselling.Furthermore, it should be noted that possiblefindings of colon polyps or cancer after the testmay have acted as potential confounding factorson risk perception.
ConclusionsThis study provides the first data on compre-hending predictive genetic test results incancer, which suggest that the majority of themutation positive subjects tend to underesti-mate their risk. Whether post-test risk percep-tion aVects behaviour in terms of compliancewith cancer surveillance is still unknown. Thereis the danger that misunderstanding the testresult may aVect adherence to surveillance. Wetherefore suggest that predictive genetic testingfor HNPCC should be oVered in conjunctionwith a well organised cancer surveillance pro-gramme to promote participation independentlyof risk perception. The small number of thoseremaining worried despite an actual low riskshould be taken into account, possibly by oVer-ing further counselling sessions with emphasison psychological support.
We thank Marjo Molin for her contribution to the pre-testcounselling and expert assistance. This study was supported bya grant from the Academy of Finland and the Finnish CancerSociety.
1 Ponder B. Genetic testing for cancer risk. Science1997;278:1050-4.
3 Giardiello FM, Brensinger JD, Petersen GM, Luce MC,Hylind LM, Bacon JA, Booker SV, Parker RD, HamiltonSR. The use and interpretation of commercial APC genetesting for familial adenomatous polyposis. N Engl J Med1997;336:823-7.
4 Hubbard R, Lewontin RC. Pitfalls of genetic testing. N EnglJ Med 1996;334:1192-4.
5 Croyle RT, Lerman C. Risk communication in genetic testingfor cancer susceptibility. J Natl Cancer Inst 1999;25:59-66.
6 Lynch HT, de la Chapelle A. Genetic susceptibility to non-polyposis colorectal cancer. J Med Gen 1999;36:801-18.
7 Vasen HF, Wijnen JT, Menko FH, Kleibeuker JH, Taal BG,GriYoen G, Nagengast FM, Meijers-Heijboer EH, Ber-tario L, Varesco L, Bisgaard ML, Mohr J, Fodde R, KhanPM. Cancer risk in families with hereditary nonpolyposiscolorectal cancer diagnosed by mutation analysis. Gastroen-terology 1996;110:1020-7.
8 Aarnio M, Sankila R, Pukkala E, Salovaara R, Aaltonen LA,de la Chapelle A, Peltomaki P, Mecklin JP, Jarvinen HJ.Cancer risk in mutation carriers of DNA-mismatch-repairgenes. Int J Cancer 1999;81:214-18.
9 Ford D, Easton DF, Stratton M, Narod, S. Goldgar D, Dev-ilee P, Bishop DT, Weber B, Lenoir G, Chang-Claude J,Sobol H, Teare MD, Struewing J, Arason A, Scherneck S,Peto J, Rebbeck TR, Tonin P, Neuhausen S, BarkardottirR, Eyfjord J, Lynch H, Ponder BA, Gayther SA,Zelada-Hedman M. Genetic heterogeneity and penetranceanalysis of the BRCA1 and BRCA2 genes in breast cancerfamilies. The Breast Cancer Linkage Consortium. Am JHum Genet 1998;62:676-89.
10 Thorlacius S, Struewing JP, Hartge P, Olafsdottir GH,Sigvaldason H, Tryggvadottir L, Wacholder S, Tulinius H,Eyfjord JE. Population-based study of risk of breast cancerin carriers of BRCA2 mutation. Lancet 1998;352:1337-9.
11 Warner E, Foulkes W, Goodwin P, Meschino W, Blondal J,Paterson C, Ozcelik H, Goss P, Allingham-Hawkins D,Hamel N, Di Prospero L, Contiga V, Serruya C, Klein M,Moslehi R, Honeyford J, Liede A, Glendon G, Brunet JS,Narod S. Prevalence and penetrance of BRCA1 and BRCA2gene mutations in unselected Ashkenazi Jewish women withbreast cancer. J Natl Cancer Inst 1999;91:1241-7.
12 Jarvinen HJ, Mecklin JP, Sistonen P. Screening reducescolorectal cancer rate in families with hereditary nonpoly-posis colorectal cancer. Gastroenterology 1995;108:1405-11.
13 Vasen HF, van Ballegooijen M, Buskens E, Kleibeuker JK,Taal BG, GriYoen G, Nagengast FM, Menko FH, MeeraKP. A cost-eVectiveness analysis of colorectal screening ofhereditary nonpolyposis colorectal carcinoma gene carri-ers. Cancer 1998;82:1632-7.
14 Codori AM, Petersen GM, Miglioretti DL, Larkin EK,Bushey MT, Young C, Brensinger JD, Johnson K, BaconJA, Booker SV. Attitudes toward colon cancer gene testing:factors predicting test uptake. Cancer Epidemiol BiomarkersPrev 1999;8:345-51.
15 Glanz K, Grove J, Lerman C, Gotay C, Le Marchand L.Correlates of intentions to obtain genetic counseling andcolorectal cancer gene testing among at-risk relatives fromthree ethnic groups. Cancer Epidemiol Biomarkers Prev1999;8:329-36.
16 Lerman C, Hughes C, Trock BJ, Myers RE, Main D, Bon-ney A, Abbaszadegan M, Harty A, Franklin B, Lynch J,Lynch H. Genetic testing in families with hereditarynonpolyposis colon cancer. JAMA 1999;281:1618-22.
17 Aktan-Collan K, Mecklin JP, Jarvinen H, Nystrom-Lahti M,Peltomaki P, Soderling I, Uutela A, de la Chapelle A, Kaar-iainen H. Predictive genetic testing for hereditary non-polyposis colorectal cancer: uptake and long-term satisfac-tion. Int J Cancer 2000;89:44-50.
18 Johnson KA, Rosenblum-Vos L, Petersen GM, BrensingerJD, Giardiello FM, GriYn CA. Response to genetic coun-seling and testing for the APC I1307K mutation. Am J MedGenet 2000;91:207-11.
19 Stanley A, GaV C, Aittomäki K, Fabre L, Macrae F, St JohnD. Value of predictive genetic testing in management ofhereditary non-polyposis colorectal cancer (HNPCC). MedJ Aust 2000;172:313-16.
20 Nystrom-Lahti M, Kristo P, Nicolaides NC, Chang SY, Aal-tonen LA, Moisio AL, Jarvinen HJ, Mecklin JP, KinzlerKW, Vogelstein B, de la Chapelle A, Peltomaki P. Foundingmutations and Alu-mediated recombination in hereditarycolon cancer. Nat Med 1995;1:1203-6.
21 Holmberg M, Kristo P, Chadwicks RB, Mecklin JP, JarvinenH, de la Chapelle, Nystrom-Lahti M, Peltomaki P.
+ Comprehension of predictive genetic testresults and factors influencing it weremeasured with a questionnaire surveyone and 12 months after disclosure in271 subjects tested for HNPCC.
+ Nearly all respondents recalled correctlywhether or not they had inherited themutation predisposing to HNPCC.
+ A majority of the mutation positivesubjects incorrectly interpreted (underes-timated) their test result in terms of can-cer risk at both follow ups.
+ A great majority of the mutation negativesubjects interpreted their likelihood ofdeveloping cancer correctly.
+ Perceived pre-test risk seemed to be thebest predictor of misunderstanding, irre-spective of the test result.
+ As misunderstanding the test result maycomplicate adherence to cancer surveil-lance, oVering predictive testing in con-junction with a well organised cancerprogramme might be beneficial.
Mutation sharing, predominant involvement of the MLH1gene and description of four novel mutations in hereditarynonpolyposis colorectal cancer. Mutations in brief No 144.Online. Hum Mutat 1998;11:482.
22 Aktan-Collan K, Mecklin JP, de la Chapelle A, Peltomaki P,Uutela A, Kaariainen H. Evaluation of a counselling proto-col for predictive genetic testing for hereditary non-polyposis colorectal cancer. J Med Genet 2000;37:108-13.
23 Vasen HF, Mecklin JP, Khan PM, Lynch HT. TheInternational Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer (ICG-HNPCC). Dis ColonRectum 1991;34:424-5.
24 Spielberger C, Lushene R. Manual for the state trait anxietyinventory. Palo Alto: Consulting Psychologists Tests, 1970.
25 Axworthy D, Brock DJ, Bobrow M, Marteau TM.Psychological impact of population-based carrier testingfor cystic fibrosis: 3-year follow-up. UK Cystic FibrosisFollow-Up Study Group. Lancet 1996;347:1443-6.
26 Denayer L, Welkenhuysen M, Evers-Kiebooms G, Cassi-man JJ, Van den Berghe H. Risk perception after CF carriertesting and impact of the test result on reproductivedecision making. Am J Med Genet 1997;69:422-8.
27 Lerman C, Hughes C, Croyle RT, Main D, Durham C, Sny-der C, Bonney A, Lynch JF, Narod SA, Lynch HT. Prophy-lactic surgery decisions and surveillance practices one yearfollowing BRCA1/2 testing. Prev Med 2000;31:75-80.
28 Lerman C, Daly M, Sands C, Balshem A, Lustbader E,Heggan T, Goldstein L, James J, Engstrom P. Mammogra-phy adherence and psychological distress among women atrisk for breast cancer. J Natl Cancer Inst 1993;85:1074-80.
29 Lerman C, Trock B, Rimer BK, Boyce A, Jepson C, EngstromPF. Psychological and behavioral implications of abnormalmammograms. Ann Intern Med 1991;114:657-61.
30 Collins V, Halliday J, Williamson R. Cancer worries, riskperceptions and associations with interest in DNA testingand clinic satisfaction in familial colorectal cancer clinic.Clin Genet 2000;58:460-8.
31 Huggins M, Bloch M, Wiggins S, Adam S, Suchowersky O,Trew M, Klimek M, Greenberg CR, EleV M, ThompsonLP. Predictive testing for Huntington disease in Canada:adverse eVects and unexpected results in those receiving adecreased risk. Am J Med Genet 1992;42:508-15.
32 Tibben A, Frets PG, van de Kamp JJ, Niermeijer, MF,Vegtervan der Vlis M, Roos RA, Rooymans HG, vanOmmen GJ, Verhage F. On attitudes and appreciation 6months after predictive DNA testing for Huntingtondisease in the Dutch program. Am J Med Genet1993;48:103-11.
33 Lerman C, Lustbader E, Rimer B, Daly M, Miller S, SandsC, Balshem A. EVects of individualized breast cancer riskcounseling: a randomized trial. J Natl Cancer Inst 1995;87:286-92.
Population prevalence and estimated birthincidence and mortality rate for people withPrader-Willi syndrome in one UK Health Region
J E Whittington, A J Holland, T Webb, J Butler, D Clarke, H Boer
EDITOR—Prader-Willi syndrome (PWS) is agenetically determined disorder in which theabsence of expression of one or more mater-nally imprinted gene(s) in the chromosomalregion 15q11-13 results in a characteristicfacial appearance, learning disabilities (mentalretardation), and severe overeating behaviourowing to an abnormal satiety response to foodintake, together with a range of other behav-iours. Initially, as reported by Prader et al,1
PWS was conceived as a syndrome of obesity,short growth, cryptorchidism, and mentalretardation following hypotonia in the neonatalperiod. As more and more people with PWSwere reported and research into the syndromebegan, behavioural characteristics and otherclinical features were added, culminating in theconsensus diagnostic criteria.2 Concurrently,the genetics of the disorder were receivingattention. First was the discovery that for manythere was a visible chromosomal deletion in theproximal part of the long arm of chromosome15 (15q11-13). Reports of an apparently simi-lar deletion being associated with a phenotypi-cally very diVerent syndrome (Angelman syn-drome, AS),3 and the observation that PWSwas the result of a deletion on the chromosome15 of paternal origin, and AS the chromosome15 of maternal origin, led to the recognitionthat gender specific imprinting of genes at thatlocus accounted for two diverse syndromesbeing associated with apparently similar chro-mosomal deletions.4 Maternal chromosome 15disomies, mutations of an imprinting centre,and chromosomal translocations accounted fornon-deletion cases of PWS.5
In published reports on Prader-Willi syn-drome (PWS), prevalence has been variouslyquoted as “about 1 in 25 000 live births”,6
“between one in 25 000 and one in 10 000 liveborn children”,7 “[estimates] vary 6-fold from1 in 5000 to 10 000; 1 in 10 000; 1 in 15 000;1 in 25 000; to 1 in 10 000 to 30 000”.8 Onlytwo estimates appear to be based on epidemio-logical data, those of Akefeldt et al7 and Burd etal.8 In the latter North Dakota study, theauthors surveyed paediatricians, neurologists,and clinical geneticists and also contacted thestate’s comprehensive evaluation centre, thestate hospital, the state institution for the“mentally retarded”, and group homes for thedevelopmentally disabled, including one forpeople with PWS. In most communities, atleast four of these sources of information wereconsulted. Each was sent a one page question-naire pictorially illustrating the signs of PWS toaid identification. The response rate was 99%.These procedures yielded eight males, eightfemales, and one person whose gender was notgiven, with an age range from 9 to 30 years. Atthat time the population of North Dakota forthat age range was 263 444, giving a prevalencerate of 1:16 062, equivalent to 1:38 395 in theentire population. No figures were given for thenumber of cases with a genetic diagnosis.
In the study of Akefeldt et al,7 the authorsestimated the prevalence of PWS in the agerange 0 to 25 years in the rural Swedish countyof Skaraborg, by surveying paediatricians, neu-ropaediatricians, child psychiatrists, schoolhealth visitors, general practitioners, and doc-tors working in the fields of general medicine,rehabilitation, and mental disabilities. The
authors circulated diagnostic criteria for PWSto these professionals and also invited all schoolnurses in the county to seminars where PWSwas described. Requests were made that allpeople with possible PWS be notified to them.All people with possible PWS were examinedby a neuropaediatrician, a child psychiatrist, achild psychologist, and a speech pathologist.Clinical diagnoses were made on the basis oftheir findings. Eleven people (seven male andfour female) were considered definitely to havePWS and a further five (two male and threefemale) were considered probable. Thesenumbers gave a population prevalence of “clearPWS” up to the age of 25 years of 1:8 500 andbetween 7 to 25 years 1:8000. If the peoplewith suspected PWS were also included, theseprevalence rates became 1:6700 and 1:5000,respectively.
These diVerent population estimates can bereconciled in several ways, including samplingerrors, diVerences in ascertainment, high mor-tality rates between the ages of 25 and 30, orvarying prevalence rates in diVerent popula-tions. The first two of these probably apply, atleast to some extent. High mortality rates havenever been reported as specific to a particularage group. The final explanation is at odds withthe accepted view of PWS as a randomlyoccurring de novo genetic mutation. Thisaccepted view is supported by the fact that in
both studies the gender ratio is close to 1:1, if,in the Swedish study, people with “probable”as well as “definite” PWS are counted. Clearly,if the explanation was that prevalence rates varydepending on the population studied, thenthese estimates cannot be generalised to othercountries.
We describe a population prevalence study inthe UK that was undertaken as part of a largerstudy investigating the heterogeneity of thePWS phenotype. The aim was to locate alldefinite or possible cases of PWS in the previ-ous Anglia & Oxford Health Region, compris-ing the counties Bedfordshire, Berkshire, Buck-inghamshire, Cambridgeshire, Norfolk,Northamptonshire, Oxfordshire, and SuVolk(population approximately 5 million people).All ages were to be included, so as to obtainsome estimate of mortality rates in the PWSpopulation, and as far as possible the diagnosiswas confirmed by the routine diagnosticmethod using SNRPN methylation,9 or, if thatwas not possible, the presence of accepteddiagnostic criteria.
MethodsSince the ideal method of random sampling isimpractical for rare disorders requiring clinicaldiagnosis, the method of counting all knowncases in the region was used, as it was in thestudies quoted above. The size of the regionstudied was chosen after administrative andstatistical considerations. The region chosencomprises approximately five million people,about one tenth of the population of Englandand Wales. Using a Health Region is advanta-geous administratively, since ethics committeesand laboratory and clinical facilities are regionbased. The population prevalence rates foundin the previous studies quoted above, togetherwith statistical power considerations, suggestedthat the study of the phenotype would require apopulation of at least this size.
Health, education, and social services pro-fessionals (as detailed in the appendix) werecontacted and sent a list of common PWScharacteristics and asked if they knew ofanyone fulfilling at least half of these (seeappendix). In addition, the UK PWS Associ-ation was asked to advise about numbersknown to the Association living in the studyarea. To preserve anonymity, initials, gender,and date of birth (or age) of their nomineeswere requested; these were necessary in orderto avoid double counting of people who possi-bly had the syndrome. They were also asked ifthey were prepared to forward a letter from usto the person with PWS inviting their participa-tion in the phenotypic study. (Note thatanonymity and indirect contact were condi-tions of ethical approval for the study; thereforeno other details are known of people nomi-nated who did not contact us). All responderswho were willing were recruited into thephenotypic study for cognitive assessment ofthe person with PWS and in depth interviews(including a section asking about the presenceor not of the symptoms included in theconsensus diagnostic criteria) with a parent ormain carer or both. Blood samples were later
Figure 1 Breakdown of group.
NomineesTotal = 167
Possible PWS= 123
Not traced= 8
Dead = 2
Refused = 31
31 = No blood8 = Genetic, 23 = clinical
67 = Blood48 = Genetic19 = Not PWS
25 = No blood12 = Genetic, 5 = clinical8 = Not PWS
Seen = 92
Out of area= 22
Not PWS:Parent says = 4Description = 8
Table 1 Population classification of nominees by source
Genet PWS Clinic PWS Not seen Non-PWS% positive ofthose seen
collected from all those who were willing andable to provide samples. In some cases, whenthe subject was unwilling or there werediYculties obtaining a sample, permission wasgiven for access to previous genetic records.For most participants in the phenotypic study,therefore, a diagnosis of PWS could beconfirmed either clinically (parents/carers wereasked at interview about the presence or not of
consensus diagnostic criteria) and/or geneti-cally. In the region, there is a collection ofgroup homes for people with PWS. Only thosefunded from within the study region wereincluded in the population part of the study,the others coming from outside the region.
During the active period of the study,between September 1998 and June 2000, 167people with a diagnosis of PWS or with PWSsymptoms were notified to us from the sourceslisted earlier, including nominations by parentsand other relatives of people with eating prob-lems and other PWS symptoms who respondedto media appeals. Fig 1 shows the breakdownof this group in terms of participation in thephenotypic study, blood samples obtained,genetic test results, and clinical diagnoses. Thisfigure shows that after eliminating those nomi-nees who were found to have died, wereuntraceable, found on preliminary investiga-tion not to have PWS, or had moved out of theregion, 123 people with possible PWS wereincluded. Of these 123, 31 with a previousdiagnosis of PWS did not wish to participate inthe phenotypic study, and therefore the onlydata are those provided by those who referredthe person to the research study. Eight of thesewere from genetic laboratories, and in theremaining 23 cases the genetic status was notgiven. Many (75) of the 123 nominees werefrom multiple sources, as detailed in table 1;many were nominated as “maybe”, particularlyparental nominations. None of the 31 not seenwere “maybe”. Table 1 also gives the hit rates(that is, number of positive cases) for the casesseen among the various sources.
Among the 92 who agreed in principle toparticipate in the phenotypic study, genetictesting was possible for 67 participants. Of theother 25, eight were considered not to havePWS at interview. With 12, results of previousgenetic tests were obtained confirming thediagnosis. The remaining five were included asthey met clinical criteria. Altogether, 27 peopleout of 92 were deemed not to have PWS (theabove eight on clinical grounds only, 11 onboth clinical and genetic grounds, a furthereight on genetic grounds) and 65 met clinicaland/or genetic criteria. Our report is based onthe latter group (65 people) and the 31 who didnot wish to be involved in the larger study, giv-ing a total of 96.
Fig 2 shows the frequencies by age (stand-ardised to midnight 31.12.99) of these 96people.
ResultsThere are two obvious potential sources oferror in our results, false positives and missingcases. With respect to the former, among thosepeople seen, we have 60 with genetic confirma-tion and five who did not have a genetic test(three refused, two tried and failed) but metclinical diagnostic criteria. Among the 31 notseen, eight had genetic confirmation; table 1suggests that at least 80% of the others areexpected to be positive, so that we shouldexpect at most five false positives even beforeconsideration of the firmness of the nomina-tions, which led us to include all cases.
Figure 2 Frequencies by age.
15
10
5
0
Age
Freq
uen
cy
0–2
3–5
6–8
9–11
12–1
4
15–1
718
–20
21–2
3
24–2
627
–29
30–3
2
33–3
536
–38
39–4
142
–44
45–4
7
7
6
5
4
3
2
1
00 10 20 30 40 50
Freq
uen
cy
30
20
10
00–8 9–17 18–26 27–35 36–44 45+
Freq
uen
cy
Table 2 Numbers, gender diVerences and ages of people with PWS identified by county
With respect to missing cases, we believe thisto be more of a problem for several reasons.The weakest evidence for missing cases isregional variations in prevalence, muchstronger are age by county and gendervariations.
Table 2 gives the numbers of people withPWS divided according to each county. Theoverall population rate was 1:52 000, withcounty rates varying from 1:42 000 to 1:67 000(not statistically significant). This could indi-cate that identification was better in some areasthan in others, and a conservative estimatesuggests that there may be another sevenpeople with PWS within those counties withlow reported rates. But such variation may alsoindicate random variation and that the twoprevious prevalence study estimates can bereconciled by random sampling. North Dakotais similar in population size to a single county,while Skaraborg is only half that populationsize.
Ages were standardised to midnight on 31December 1999. Table 3 gives the numbers ofpeople with probable or definite PWS acrossthe age range. Assuming that roughly the samenumber of babies with PWS are born everyyear, an examination of the overall agestructure of our sample suggests that some veryyoung children with PWS may not yet have
been diagnosed (table 3), awaiting the onset ofeating and behavioural problems to promptinvestigations. (This is supported by age atdiagnosis data collected in interviews; newdiagnoses were common in the first year andbetween 3 and 5 years). If we assume a zeromortality rate for children under 6, table 3suggests that as many as five children withPWS may not yet have been diagnosed. Theage structure of identified cases within countiesalso supports the “missing cases” hypothesis.In Norfolk, for example, only one person under18 years out of 13 was identified, while inBerkshire there was only one person with PWSover 17 years out of 11 identified.
In all but one county, we found more malesthan females (55 males, 33 females, eight gen-der not stated). Assuming that the gender ratioin PWS is close to unity, as the two previousstudies found, it therefore seems that femaleswith PWS were either specifically not beingdiagnosed or were less likely to be referred tothe study than males, and we estimate theremay be between 14 and 30 more females withPWS in the region. Of the four infants aged0-2, three were male, and in both Cambridge-shire and Norfolk we found six more malesthan females (table 3).
Unlike the previous population studieswhich yielded 17 probable cases, age range9-30, in North Dakota and 16 cases, age range0-25, in Skaraborg, we have a suYciently largesample to look at age frequencies. A number ofdiVerent methods are available to estimatebirth incidence from the raw data (which isshown at the top of fig 2). All methods dependon the diVerent assumptions we are preparedto make, the chief common assumption beingthat PWS results from a randomly occurringgenetic fault, that is, the age frequenciesconstitute a random sequence about a steadyincidence rate. (We note that this sequencewould vary slightly if we chose a diVerent datefor age standardisation.) This sequence can besmoothed in various ways, as shown in fig 2,where the data have been grouped in three yearintervals and in nine year intervals. In the lattercase, the data were fit by a quadratic with anR2=0.996 (an exceptionally good fit). We havealso assumed no false positives in ascertain-ment. Assuming no missing data would thengive lower bounds for the prevalence and birthincidence for a particular method of calcula-tion of the latter.
Assuming no missing data and using thesmoothed data obtained from nine year group-ings, we find rates of 22, 23, 21 cases in the firstthree nine year periods, a rate of 1:28 000, witha zero mortality rate to age 27 and thereafter anaverage rate of 6.1% (from 2.44 to 0 in 20years).
A similar calculation with eight year group-ing (fit by a quadratic with R2=0.85) and ratesof 17, 17, 25, 17 cases in the first four eight yearperiods, gives a rate of 1:29 000, with a zerorate of mortality to age 32 and thereafter anaverage rate of 7.9% (from 2.375 to 0 in 15years).
The paragraphs above suggest that data aremissing (females) but may be fairly evenly
Table 3 Numbers across the Region divided by age andsex
spread across the age groups (age by county). Ifthis is the case, then the falling number of caseswith age is probably because of mortality andwe can use the whole distribution to estimatelower bounds for birth incidence and mortalityrate. Fig 3 illustrates this method, where thefitted line is the result of ignoring ages 0-5(since we hypothesised that cases might bemissing in this age range, their eventualdiagnosis awaiting the onset of the secondphase of PWS symptoms). A straight line, asshown, was found to be near the optimum fitwith R2=0.41 (R2 could be increased to 0.44 byfitting a fourth order polynomial). This wouldimply a steady death rate across the age groups,rather than an increasing rate with age. (Anec-dotally, we were told of two people with PWSwho died before we could see them, a 3-4 yearold girl and a 13 year old boy). From fig 3, wesee that 10.2 children are expected in the agerange 0-2 (that is, in three years), giving a birthincidence of 1:20 000 in the populationstudied (68 000 births per year). We also seethat the overall death rate in this PWS popula-tion is about 3% per year (that is, from 3.4 to 0over 56 years). This compares with an overalldeath rate in the population of England andWales of about 1% per year, and only about0.13% per year in those up to 55 years.
Another set of assumptions is that advancesin genetics would result in increasing numbersof accurate diagnoses, so that ascertainment byage would be a step function, with steps corre-sponding to identification of deletions (c 1981,age 18), maternal disomy (c 1989, age 10), andthe imprinting centre methylation tests (c1995, age 4). This is clearly not supported byour data.
A final set of assumptions are that data aremissing only, or mainly, from the older agegroups. This is consistent with the 8 year and 9year groupings if we assume little or nomortality. But there are two arguments againstthis. First, the county data and age at diagnosisdata suggest that missing cases are not confinedto older age groups; in some counties such agegroups are over-represented. Secondly, severalolder people were nominated as possible PWSand were interviewed but were found to begenetically negative, so failure to find olderpeople was not because of lack of candidates.
DiscussionThere are a number of problems related to thismethod of conducting a prevalence study ofPWS. The three main questions are: howshould PWS be defined?, do all nominees havePWS?, and do all other people in the region nothave PWS?
The first question still remains to someextent unresolved as neither the genetic testnor the clinical diagnosis is 100% definitive.Although it is generally agreed that geneticabnormalities of the Prader-Willi critical re-gion on the chromosome 15 of paternal originunderlie the clinical manifestations of the syn-drome, and that the dual genetic tests used willdetect probably 99% of such abnormalities, theprecise gene(s) involved have not been identi-fied and therefore a definitive genetic test is not
available. From the clinical perspective, it isclear that there are other chromosomal abnor-malities, such as deletions of part of chromo-some 6 leading to PWS-like symptoms (onesuch case is known to us), which show thatreliance on clinical diagnosis may result in theinclusion of those whose clinical phenotyperesults from a diVerent genotype. There wasone person in our population study with a scoreof 9 on the diagnostic criteria who wasexcluded when extensive genetic tests provednegative.
With respect to the second question, we can-not be certain whether everyone included inthis study definitely has PWS. Some nomi-nated did not respond to requests to take partin the phenotype study, and therefore we havehad to rely on previous clinical and/or geneticdata. Others did not wish for a blood test and,in the absence of a previous test, a geneticdiagnosis was not possible. However, we havebeen restrictive about who was or was notincluded. We excluded all those 27 nomineeswho did not meet either clinical criteria (noblood), genetic criteria, or both clinical andgenetic criteria for PWS. These people hadnormal maternal and paternal methylationpatterns and gene expression at the SNRPNlocus (where blood was available) and typicallyhad less than 50% of the required diagnosticcriteria. We have included 60 people who haddefinite genetic diagnoses and five others whohad definite clinical diagnoses (in the absenceof genetic confirmation). The 31 who did notwish to participate in the phenotypic studywere all included as having PWS, as detailedearlier.
Our assessment of findings relevant to thethird question leads us to believe that there areother people in the region who do have PWS. Ifit is assumed that the syndrome is indeedcaused solely by random genetic error (that is,with no environmental component), it followsthat prevalence should not be biased withrespect to gender, race, or environment.Although some relationship to paternal occu-pation in the hydrocarbon industry has beendescribed in the past, which might possiblylead to clustering,10 this has not been repli-cated. With the above assumptions, it is possi-ble to look for internal consistencies within ourregional sample, such as by county, by gender,or by race. There is not enough variation inracial background to test our sample withrespect to race. In all but one county, we foundmore males than females (55 males, 33females, eight gender not stated); this agreeswith previously published reports.11 Assumingthat the gender ratio in PWS is close to unity, asthe two previous studies found, it thereforeseems that most of any missing cases (estimate14 to 30) will be female. This is not surprising;at all ages, suspicion of PWS is more likely inmales (hypogonadism in childhood, short stat-ure with obesity in adulthood). However, intwo of the three age groups in which femalesdid in fact outnumber males, the age group wasin the over 30s raising the possibility that, withPWS, as in the general population, womenoutlive men. Given the assumptions about
those with PWS not identified, the findingssuggest that a population prevalence of1:52 000 is a lower bound and that the trueprevalence is somewhat higher, nearer to1:45 000.
Unlike the two previous population surveysof PWS, the present survey was large enough toallow the age structure of the PWS populationto be examined. We were able to estimate bothbirth incidence and death rate. The estimate ofmortality rate assumes that there is no system-atic age related ascertainment bias. For exam-ple, the age data might have been stepped cor-responding to advances in awareness anddiagnostic refinements, but the age structurefound argues against such a cohort eVect.However, it is possible that there are olderadults with PWS born before the syndromewas well known and in whom the diagnosis hasnever been considered. While this possibilitycannot be ruled out, we would argue that anincreased mortality rate is the most likelyexplanation of the age profile of this populationbased cohort of people with PWS. High rates ofboth physical and psychiatric disorder havebeen reported12 and preliminary analysis offurther data from this study provides evidenceof significant morbidity. It would seem likelythat obesity and associated complications arethe main contributory factor. However, asystematic prospective study is required toestablish whether this is the case or not. Thephysical health needs of people with learningdisability are often not properly assessed oradequately met,13 and disorders such as type 2diabetes might easily be missed in this popula-tion. There is also concern that misguided sup-port strategies in adult life may result inmarked weight increase and that resultantobesity related problems then occur.14 15
JW and JB were supported by a grant from the Wellcome Trustwith further support from the PWS Association. We are verygrateful to both organisations for their funding of this study. Wewould specifically like to thank Rosemary Johnson of the PWSAfor all her administrative support, and all those people withPWS, their families, and other carers and many people whohelped us with this study. This includes medical professionals,those working in genetic clinics, social services placements, edu-cation and learning disability services, and the media forcontacting people on our behalf.
1 Prader A, Labhart A, Willi H. Ein Syndrom von Adipositas,Kleinwuchs, Kryptorchismus und Oligophrenie nach mya-tonieartigem Zustandim Neugeborenenalter. Schweiz MedWochenschr 1956;86:1260-1.
2 Holm VA, Cassidy SB, Butler MG, Hanchett JM,Greenswag LR, Whitman BY, Greenberg F. Prader-Willisyndrome: consensus diagnostic criteria. Pediatrics 1993;91:398-402.
3 Magenis RE, Brown MG, Lacy DA, Budden S, LaFranchiS. Is Angelman syndrome an alternate result ofdel(15)(q11q13)? Am J Med Genet 1987;28:829-38.
5 Nicholls RD. Genomic imprinting and candidate genes inPrader-Willi and Angelman syndromes and mouse models.Curr Opin Genet Dev 1993;3:445-56.
6 Butler MG. Prader-Willi syndrome: current understandingof cause and diagnosis. Am J Med Genet 1990;35:319-32.
7 Akefeldt A, Gillberg C, Larsson C. Prader-Willi syndromein a Swedish rural county: epidemiological aspects. DevMed Child Neurol 1991;33:715-21.
8 Burd L, Vesely B, Martsolf J, Kerbeshian J. Prevalence studyof Prader-Willi syndrome in North Dakota. Am J MedGenet 1990;37:97-9.
9 Glenn CC, Saitoh S, Jong MTC, Filbrandt MM, Surti U,Driscoll DJ, Nicholls RD. Gene structure, DNA methyla-tion, and imprinted expression of the human SNRPN gene.Am J Hum Genet 1996;58:335-46.
10 Akefeldt A, Anvret M, Grandell U, Norliner R, Gillberg C.Parental exposure to hydrocarbons in Prader-Willi syn-drome. Dev Med Child Neurol 1995;37:1101-9.
11 Cassidy SB, Forsythe M, Heeger S, Nicholls RD, Schork N,Benn P, Schwartz S. Am J Med Genet 1997;68:433-40.
12 Greenswag LR. Adults with Prader-Willi syndrome: Asurvey of 232 cases. Dev Med Child Neurol 1987;29:152-4.
13 Lennox NG, Kerr MP. Primary health care and people withan intellectual disability: the evidence base. J Intellect Disa-bil Res 2000;41:365-72.
14 Dykens EM, GoV BJ, Hodapp RM, Davis L, Devanzo P,Moss F, Halliday J, Shah B, State M, King B. Eating them-selves to death: have “personal rights” gone too far in treat-ing people with Prader-Willi syndrome? Ment Retard 1997;35:312-14.
15 Holland AJ, Wong J. Genetically-determined obesity inPrader-Willi syndrome: the ethics and legality of treatment.J Med Ethics 1999;25:230-6.
Appendix. Contacts with groups andprofessionals in search for people withPWS: PWSA (UK)Genetics Departments - all known postholders.Paediatricians - registered members of the RoyalCollege of Paediatricians.Paediatricians - listed in the Anglia & Oxford HealthRegion Handbook.Psychiatrists in learning disabilities - registered mem-bers of the Royal College of Psychiatrists.Psychiatrists in learning disabilities - listed in the Anglia& Oxford Health Region Handbook.GPs - Cambridge area only (not cost eVective).Dieticians.Community nurses.County CouncilsSpecial education.Social services.Special schools for learning disabilities.Residential homes for learning disabilities.Media - newspapers, television, radio.Information given to contacts
+ The identification of all people in oneHealth Region, who may have the geneti-cally determined Prader-Willi syndrome,was attempted by direct contact withhealth, social, and educational servicesand through the PWS Association and byappeals through the local media.
+ From a total of 167 people referred withpossible PWS, 96 were included havingPWS on genetic and/or clinical grounds.Variation across the eight counties in theHealth Region and observed genderdiVerences enabled us to estimate thenumbers of people with PWS which weremissed by the ascertainment methodsused. From the age related prevalencedata, both birth incidence and mortalityrate were estimated.
+ We concluded that a lower bound for thepopulation prevalence was 1:52 000. Alower bound for the birth incidence wasestimated to be 1:29 000. The meanmortality rate was estimated to be 3% peryear across the age range or about 7% peryear above age 30, depending on assump-tions made.
+ While diVerential ascertainment acrossthe age range might be a factor, weconclude that these findings largely indi-cate significant morbidity and associatedmortality. We predict that this is becauseof a failure to manage weight gain in laterlife and thereby prevent associated obes-ity related health problems.
Do you know anyone who has been diagnosed as havingPWS or may have had five of the following characteris-tics?Floppiness at birth.Initial failure to thrive or diYculty in sucking.The development of severe overeating and rapid weightgain in early childhood.
Obesity or the need for weight control.Problems with sexual development (for example, unde-scended testes, delayed periods).Some learning disability (mental handicap).Small hands and feet.Short stature or the need for growth hormone.An abnormality of chromosome 15.
Mosaicism for 45,X cell line may accentuate theseverity of spermatogenic defects in men withAZFc deletion
Jadwiga Jaruzelska, Aleksandra Korcz, Alina Wojda, Piotr Jedrzejczak, Joanna Bierla,Tatiana Surmacz, Leszek Pawelczyk, David C Page, Maciej Kotecki
EDITOR—Over the past 10 years, severalauthors have reported microdeletions in thelong arm of the Y chromosome (Yq) in menwith idiopathic, non-obstructive azoospermiaor severe oligospermia. These microdeletionswere clustered on the Yq fragment previouslydescribed as the azoospermia factor region(AZF).1 More recently, a number of genesexpressed specifically in the testes and mappingto AZFa, AZFb, or AZFc subregions have beencloned.2–4 One of the approaches to under-standing the role of these genes in human sper-matogenesis is to look for a correlation betweenthe lack of given AZF genes and the particularspermatogenic defect in the phenotypes of thepatients. However, attempts to find such a cor-relation have failed so far. Instead, a broadspectrum of phenotypes ranging clinically fromazoospermia to severe oligospermia and histo-logically from Sertoli cell only syndrome(SCOS) to hypospermatogenesis has beendescribed in association with AZFc dele-tions.5 6
A recent study found chromosomal aberra-tions in 15% of azoospermic patients.7 How-ever, in papers focusing on the analysis of AZFmicrodeletions in patients with idiopathicinfertility,2 3 5 8–30 systematic, bilateral, histologi-cal, molecular, and cytogenetic analyses in thesame large group of patients was rarely carriedout, thus limiting information on the coexist-ence of AZF deletions and chromosomal aber-rations.
In this study, we propose and test thehypothesis that chromosomal defects mayoften accompany AZF deletions and cause thelack of a genotype-phenotype correlation inhuman male idiopathic infertility. We alsoattempt to evaluate the nature of the sperma-togenetic failure associated with isolated AZFcdeletions. For this purpose, we performed adual genetic analysis of karyotypes and mo-lecular status of the AZF region along withbilateral testicular histological evaluation in 94patients with non-obstructive, idiopathic infer-tility and azoospermia, severe oligospermia, oroligospermia.
Material and methodsSixty five men with azoospermia (lack of spermcells in semen), 23 men with severe oligosper-mia (fewer than 5 × 106 sperm cells/ml semen),and six with oligospermia (5-10 × 106 spermcells/ml semen), all of them of Polish origin,were included in the study.
Histological analyses of biopsies from bothtestes of 77 patients were performed informalin fixed paraYn embedded tissue blocks.Sections were cut at 4 µm thickness and stainedwith haematoxylin-eosin.
Chromosome studies were carried out onperipheral blood lymphocytes of 93 out of 94patients using GTG, FPG, CBG, and QFQbanding. Karyotypes were analysed in at least100 metaphases.
DNA was isolated from 10 ml of peripheralblood leucocytes of the patients and, whenavailable, also from the fathers or other malerelatives on the paternal side. For molecularanalysis, genomic DNA was amplified by PCRusing primers specific for 23 Y chromosomespecific sequence tagged site (STS) markers(19 mapping to AZFa, AZFb, and AZFc andfour mapping to short arm of the Y chromo-some) according to conditions described in theGenebank entry.
ResultsHISTOLOGICAL PHENOTYPES
The histological evaluation of testicular biop-sies was performed in all 77 patients, 62 withazoospermia, 14 with severe oligospermia, and
J Med Genet2001;38:798–802
Institute of HumanGenetics, PolishAcademy of Sciences, 32Strzeszynska, 60-479Poznan, PolandJ JaruzelskaA KorczA WojdaT SurmaczM Kotecki
Clinic of Infertility andReproductiveEndocrinology, MedicalAcademy, Poznan,PolandP JedrzejczakL Pawelczyk
Department ofHistology, MedicalAcademy, Poznan,PolandJ Bierla
Howard HughesMedical Institute,Whitehead Institute forBiomedical Researchand Department ofBiology, MassachusettsInstitute of Technology,Cambridge,Massachusetts, USAD C Page
Table 1 Frequency of diVerent histological phenotypes inpatients with spermatogenic failure. Testis biopsies wereevaluated in 77 patients. Ten patients found with Klinefeltersyndrome karyotype 47,XXY are included within thegroup diagnosed as Sertoli cell only syndrome
one with oligospermia. Histological abnormali-ties were detected in 63 patients and in 14patients the histology was normal. The histo-logical phenotypes are summarised in table 1.Among 14 patients with normal histology, 10were azoospermic, three severely oligospermic,and one was oligospermic.
AZF DELETIONS
Based on the PCR amplification of 23 STSmarkers specific to the Y chromosome (mostlyAZF region), deletions in six patients (IHG8,IHG18, IHG22, IHG67, IHG82, and IHG120)were detected (table 2). In two cases, IHG8 andIHG22, the deletions were large, terminal, andsimilar in size encompassing AZFb, AZFc, and
the heterochromatin region. Thus, they weredetectable cytogenetically. Four other deletions,IHG18, IHG67, IHG82, and IHG120, weresmall, interstitial, and similar in size andspanned the AZFc subregion (table 2).
In four cases (IHG22, IHG67, IHG82, andIHG120), deletions were found to be de novo,and all STS markers absent in the patients werepresent in their father. DNA samples of closemale relatives of patients IHG8 and IHG18were not available and therefore the de novostatus could not be tested.
KARYOTYPING
Patients IHG8, IHG22, and IHG120 weremosaic: They were carrying one cell line 46,XY
Table 2 Results of the genetic analyses of the patients: karyotyping of blood lymphocytes and PCR detection of AZF deletions using pairs of primerscomplementary to 23 Y specific STS markers
+ presence, − absence of STS.*The presence of CDY2 and the absence of CDY1 was detected by SSCP analysis of PCR product corresponding to the sY638 (data not shown). The sequence dif-ference enables SSCP distinguishing between CDY2 and CDY1 copies.
Table 3 Phenotypes of the five patients with spermatogenic failure carrying deletions within AZF region
Evaluation of the germinal epitheliumSpermatogonia ↓ ↓ ↓ ↓ ↓ ↓ N N ↓ NSpermatocytes 1st order ↓ ↓ Single Few ↓ ↓ N N ↓ ↓Spermatocytes 2nd order − − − − Single Single Single Single ↓ ↓Spermatides − − − − − − − − ↓ ↓Sperm cells − − − − − − − − ↓ ↓Sertoli cells N N N N N N N N N N
Evaluation of the interstitial tissueLeydig cells N N N N N N N N N NInflammatory infiltrates − − − − − − − − − −Fibrosis − − − − − − − − − −
Evaluation of extension of the lesionsFocal + +DiVused (generalised) + + + + + + + +
+ presence, − absence, N = normal appearance, ↓ = decrease in number or size, L = left, R = right testis. Patient IHG120 had azoospermia but no histological evalu-ation of the testis.
with a large terminal deletion (IHG8 andIHG22) or small interstitial deletion (IHG120)and a second cell line (45,X) lacking the entireY chromosome (table 2). Among patients withno AZF deletions, chromosomal aberrationswere detected in 30% of cases. In 18 subjects,aberrations of the sex chromosomes werefound including 10 47,XXY cases, whereasautosomal aberrations were present in 10males (Wojda et al, submitted).
GENOTYPE-PHENOTYPE CORRELATION
Both patients with large terminal deletions(IHG8 and IHG22) were found to haveazoospermia and complete maturation arrestlacking secondary spermatocytes, whereas threepatients with small interstitial AZFc deletions(IHG18, IHG67, and IHG82) had milder phe-notypes: azoospermia or severe oligospermiawith incomplete maturation arrest (IHG18 andIHG67) or hypospermatogenesis (IHG82)(table 3). In addition, the pattern of incompletematuration arrest of patient IHG18 was focaland not generalised (table 3). Despite significantrepresentation of the 45,X cell line, especially inpatients IHG8 and IHG22 (table 2), no gonadaldegeneration, short stature, webbing of the neck,lymphoedema, mental retardation, or any otherTurner stigmata, or other somatic abnormalities,such as genital ambiguity or gynaecomastia,were present in any of our six Y deleted patients.The height of the men with Y deletions was: 170cm (IHG8), 173 cm (IHG18), 168 cm(IHG22), 175 cm (HG67), 170 cm (IHG82),and 180 cm (IHG120).
DiscussionGENOTYPE-PHENOTYPE CORRELATION
Among 94 infertile but otherwise normalmales, we describe six with two types of AZFdeletions, large and terminal encompassingAZFb/AZFc/heterochromatin (IHG8, IHG22)and small and interstitial encompassing AZFc(IHG18, IHG67, IHG82, IHG120). In confir-mation of our hypothesis that the lack of corre-lation between the size of AZF deletions andthe phenotype of infertile men may be causedby coexistence of the AZF deletion with achromosomal aberration, we found a mosaic45,X cell line in three of five karyotyped AZFdeleted patients (a frequency of at least 50%).Several cases of infertile males with 45,X and46,XY cell lines with the abnormal Ychromosome were described before molecularstudies of male infertility were available.31–45
Among these cases, the aberrant Y chromo-some was described as being isodicentric,33 38–41
or a ring chromosome,31 32 44 45 or just lackingYq,1 34–37 40 42 43 as in the two cases, IHG8 andIHG22, described in this study. Most of thesepreviously described infertile males with 45,Xmosaic karyotypes and a non-fluorescent Yrarely had a “pure sterility” phenotype, but usu-ally manifested several somatic features, mostlyshort stature,35 36 42 43 gynaecomastia,34 36–38 orgenital ambiguity.42 43 In our study, however, allthree patients with a 45,X mosaic cell line hada pure sterility phenotype, which was the caseeven in those with a high proportion of 45,X(patients IHG8 and IHG22). Another patient
of this type was previously published and had45,X in 50% of blood cells, no somatic abnor-malities, and a height of 183 cm.39 However,that report did not describe the patient’stesticular histology, so phenotype-genotypecorrelation could not be analysed to the extentto which it was possible in our patients.
Pure 45,X chromosomal constitution isknown to result in incomplete ovarian develop-ment (streak gonads in the adult). Phenotypicvariability of these published 45,X mosaicinfertile patients, including our patients, isprobably the result of the tissue representationof 45,X mosaicism, which may to a varyingdegree aVect the gonadal development and dif-ferentiation processes.46 Similarly, 45,X mosai-cism could enhance the infertility phenotypecaused by a coexisting AZF deletion of anysize. Since the representation of the 45,X linein the target tissue, that is, primordial germcells or spermatogonia, may be diVerent fromthe one observed in blood cells, the phenotypeenhancement in AZFdel/45,X subjects cannotbe accurately evaluated, making the correlationof the size of the AZF deletion and the infertil-ity phenotype very diYcult. Thus, patientIHG120 with 45,X mosaicism should not becompared to those who have a similar sizeddeletion and no 45,X cell line in blood (IHG67and IHG82). Owing to the same limitations,45,X mosaic patients IGH8 and IGH22carrying large terminal deletions should not bedirectly compared to the rest of the patientscarrying smaller interstitial deletions.
However, this study does show that AZFcmicrodeletions alone may result in incompletematuration arrest in which some secondaryspermatocytes and normal appearing sperma-togonia are present (IHG67), or even in hypo-spermatogenesis (IHG82) in which all devel-opmental stages of germ cells are found,although in fewer numbers. We were able tofind this despite a broad spectrum of observedphenotypes (tables 1 and 3) and sizes of AZFdeletions (table 2), which were typically asso-ciated with AZFc deletions in other studies.5 6
Detection of such a clear genotype-phenotypecorrelation in our patients with isolated AZFcmicrodeletion was possible only when thekaryotypic defect was excluded or confirmedin the patients studied and a systematicbilateral histological evaluation accompaniedthe AZFc deletion molecular screening. Sinceinterphase FISH with Y probes in testicularsamples from AZF deleted patients was notperformed, we cannot exclude the possibilityof a 45,X cell line in the target tissue.However, testicular biopsy is an invasiveprocedure and therefore is limited to azoosper-mic patients when obtaining spermatozoa forICSI in vitro fertilisation.
The association of incomplete maturationarrest phenotype with isolated AZFc deletionsfound in this study could also indicate thatgenes encoded by AZFc, for example, DAZ orCDY1, are not crucial for the establishment ofthe germline stem cells, the spermatogonia.
ORIGIN OF 45,X CELL LINE MOSAICISM WITH AZFcMICRODELETIONS
It was previously reported that Y chromosomescarrying several types of cytogenetically detect-able aberrations, including the lack of Yq, havea tendency to be lost in the course of cell divi-sion and lead to the appearance of the 45,Xmosaic cell line.47 48 Therefore, it can beassumed that large AZF terminal deletions andthe mosaic 45,X cell line in patients IHG8 andIHG22 are not independent events, and thatthe 45,X line resulted from a loss of theaberrant Yq chromosome.
Interestingly, in patient IHG120, the mosaic45,X line coexists with an interstitial AZFcmicrodeletion. This finding seems to indicatethat smaller, cytogenetically undetectable mo-lecular defects could predispose to the loss ofthe entire Y chromosome too. So far, only asingle case of that type has been reported,20
although most papers describing patientsscreened for AZFc microdeletions did notinclude cytogenetic analysis, so 45,X mosai-cism could not be excluded. Taken together,our data (one patient out of three) and the onesof Oliva et al20 (one patient out of 10) show thatat least 15% of karyotyped patients with AZFcmicrodeletions do carry a mosaic 45,X cellline. This seems to be less frequent than thecoexistence of 45,X with the terminal deletionsof Yq (both patient IHG8 and IHG22), butfrequent significantly enough to be addressedin future or even retrospective studies.
In conclusion, our results underline theimportance of a combined molecular andkaryotypic approach as well as thorough histo-logical analysis for proper evaluation ofgenotype-phenotype correlation in patients withspermatogenic failure carrying AZFc deletions.
We thank Professor Helena Kedzia for helpful discussion on thehistological categorisation of the patients and Renata Matuszakfor technical help in screening for AZF deletions. This work wassupported by an International Collaborative Grant from theHoward Hughes Medical Institute 75195-543601 to MK andDCP and a grant from the Polish State Committee for ScientificResearch 4 PO5E 045 12 to AK.
1 Tiepolo L, ZuVardi O. Localization of factors controllingspermatogenesis in the nonfluorescent portion of the humanY chromosome long arm. Hum Genet 1976;34:119-24.
2 Ma K, Inglis JD, Sharkey A, Bickmore WA, Hill RE, ProsserEJ, Speed RM, Thomson EJ, Jobling M, Taylor K, Wolfe J,Cooke HJ, Hargreave TB, Chandley AC. A Y chromosomegene family with RNA-binding protein homology: candi-dates for the azoospermia factor AZF controlling humanspermatogenesis. Cell 1993;75:1287-95.
3 Reijo R, Lee TY, Salo P, Alagappan R, Brown LG,Rosenberg M, Rozen S, JaVe T, Straus D, Hovatta O, de laChapelle A, Silber S, Page DC. Diverse spermatogenicdefects in humans caused by Y chromosome deletionsencompassing a novel RNA-binding protein gene. NatGenet 1995;10:383-92.
4 Lahn BT, Page DC. Functional coherence of the human Ychromosome. Science 1997;278:675-80.
5 Reijo R, Alagappan RK, Patrizio P, Page DC. Severe oligo-zoospermia resulting from deletions of azoospermia factorgene on Y chromosome. Lancet 1996;347:1290-3.
6 Vogt PH. Human chromosome deletions in Yq11, AZFcandidate genes and male infertility: history and update.Mol Hum Reprod 1998;4:739-44.
7 Chandley AC. Chromosome anomalies and Y chromosomemicrodeletions as causal factors in male infertility. HumReprod 1998;suppl 1:45-50.
8 Brandell RA, Mielnik A, Liotta D, Ye Z, Veeck LL, PalermoGD, Schlegel PN. AZFb deletions predict the absence ofspermatozoa with testicular sperm extraction: preliminaryreport of a prognostic genetic test. Hum Reprod 1998;13:2812-15.
9 Chandley AC, Cooke HJ. Human male fertility - Y-linkedgenes and spermatogenesis. Hum Mol Genet 1994;3:1449-52.
10 Foresta C, Ferlin A, Garolla A, Rossato M, Barbaux S, DeBorcoli A. Y-chromosome deletions in idiopathic severetesticulopathies. J Clin Endocrinol Metab 1997;82:1075-80.
11 Foresta C, Ferlin A, Garolla A, Moro E, Pistorello M, Bar-baux S, Rossato M. High frequency of well-definedY-chromosome deletions in idiopathic Sertoli cell-onlysyndrome. Hum Reprod 1998;13:302-7.
12 Girardi SK, Mielnik A, Schlegel N. Submicroscopicdeletions in the Y chromosome of infertile men. HumReprod 1997;12:1635-41.
13 Grimaldi P, Scarponi C, Rossi P, Rocchietti March M, FabbriA, Isidori A, Spera G, Krausz C, Geremia R. Analysis of Yqmicrodeletions in infertile males by PCR and DNA hybridi-zation techniques. Mol Hum Reprod 1998;4:1116-21.
14 Kent-First MG, Kol S, Muallem A, Ofir R, Manor D, BlazerS, First N, Itskovitz-Eldor J. The incidence and possiblerelevance of Y-linked microdeletions in babies born afterintarcytoplasmic sperm injection and their infertile fathers.Mol Hum Reprod 1996;2:943-50.
15 Kobayashi K, Mizuno K, Hida A, Komaki R, Tomita K,Matsushita I, Namiki M, Iwamoto T, Tamura S, MinowadaS, Nakahori Y, Nakagome Y. PCR analysis of the Ychromosome long arm in azoospermic patients: evidencefor a second locus required for spermatogenesis. Hum MolGenet 1994;3:1965-7.
16 Liow SL, Ghadessy FJ, Ng SC, Yong EL. Y chromosomemicrodeletions, in azoospermic or near-azoospermic sub-jects, are located in the AZFc (DAZ) subregion. Mol HumReprod 1998;4:763-8.
17 Mulhall JP, Reijo R, Algappan R, Brown L, Page D, CarsonR, Oates RD. Azoospermic men with deletion of the DAZ
+ In infertile men with deletions within theY chromosome azoospermia factor re-gion (AZF), investigators have observedonly a weak correlation between the sizeof the deletions and the severity of theassociated spermatogenic defects. Wehypothesise that this might result from acoexistence of the AZF deletion with achromosomal aberration.
+ A thorough bilateral analysis of testicularhistology combined with genetic testsincluding blood karyotyping and screen-ing for AZF deletions was performed in94 patients with non-obstructive infertil-ity, including 65 azoospermic, 23 severelyoligospermic, and six oligospermic men.
+ Abnormalities of the AZF region wereidentified in six patients: large terminaldeletions of AZFb/AZFc/heterochromatin(two patients) and small interstitial dele-tions of just AZFc (four patients).Surprisingly, in five of these AZF deletedmales who were karyotyped, a second cellline, 45,X, was found in three of them (afrequency of at least 50%). Interestingly,one patient had a 45,X line coexistingwith just an interstitial AZFc deletion,indicating mosaic loss of an entire Ychromosome secondary to this deletion.The two patients with large deletions anda coexisting 45,X cell line had bothazoospermia and complete maturationarrest with no secondary spermatocytes,whereas patients with just AZFc dele-tions and a pure 46,XY karyotype hadless severe phenotypes, namely azoosper-mia or severe oligospermia and incom-plete maturation arrest with the presenceof secondary spermatocytes or hyposper-matogenesis.
+ These results show the importance of acombined histological, cytogenetic, andmolecular approach, which in this studyallowed the observation of an associationbetween the incomplete maturation arrestphenotype and isolated AZFc deletions.
gene cluster are capable of completing spermatogenesis:fertilization, normal embryonic development and preg-nancy occur when retrieved testicular spermatozoa areused for intracytoplasmic sperm injection. Hum Reprod1997;12:503-8.
18 Najmabadi H, Huang V, Yen P, Subbarao MN, Bhasin D,Banaag L, Naseeruddin S, De Kretser DM, Gordon BakerHW, McLachlan RI, Loveland KA, Bhasin S. Substantialprevalence of microdeletions of the Y-chromosome in infer-tile men with idiopathic azoospermia and oligozoospermiadetected using a sequence-tagged site-based mappingstrategy. J Clin Endocrinol Metab 1996;81:1347-52.
19 Nakahori Y, Kuroki Y, Komaki R, Kondoh N, Namiki M,Iwamoto T, Toda T, Kobayashi K. The Y chromosomeregion essential for spermatogenesis. Horm Res 1996;46:20-3.
20 Oliva R, Margarit E, Ballesca JL, Carrio A, Sanchez A, MilaM, Jimenez L, Alvarez-Vijande JR, Ballesta F. Prevalence ofY chromosome microdeletions in oligospermic andazoospermic candidates for intracytoplasmic sperm injec-tion. Fertil Steril 1998;70:506-10.
21 Qureshi SJ, Ross AR, Cooke HJ, Intyre MAM, ChandleyAC, Hargreave TB. Polymerase chain reaction screeningfor Y chromosome microdeletion: a first step towards thediagnosis of genetically-determined spermatogenic failurein men. Mol Hum Reprod 1996;2:775-9.
22 Pryor JL, Kent-First M, Muallem A, Van Bergen AH, NoltenWE, Meisner L, Roberts KP. Microdeletions in the Y chro-mosome of infertile men. N Engl J Med 1997;336:534-9.
23 Silber SJ, Alagappan R, Brown LG, Page DC. Ychromosome deletions in azoospermic and severely oligo-zoospermic men undergoing intracytoplasmic sperm injec-tion after testicular sperm extraction. Hum Reprod 1998;13:3332-7.
24 Simoni M, Gromol J, Dworniczak B, Rolf C, Abshagen K,Kamischke A, Carani C, Meschede D, Behre HM, Horst J,Nieschlag E. Screening for deletions of the Y chromosomeinvolving the DAZ (deleted in azoospermia) gene inazoospermia and severe oligospermia. Fertil Steril 1997;67:542-7.
25 Stuppia L, Calabrese G, Guanciali Franchi P, Mingarelli L,Gatta V, Palka G, Dallapiccola B. Widening of theY-chromosome interval 6 - deletion transmitted from thefather to his infertile sun accounts for an oligozoospermiacritical region distal to the RBM1 and DAZ genes. Am JHum Genet 1996;59:1393-5.
26 Stuppia L, Gatta V, Mastroprimiano G, Pompetti F,Calabrese G, Guancialli Franchi P, Morizio E, MingarelliR, Nicolai M, Tenaglia R, Improta L, Sforza V, Bisceglia S,Palka G. Clustering of Y chromosome deletions insubinterval E of interval 6 supports the existence of an oli-gozoospermia critical region outside the DAZ gene. J MedGenet 1997;34:881-3.
27 Stuppia L, Gatta V, Calabrese G, Guanciali Franchi P,Morizio E, Bombieri C, Mingarelli R, Sforza V, Frajese G,Tenaglia R, Palka G. A quarter of men with idiopathicoligo-azoospermia display chromosomal abnormalities andmicrodeletions of diVerent types in interval 6 of Yq11.Hum Genet 1998;102:566-70.
28 van de Ven K, Montag M, Peschka B, Leygraaf J, SchwanitzG, Haidl G, Krebs D, van der Ven H. Combinedcytogenetic and Y chromosome microdeletion screening inmales undergoing intracytoplasmic sperm injection. MolHum Reprod 1997;3:699-704.
29 Vereb M, Agulnik AI, Houston JT, Lipschultz LI, Lamb DJ,Bishop CE. Absence of DAZ gene mutations in cases ofnon-obstructed azoospermia. Mol Hum Reprod 1997;3:55-9.
30 Vogt PH, Edelmann A, Kirsch S, Henegariu O, HischmannP, Kiesewetter F, Koehn FM, Schill WB, Farah S, RamosC, Hartmann M, Hartschuh W, Meschede D, Behre HM,
Castel A, Nieschlag E, Weidner W, Groene HJ, Jung A,Engel W, Haidl G. Human Y chromosome azoospermiafactors (AZF) mapped to diVerent subregions in Yq11.Hum Mol Genet 1996;5:933-43.
31 Chandley AC, Edmond P. Meiotic studies on a subfertilepatient with a ring Y chromosome. Cytogenetics 1971;10:295-304.
32 Maeda T, Ohno M, Ishibashi A, Samejima M, Sasaki K.Ring Y chromosome: 45,X/46,Xr(Y) chromosome mosai-cism in a phenotypically normal male with azoospermia.Hum Genet 1976;34:99-102.
33 Haaf T, Schmid M. Y isochromosome associated with amosaic karyotype and inactivation of the centromere. HumGenet 1990;85:486-90.
34 Interlandi JW, Russell MH, Kirchner F, Rabin D. Geneticand endocrine findings in a 31-year-old 45,X/46,Xdel(Y)(q12) male. J Clin Endocrinol Metab 1981;53:1047-55.
35 Kaluzewski B, Jokinen A, Hortling H, de la Chapelle A. Atheory explaining the abnormality in 45,X/46,XY mosai-cism with non-fluorescent Y chromosome: presentation ofthree cases. Ann Genet 1978;21:5-11.
36 Beverstock GC, MacFarlane JD, Veenema H, Hoekman H,Goodfellow PJ. Y chromosome specific probes identifybreakpoint in a 45,X/46,X,del(Y)(pter-q11.1:) karyotypeof an infertile male. J Med Genet 1989;26:330-3.
37 Madan K, Gooren L, Schoemaker T. Three cases of sexchromosome mosaicism with a nonfluorescent Y. J HumGenet 1979;46:295-304.
38 Taylor MC, Gardner HA, Ezrin C. Isochromosome for thelong arm of the Y in an infertile male. Hum Genet 1978;40:227-30.
39 Smith A, Conway A, Robson L. The use of fluorescencein-situ hybridisation to clarify abnormal Y chromosomes intwo infertile men. Med J Aust 1994;160:545, 548-9, 552.
40 Kaluzewski B, Jakubowski L, Debiec-Rychter M, GrzeschikKH, Limon J, Gibas Z. Two mosaic cases with nonfluores-cent Y chromosome analysed with Y-specific DNA probes.Am J Med Genet 1988;31:489-503.
41 Chandley AC, Ambros P, McBeath S, Hargreave TB,Kilanowski F, Spowart G. Short arm dicentric Y chromo-some with associated statural defects in a sterile man. HumGenet 1986;73:350-3.
42 Ganshirt D, Pawlowitzki IH. Hae III restriction of DNAfrom three cases with nonfluorescent Y chromosomes(45XO/46XYnf) Hum Genet 1984;67:241-4.
43 Ganshirt-Ahlert D, Pawlowitzki IH, Gal A. Three cases of45,X/46,XYnf mosaicism. Molecular analysis revealedheterogeneity of the nonfluorescent Y chromosome. HumGenet 1987;76:153-6.
44 Chandley AC, Edmond P, Christie S, Gowans L, Fletcher J,Frackiewicz A, Newton M. Cytogenetics and infertility inman. I. Karyotype and seminal analysis: results of afive-year survey of men attending a subfertility clinic. AnnHum Genet 1975;39:231-54.
45 Chandley AC, Maclean N, Edmond P, Fletcher J, WatsonGS. Cytogenetics and infertility in man. II. Testicularhistology and meiosis. Hum Genet 1976;40:165-76.
46 Rosenberg C, Frota-Pessoa O, Vianna-Morgante AM, ChuTH. Phenotypic spectrum of 45,X/46,XY individuals. AmJ Med Genet 1987;27:553-9.
47 Hsu LYF. Phenotype/karyotype correlations of Y chromo-some aneuploidy with emphasis on structural aberrationsin postnatally diagnosed cases. Am J Med Genet 1994;53:108-40.
48 Yoshida A, Nakahori Y, Kuroki Y, Motoyama M, Araki Y,Miura K, Shirai M. Dicentric Y chromosome in anazoospermic male. Mol Hum Reprod 1997;3:709-12.
Alternative centromeric inactivation in apseudodicentric t(Y;13)(q12;p11.2) translocationchromosome associated with extremeoligozoospermia
Jean Pierre SiVroi, Brigitte Benzacken, Roxani Angelopoulou, Corine Le Bourhis,Isabelle Berthaut, Samia Kanafani, Asmae Smahi, Jean Philippe Wolf, Jean Pierre Dadoune
EDITOR—Centromeres are the specialised re-gions of chromosomes that ensure normaltransmission of sister chromatids to each daugh-ter cell after mitosis. Alphoid satellite DNAsequences, consisting of tandemly repeated
≅170 bp units present at all human centromeres,contain the information necessary for centro-meric function,1 despite the observation ofmarker chromosomes lacking detectable alphoidDNA.2–4 Dicentric chromosomes, resulting from
some Robertsonian or Y;autosome transloca-tions, represent a valuable tool for studying fac-tors which ensure that only one of the centro-meres is mitotically active, thus preventingchromosomal bridges and breakages to occur atanaphase. It has been shown that centromericinactivation is largely an epigenetic event5 basedon the ability of alphoid sequences to bind spe-cific centromeric proteins (CENPs), particularlythe CENP-C protein which is necessary forproper kinetochore assembly.6
Here we describe a de novo dicentric Y;13(q12;p11) translocation chromosome found ina severely oligozoospermic patient and exhibit-ing a variable pattern of centromeric activity, asdefined by the localisation of the primary con-striction.
Methods and resultsThis patient was a healthy, 20 year old, WestIndian man who referred himself to the labora-tory because of ejaculation problems. Spermanalysis showed first an abnormal viscosity ofejaculate which took as long as six hours to lique-fy and, second, an extreme oligozoospermia at0.1 million spermatozoa/ml. Biochemicalparameters of the semen were normal. Furthersperm counts showed a similar constitution ofejaculate and testicular impairment varyingfrom severe oligozoospermia to azoospermia orcryptozoospermia. Testicular biopsy was notproposed.
Karyotyping was performed on blood lym-phocytes by conventional cytogenetic methodsusing R and G banding and BrdU incorpora-tion after cell culture synchronisation. Itshowed an apparently balanced reciprocaltranslocation between the distal region of the Ychromosome long arm and the short arm ofone chromosome 13. Paternal chromosomeswere normal.
Curiously, this abnormal chromosome ex-hibited two diVerent features with regard to theposition of its primary constriction. In 50 cellsanalysed, this was localised at the Y centromerein about half the cells, giving a characteristicaspect of a large acrocentric chromosome, andat the chromosome 13 centromere in others,leading to an abnormal metacentric rearrangedchromosome (fig 1). On the basis of thesemorphological data, the translocated chromo-some was considered as pseudodicentric. Cbanding indicated that Y heterochromatin wasapparently preserved and included chromo-some 13 centromere labelling while the Y cen-tromere was normally present (fig 1). Silverstaining failed to detect any nucleolar organiserregion (NOR) on the translocated chromo-some (data not shown). Fluorescence in situhybridisation (FISH), using Y and chromo-some 13 specific probes, confirmed cytogeneticresults (data not shown). Thus, chromosomalbreakpoints were localised in Yq12 and13p11.2 and the proband’s karyotype was45,X,−13,−Y, +psu dic(Y),t(Y;13)(q12;p11.2)[27]/45,X,−13,−Y,+psu dic(13),t(13;Y)(p11.2;q12) [23]. No normal cell line wasobserved.
After genomic DNA extraction, PCR reac-tions were performed for checking the integrity
of the Y chromosome euchromatic region.Twelve diVerent STSs (sequence tagged sites),corresponding to the three AZF loci (azoosper-mia factor, AZFa: sY85, sY95; AZFb: sY114,sY116, sY125, sY127; AZFc: sY135, sY149,sY152, sY254), to SRY (sY14), and to the het-erochromatic distal Yq region (sY160), wereamplified and gave positive results in theproband’s DNA samples, indicating the ab-sence of interstitial microdeletion (data notshown).
Because the distal part of the Y chromosomelong arm is diYcult to analyse by conventionalcytogenetic techniques, both molecular andFISH approaches were used for a bettercharacterisation of the chromosomal break-points. However, FISH alone could givepositive results if this breakpoint was localisedwithin the region recognised by the probe. Amolecular polymorphic marker (DXYS154),corresponding to the pseudoautosomal region2 (PAR2), localised at the tip of sex chromo-some long arms, as well as two autosomalmarkers (D7S1779 and D14S983), were am-plified both in our patient and his father andcompared with each other (maternal bloodsample was not available). These three markersare dinucleotide repeats, (CA)n, and wereamplified by PCR, separated on a 6%polyacrylamide/urea gel, then transferred to anylon membrane, and hybridised to a labelledGT probe (ECLTM, Amersham PharmaciaBiotech). Results showed that the probandreceived a copy of one of each autosomalmarkers from his father and only one copy ofthe DXYS154 marker, which was diVerent insize from the paternal one (fig 2A), thusindicating the lack of PAR2 on the abnormal Ychromosome. This was confirmed by FISHusing a cosmid probe of this region which indi-cated that only the X chromosome exhibited afluorescent signal in our patient whereas bothgonosomes were labelled in paternal met-aphases (fig 2B, C). Therefore, the transloca-tion breakpoint in the Y chromosome waslocalised in the distal part of the heterochro-matic region (DYZ1).
Functional activity of one or both centro-meres was investigated using an antibodyagainst the kinetochore associated proteinCENP-C (a gift from Professor W C Earn-shaw, University of Edinburgh, Scotland). Forthis purpose, freshly prepared chromosomes
J Med Genet2001;38:802–806
Service d’Histologie,Biologie de laReproduction etCytogénétique, HôpitalTenon, 4 rue de laChine, 75020 Paris,France, and Laboratoirede Cytologie-Histologie,Centre Universitairedes Saints Pères, 45 ruedes Saints Pères, 75270Paris, FranceJ P SiVroiC Le BourhisI BerthautS KanafaniJ P Dadoune
Service d’Histologie,Embryologie,Cytogénétique etBiologie de laReproduction, HôpitalJean Verdier, 93 Bondy,FranceB BenzackenJ P Wolf
Laboratory of Histologyand Embryology,University of Athens,Medical School, Athens,GreeceR Angelopoulou
Département deGénétique, INSERMU393, HôpitalNecker-EnfantsMalades, Paris, FranceA Smahi
Figure 1 Partial view of proband’s karyotype after Rbanding (A, B, D) and C banding (C, E) showing thevariable aspect of the translocation chromosome accordingto centromeric activation (arrows). (A) Normalchromosome 13. (B) Translocation chromosome with the Ychromosome active centromere giving a large acrocentricabnormal chromosome. (C) View after C banding. (D)Chromosome 13 active centromere leading to a metacentrictranslocation chromosome. (E) View after C banding.
were first hybridised with a Y centromericfluorescent probe (Oncor, USA), labelled withrhodamine, and observed under UV light foridentifying precisely the Y chromosome inmitosis and the position of its primaryconstriction. The localisation of each mitosisobserved was then carefully recorded. Afterseveral washes in PBS and TEEN buVer (0.2mmol/l EDTA, 25 mmol/l NaCl, 1.0 mmol/ltriethanolamine, 0.5 % Triton, 0.1 % BSA),slides were allowed to incubate with the rabbitanti-CENP-C antibody, diluted 1/1000, forone hour at room temperature. They were thenrinsed three times in KB buVer (10 mmol/l TrisHCl, pH 7.7, 150 mmol/l NaCl, 0.1% BSA)
and incubated with a biotinylated goat anti-rabbit Ig, diluted 1/500, for 30 minutes at37°C. After washing in KB buVer, immunola-belling was performed by incubating slideswith FITC conjugated avidin. Mitoses, whichhad been studied by FISH, were reanalysedand immunolabelling of CENP-C was com-pared to the position of the primary constric-tion. This method was reliable and led to aslight decrease of immunolabelling intensity.
The results showed clearly that proteinCENP-C labelling is exclusively localised at thesite of the primary constriction in the rear-ranged chromosome, thus indicating a variableactivation of one of the centromeres in theproband’s cells (fig 3).
DiscussionApparently balanced Y;autosome transloca-tions can be found either in fertile7 8 orinfertile9 10 patients with, in some cases, pheno-typic diVerences between carriers of the sametranslocation in a family.11 12 Usually, infertilityis the consequence of a Y chromosome break-point occurring in the euchromatic long armsegment and leading to the loss of genes impli-cated in the azoospermia factor (AZF). How-ever, in cases with an apparently intact Y chro-mosome translocated onto an autosome,spermatogenetic impairment is thought toresult from abnormal meiotic behaviour oftranslocated chromosomes, which interact withthe XY body in most germ cells.10 Spreading ofX chromosome inactivation to autosomalsegments or an abnormal sex vesicle constitu-tion are the main explanations for meiotic fail-ure and spermatocyte degeneration. In ourcase, the Y chromosome breakpoint was local-ised at the end of the heterochromatic regionand integrity of the euchromatic long arm seg-ment was ascertained by molecular analysis.Therefore, despite the lack of testicular biopsyin our patient, oligozoospermia was probablythe consequence of these meiotic events,although it was not possible to explain theabnormal constitution of his semen. Loss ofPAR2, which is observed in infertile mencarrying a Y chromosome terminal deletionbut not in those with an interstitial one, isunlikely to be responsible for impairment ofspermatogenesis in our patient.
More interesting was the variable nature ofthe translocated chromosome in relation to theposition of its primary constriction. Such apolymorphism is very rarely found duringcytogenetic investigations but a case similar toours has already been described in a child withcongenital malformations owing to the exist-ence of an additional isochromosome 13q in23% of cells in blood and 5% in skin.13 Variablecentromeric activity has also been observed inan infertile patient carrying a t(Y;14) transloca-tion chromosome.14
However, the occurrence of dicentric chro-mosomes is a common event in Robertsoniantranslocations, whole arm reciprocal transloca-tions, or in structurally abnormal chromo-somes like isochromosomes. Mitotic stability ofsuch chromosomes implies that only onecentromere is active, which has been related to
Figure 2 Molecular characterisation of the translocationbreakpoint. (A) (a) Amplification of a polymorphicmarker localised in PAR2 in the father (F) and proband(P). Two bands corresponding to X and Y chromosomeswere present in the father’s DNA while only the maternalcopy of a diVerent size was amplified from the proband’sDNA, thus indicating the lack of PAR2 on his Ychromosome. (b, c) Amplification of autosomal polymorphicmarkers. The proband received a paternal copy of eachmarker. (B, C) Localisation of chromosomal breakpoints byFISH using a cosmid probe coding for the PAR2 region.Both gonosomes were labelled at the tip of their long arm inpaternal metaphases (B, thick arrows) whereas, in theproband, only the X chromosome exhibited a fluorescentsignal (C, thin arrow). Arrowhead indicates the unlabelledtranslocation chromosome).
the specific binding of the centromeric proteinCENP-C at this site.15 Indeed, while inactivecentromeres retain their ability to bind somecentromeric proteins like CENP-B, kineto-chore assembly and centromeric activationrequire at least CENP-C and/or CENP-Ebinding.16 Molecular analysis of a de novodicentric Y;21 translocation chromosome, withseveral clones exhibiting variable patterns ofcentromeric activation like our case, has shownmultiple forms of alphoid DNA deletions ofthe Y centromere.5 However, in this latter case,deletions were not systematically associatedwith Y centromeric inactivation. These resultsindicate that the centromeric activation/inactivation process is largely dependent onepigenetic factors but that it can occasionallybe associated with changes in alphoid DNAstructure.
The physical distance between the twocentromeres in a dicentric chromosome may bean additional factor for predisposing cells toinactivate one of the centromeric structures.Dicentric Xq chromosomes, in which centro-meres were separated by 4-12 Mb of Xpmaterial, have been shown to bind CENP-C onboth centromeres in most cells and, therefore, to
present a high degree of coordination betweenthe two sets of active kinetochores at mitosis.17
Such mitotic behaviour diVers from that ob-served in dicentric X chromosomes with wellseparated centromeres, by 34 Mb or more, inwhich one of them is systematically inactivated.17
However, in these cases, the symmetricalappearance of the abnormal chromosomes didnot allow the authors to determine if the inacti-vation process always occurred at the same cen-tromere or alternately at one or the other.
The same question arises from the analysisof dicentric chromosomes in which centro-meres are very close together. This may be thecase in some autosomal translocation chromo-somes like those involving acrocentric pairs.Mitotic instability generated by the activationof both centromeres would lead to the loss ofautosomes and to the death of the abnormalcell lines in which such an event occurred. Thesurviving clones would be only those in whichone centromere has been inactivated regardlessof its nature and the fact that it may be alwaysthe same one or not. Indeed, the centromerevicinity makes a possible alternative inactiva-tion of one of them indistinguishable bycurrent cytogenetic methods and immuno-chemical techniques. Only dicentric transloca-tion chromosomes with a non-symmetricalappearance and distant centromeres, like thatof our patient, would allow an alternative inac-tivation process to be diagnosed.
In conclusion, the case described hereemphasises the close relationship betweenstructural chromosomal rearrangements, espe-cially those involving the Y chromosome, andmale infertility. Moreover, it underlines an epi-genetic mechanism of alternative centromericinactivation which may be a common phenom-enon in dicentric chromosomes.
The authors thank Professor W C Earnshaw for providing anti-body to CENP-C and Mr Ph N’Guyen for artwork. This workwas supported by grants from AP-HP (CRC 96053 and PHRCAOM96142).
1 Tyler-Smith C, Oakey RJ, Larin Z, Fisher RB, Crocker M,AVara NA, Ferguson-Smith MA, Muenke M, ZuVardi O,Jobling MA. Localization of DNA sequences required forhuman centromere function through an analysis ofrearranged Y chromosomes. Nat Genet 1993;5:368-75.
2 Voullaire LE, Slater HR, Petrovic V, Choo KHA. Afunctional marker centromere with no detectable alpha-satellite, satellite III, or CENP-B protein: activation of alatent centromere? Am J Hum Genet 1993;52:1153-63.
3 Blennow E, Telenius H, de Vos D, Larsson C, Henriksson P,Johansson O, Carter NP, Nordenskjold M. Tetrasomy 15q:two marker chromosomes with no detectable alpha-satelliteDNA. Am J Hum Genet 1994;54:877-83.
4 Bukvic N, Susca F, Gentile M, Tangari E, Ianniruberto A,Guanti G. An unusual dicentric Y chromosome with afunctional centromere with no detectable alpha-satellite.Hum Genet 1996;97:453-6.
5 Fisher AM, Al-Gazali L, Pramathan T, Quaife R, CockwellAE, Barber JCK, Earnshaw WC, Axelman J, Migeon BR,Tyler-Smith C. Centromeric inactivation in a dicentrichuman Y;21 translocation chromosome. Chromosoma1997;106:199-206.
6 Tomkiel JE, Cooke CA, Saitoh H, Bernat RL, EarnshawWC. CENP-C is required for maintaining proper kineto-chore size and for a timely transition to anaphase. J Cell Biol1994;125:531-45.
7 Laurie DA, Palmer RW, Hulten MA. Studies on chiasmafrequency and distribution in two fertile men carryingreciprocal translocations; one with a t(9 ;10) karyotype andone with a t(Y;10) karyotype. Hum Genet 1984;68:235-47.
8 Eliez S, Morris MA, Dahoun-Hadorn S, DeLozier-BlanchetCD, Gos A, Sizonenko P, Antonarakis SE. Familial translo-cation t(Y;15) (q12;p11) and de novo deletion of thePrader-Willi syndrome (PWS) critical region on 15q11-q13. Am J Med Genet 1997;70:222-8.
9 Maraschio P, Tupler R, Dainotti E, Cortinovis M, TiepoloL. Molecular analysis of a human Y;1 translocation in anazoospermic male. Cytogenet Cell Genet 1994;65:256-60.
Figure 3 Binding of CENP-C protein to the active centromere. The translocationchromosome was identified by DAPI staining and FISH with a Y centromeric probeshowing either two well separated fluorescent spots or a more compact signal according tochromosome 13 (top) or Y (bottom) centromere activation. Labelling with an antibody toCENP-C (right) was always localised at the site of the primary constriction on therearranged chromosome (white arrows).
10 Delobel B, Djlelati R, Gabriel-Robez O, Croquette MF,Rousseaux-Prevost R, Rousseaux J, Rigot JM, Rumpler Y.Y-autosome translocation and infertility: usefulness ofmolecular, cytogenetic and meiotic studies. Hum Genet1998;102:98-102.
11 Doneda L, Magnani I, Tibiletti MG, Dalpra L, Larizza L.DiVerent phenotypes in two cases of an apparently identicalfamilial (Yq;13p) translocation. Hum Reprod 1992;7:495-9.
12 Teyssier M, Rafat A, Pugeat M. Case of (Y;1) familial trans-location. Am J Med Genet 1993;46:339-40.
13 Ing PS, Smith SD. Cytogenetic studies of a patient withmosaicism of isochromosome 13q and a dicentric (Y;13)translocation showing diVerential centromeric activity. ClinGenet 1983;24:194-9.
14 Gentile M, Cariola F, Buonadonna AL, Caroppo E, DiCarlo A, Carone D, D’Amato G. Alternate centromereinactivation in a dicentric (Y;14) (q12;p11) associated withazoospermia. Hum Reprod 2000;15:P-214.
15 Earnshaw WC, Ratrie H, Stetten G. Visualization ofcentromere proteins CENP-B and CENP-C on a stabledicentric chromosome in cytological spreads. Chromosoma1989;98:1-12.
16 Sullivan BA, Schwartz S. Identification of centromeric indicentric Robertsonian translocations: CENP-C andCENP-E are necessary components of functional centro-meres. Hum Mol Genet 1995;4:2189-97.
17 Sullivan BA, Willard HF. Stable dicentric X chromosomeswith two functional centromeres. Nat Genet 1998;20:227-8.
7th European Forum on Quality Improvement in Health Care
21–23 March 2002
Edinburgh, Scotland
We are delighted to announce this forthcoming conference in Edinburgh. Authors are invitedto submit papers (call for papers closes on Friday 5 October 2001) and delegate enquiries arewelcome.
The themes of the Forum are:
x Leadership, culture change, and change managementx Achieving radical improvement by redesigning carex Health policy for lasting improvement in health care systemsx Patient safetyx Measurement for improvement, learning, and accountabilityx Partnership with patientsx Professional quality: the foundation for improvementx Continuous improvement in education and trainingx People and improvement.
Presented to you by the BMJ Publishing Group (London, UK) and Institute for HealthcareImprovement (Boston, USA). For more information contact: [email protected] or look atthe website www.quality.bmjpg.com. Tel: +44 (0)20 7383 6409; fax: +44 (0)20 7373 6869.
