University of Groningen

Application of molecular techniques in population genetic studies of cystic fibrosis in theNetherlandsde Vries, Hendrik Gerardus

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:1996

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):de Vries, H. G. (1996). Application of molecular techniques in population genetic studies of cystic fibrosis inthe Netherlands. [S.l.]: [S.n.].

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 01-02-2020

Application of molecular techniques in population

genetic studies of cystic fibrosis in The Netherlands

Stichting Drukkerij Regenboog, Groningen

This study was supported by a grant from The Netherlands Organization for Scientific

Research (NWO). Financial support to publish this thesis was provided by the NWO and

by the "Stichting voor erfelijkheidsvoorlichting te Groningen".

RIJKSUNIVERSITEIT GRONINGEN

Application of molecular techniques in population

genetic studies of cystic fibrosis in The Netherlands

Proefschrift

ter verkrijging van het doctoraat in de

Geneeskunde

aan de Rijksuniversiteit Groningen

op gezag van de

Rector Magnificus Dr. F. van der Woude

in het openbaar te verdedigen op

woensdag 31 januari 1996

des namiddags te 2.45 uur precies

door

Hendrik Gerardus de Vries

geboren op 14 april 1965

te Roosendaal

Promotores: Prof.dr. L.P. ten Kate

Prof.dr. C.H.C.M. Buys

Referent: Dr. H. Scheffer

’Kom verder,’ zei hij (professor Sickbock) vriendelijk. ’U treft het. Ik ben net bezig

aan een onderzoek van de mutaties in de genen.’

’Van de wat?’ vroeg heer Ollie, bedremmeld om zich heen kijkend.

’Laat maar,’ hernam de professor. ’Ik heb ontdekt, dat alles wat de natuur ons

voorschotelt beuzelachtig is. Er bestaat niets, dat door een scherp geleerde niet kan

worden verbeterd. Volgt u me?’

(Marten Toonder; Een heer moet alles alleen doen; De toornviolen. De Bezige Bij,

Amsterdam, 1969)

Promotiekommissie: Prof.dr. J.J. Cassiman

Prof.dr. A. Hofman

Prof.dr. H.S.A. Heymans

Contents

Outline of this thesis

Chapter 1. General introduction

1.1 Delineation of and research into cystic fibrosis 11

1.2 The CFTR gene and its mutations 12

1.3 Possibilities for genetic modification

A. Mouse models 15

B. Somatic gene therapy for cystic fibrosis 17

1.4 Cystic fibrosis prevalence at birth 18

1.5 What causes the high frequency of cystic fibrosis ? 19

1.6 Screening for cystic fibrosis carriers 20

1.7 Identity By Descent (IBD) analysis 21

References 23

Chapter 2 Validation of determination of∆F508 mutations of the cystic 33

fibrosis gene in over 11,000 mouthwashes.

Chapter 3 Prevalence of∆F508 cystic fibrosis carriers in The Netherlands: 41

region of residence, age, and number of offspring as explanatory

variables.

Chapter 4 Relative frequencies of CFTR mutations in The Netherlands: 57

Regional variation is present even in a small country.

Chapter 5 Number and sex of offspring of∆F508 carriers outside cystic 67

fibrosis families

Chapter 6 Haplotype identity between individuals who share a CFTR 71

mutation allele Identical By Descent: demonstration of the

usefulness of the haplotype sharing concept for gene mapping

in real populations

Summary 85

Samenvatting 89

Curriculum vitae 94

Dankwoord 95

Outline of this thesis

Cystic fibrosis (CF) is a common autosomal recessive disease with an estimated

birth prevalence of about one in 2500 in Caucasian populations. Since the cloning

and characterizing of the CF gene in 1989, numerous reports have been published

on the detection of new mutations. Till now more then 500 mutations have been

described and this number is still increasing. The most common CF mutation is the

∆F508 mutation with an estimated relative frequency of 70% in Europe. The

median life expectancy of CF patients has strongly increased to 28-40 years. Since

1991 the genetic modification is intensively investigated in mouse-models and in

somatic gene therapy studies.

The birth prevalences of CF in European populations and populations of

European descent elsewhere show a range from 1 in 1700 to 1 in 7700 (excluding

Finland where CF is extremely rare). Heterozygote advantage is the most likely

explanation for the maintenance of the high CF gene frequency at current levels.

The feasibility and appropriateness of offering carrier testing to the general

population was and still is the topic for extensive discussions. This resulted in a

number of arguments against and in favour of population screening. The American

Society of Human Genetics and the National Institutes of Health stated that carrier

screening would be desirable only if 90-95% of the mutations can be detected.

The distribution of numerous less common CF mutations is limited to certain

countries or even population groups within countries. It has been well described that

some specific mutations have achieved a high frequency in certain founder

populations. An example is the A455E mutation in a French Canadian population. It

is very likely that patients with this mutation have a common not to far predecessor.

The study of the DNA of patients with this mutation has been used to demonstrate

the general feasibility of an Identity By Descent (IBD) approach.

The application of molecular techniques strongly enhances the identification of

CF patients. Direct mutation detection can also be used for epidemiologic studies

and population-based carrier screening. One of the purposes of this study was to

analyse the usefulness of a mouthwash procedure for the detection of a large

numbers of CF carriers. Chapter 2 describes the determination of the failure rate,

specificity and sensitivity of the mouthwash procedure. The samples collected for

9

Chapter 1

General introduction

1.1 Delineation of and research into cystic fibrosis

Cystic fibrosis (CF) is the most common lethal autosomal recessive disorder

affecting Caucasian populations, with a generally quoted birth prevalence of one in

2500 [1].

Table 1. Historical perspectives of cystic fibrosis research.

1936 Report of a child with the clinical features of cystic fibrosis. [2]

1938 First detailed description of the clinical features and pathology of CF: "Cystic fibrosis" [3]of the pancreas.

1945 Recognition of the role of sticky mucus as a cause of many of the symptoms: "Mucoviscidosis". [4]

1946 Suggestion of an autosomal recessive inheritance pattern for CF. [5]

1953 Investigation of acute salt loss caused by excessive sweating in babies with CF, during a [6]heat-wave.

1959 Development of the pilocarpine iontophoresis method for sweat testing. [7]

1981 Description of altered electrical properties of CF respiratory epithelium associated with [8]abnormalities of both sodium and chloride transport.

1983 Documentation of chloride impermeability of CF sweat gland ducts. [9]

1985 Mapping of the CF gene to a small segment of the long arm of chromosome 7q, using [10]restriction fragment length polymorphisms (RFLP).

1987 Linkage disequilibrium between the markers XV2C/KM19 and cystic fibrosis was observed. [11]

1989 Cloning of the CF gene and description of a cDNA sequence. [12]

1989 Identification of the∆F508 mutation in approximately 70 percent of the mutations. [13]

1991 Report of the total genomic sequence of the CFTR gene. [14]

1994 Description of more then 500 different CF mutations. [15]

11

The disease is characterized by abnormal secretions of the exocrine glands which

cause chronic obstruction and infection of the respiratory tract, meconium ileus,

pancreatic insufficiency, and elevated levels of sweat electrolytes. Rochholz

(Almanac of Children’s Songs and Games from Switzerland, 1857) described " The

child will soon die whose brows tastes salty when kissed". Table 1 shows important

data on the delineation of and research into CF during the past 60 years.

1.2 The CFTR gene and its mutations

The gene responsible for cystic fibrosis (CF) has been identified in 1989 [12,13,16].

It is located on the long arm of chromosome 7, region 7q31.2 [17]. The gene is

approximately 230,000 kb in size and contains 27 exons [14]. The encoded mRNA

is about 6.5 Kb long. The gene codes for the 1480 amino acid cystic fibrosis

transmembrane conductance regulator (CFTR) protein, with a molecular mass of

168,138 dalton. CFTR is a cAMP-induced chloride channel [18-30], which is

located in epithelial tissues, including pancreatic ductal cells, salivary glands,

intestine, lung, testis and endometrium [31-34]. Defective regulation of the chloride

channel, caused by different mutations in the CF gene, results in desiccation of

secretions, increase in the viscosity of mucus and a decrease in mucociliary clear-

ance. The primary sequence of CFTR suggests that it is a transmembrane protein

which consists of five domains (Figure 1): two membrane-spanning domains, each

composed of six putative transmembrane segments; two nucleotide-binding

domains; and a unique regulatory (R) domain [35,36]. The membrane-spanning

domains contribute to the formation of the Cl- channel pore. The nucleotide-binding

domains and the regulatory domain control the channel activity through an

interaction with cytosolic nucleotides. Phosphorylation of the regulator domain by

cAMP-dependent kinase is required for the channel to open [37-39]. In addition to

this function, the complex expression pattern of the gene and the varied degree of

the clinical manifestations in different organs and tissues suggest that CFTR may

have other activities [40-45].

Thus far more then 500 presumed mutations have been identified in the CFTR

gene in the past 5 years [15]. These mutations can be divided in missense- (42%),

frameshift- (23%), nonsense- (16%), splice defect- (16%) and deletion mutations

12

(4%). The distribution of mutations over the gene is nonrandom, with the highest

density in NBF1 (14.4%) and to a lesser extent in NBF2 (10.8%) (Figure 1). Most

of the CF mutations have been identified among patients of Caucasian ancestry. The

population variation of CF mutations in Europe has been well documented by the

Cystic Fibrosis Consortium [46].

Figure 2 shows four different mechanisms by which CF mutations disrupt the

CFTR function [36].Class 1 mutations: Mutations (G542X, 621+G->T, etc) that

cause defective protein production by premature termination signals as a

consequence of affected splice sites, frameshifts or nonsense mutations [35,47,48].

These mutations result in a shortage of CFTR Cl- channels in affected epithelia

since little or no full-length protein is produced.Class 2 mutations:Several mutant

forms of CFTR fail to be transported correctly, which results in defective protein

processing. Experiments with other proteins suggest that class 2 mutations (∆F508,

S559T, N1303K, etc) cause an incorrect folding of CFTR. Cellular mechanisms

recognise mutant CFTR as abnormal and mark the protein for degradation

Figure 1. CFTR contains 27 exons and consists of five domains: two membrane-spanning domains (MSD),two nucleotide-binding domains (NBF1 and 2), and a regulatory (R) domain. (Adapted from Tsui andBuchwald)

13

which only one has been identified. The state of pancreatic function allows

discrimination of different phenotypes. Class 1 and 2 mutations are associated with

a severe pancreatic-insufficient phenotype, while class 3 and 4 mutations show a

normal pancreas function.

Approximately 70% of the CF chromosomes harbour a three basepair deletion

in CFTR, which removes the phenylalanine residue at amino acid 508 of the

predicted CFTR polypeptide (∆F508) [46]. The estimates of the age of the∆F508

mutation vary from 3,000 to 52,000 years [52-55]. According to the latter study, in

which microsatellites were used in different CF populations in Europe, the∆F508

mutation arose at least 52,000 years ago in a population genetically distinct from

the present European population. The mutation was introduced in Europe in

different periods. The first spread was during the Palaeolithic period. The Basque

population is believed to be a Palaeolithic population [56], and has a high relative

frequency of∆F508 (88%) [53]. Later introduction of∆F508 in Central and Nor-

thern Europe (Neolithic period) increased the frequency of this mutation in these

regions. The relative∆F508 frequencies are shown in Figure 3 [46]. The highest

frequency per country can be found in Denmark (0.88), while the frequency

decreases in Southern and Eastern directions.

1.3 Possibilities for genetic modification

A. Mouse models

The mouse homolog of the human CFTR gene was cloned in 1991 [57]. The mouse

protein is 78% identical to the human CFTR, with a high conservation in the

transmembrane and nucleotide-binding domains. A mouse model for cystic fibrosis

made by gene targeting was first described in 1992 [58]. The murine CFTR gene in

embryonic stem cells was inactivated by targeted insertion of a neomycin gene. This

resulted in a frame stop codon in exon 10. Numerous mutant mice have been

produced using this strategy [59-63]. Almost all mutant CF/CF mice displayed

many features common to young CF patients, including failure to thrive, meconium

ileus, alteration of mucous and serous glands. This resulted in death due to

intestinal obstruction before 40 days of age.

15

B. Somatic gene therapy for cystic fibrosis

The clinical expressions of CF are very broad, although the manifestations in the

airways account for more then 95% of the morbidity in CF patients [1]. It is

therefore that gene transfer strategies are focused on delivery to the airways. A

variety of gene transfer techniques have been used effectively in vitro and in

animals: adenoviral vectors, liposome-mediated gene transfer, DNA-ligand

complexes and adeno-associated viral (AAV) vectors [64]. The first two therapies

are being used in phase I clinical trials involving CF patients.

Several studies indicate that even a low expression (10%) of the CF gene in airway

epithelium may provide clinical benefit [65,66]. Healthy CF carriers will have half-

normal levels of expression. There are some compound heterozygotes which show

mild or normal phenotypes, while they have very low levels of expression [47,67].

This means that there is no necessity to correct all cells and even low levels of

expression per cell may be beneficial.

The major goals for somatic gene therapy for cystic fibrosis are (1) to show

that the expression of the CF gene in airways epithelium will alleviate lung disease

in CF patients and (2) to achieve a lasting expression of the transferred gene.

Table 3. Major advantages and disadvantages of gene transfer systems.

Advantages Disadvantages

Adenoviral 1. Highly efficient gene transfer 1. Transient nature of the expressionvectors 2. Transfection of many cell types 2. Pathogenicity of the wild type virus

3. No evidence for save and effectivereadministration.

DNA-liposome 1. Standardized production of 1. No persistent expressioncomplexes large amounts of vector 2. Low gene expression

2. No risks of viruses3. Readministrations with minimal

host response

DNA ligand 1. See DNA-liposome complexes 1. See DNA-liposome complexescomplexes 2. Large size of complexes may complicate

escape from the vascular space3. Peptides can induce immune responses

AAV vectors 1. Stable virus preparations 1. Possibility for immune responses2. Lack of pathogenicity 2. No adequate system to produce virus3. High-efficiency integration at present 3. Limitation of insert size to 4.8 kb

17

This can be performed through a single delivery with persistent expression or by

use of a save and effective repetitive delivery system. The biological aspects of

various gene transfer systems have been reviewed elsewhere [68,69]. The

advantages and disadvantages of these strategies are summarized in table 3.

1.4 Cystic fibrosis prevalence at birth

Cystic fibrosis is the most common autosomal recessive disease in caucasian

populations. Birth prevalences have been determined by surveys and range from 1

in 1700 to 1 in 40,000 in various European countries (Table 2). Thus far, estimates

based on molecular screening for heterozygotes have not been described in

European populations.

The estimates of the birth prevalences show a variability which as a whole or

in part may be due to under- or overdiagnosis and under- or overreporting.

Alternatively this variability may be due to mechanisms such as foundereffect and

genetic drift.

Table 2. Birth prevalences of cystic fibrosis determined in European populations

Czechoslovakia 1 in 2600 1962 [ 70]1 in 2800 1964 [ 71]1 in 2700 1967 [ 72]1 in 3400 1970 [ 73]1 in 3300 1970 [ 74]1 in 5440 1972 [ 75]

Denmark 1 in 4500 1972 [ 76]1 in 4760 1988 [ 77]

England 1 in 3000 1967 [ 78]1 in 4100 1967 [ 79]1 in 2000 1968 [ 80]1 in 3000 1972 [ 81]1 in 2500 1988 [ 82]

Finland 1 in 40000 1972 [ 83]France 1 in 3300 1961 [ 84]

1 in 1700-2000 1971 [ 85]1 in 2000 1973 [ 86]1 in 1800 1974 [ 87]1 in 3200 1978 [ 88]

Germany 1 in 3300 1963 [ 89]1 in 3000 1968 [ 90]1 in 2000 1975 [ 91]

18

Table 2. (Continued)

Ireland 1 in 1800 1976 [ 92]Italy 1 in 1100-3500 1973 [ 93]

1 in 2700 1976 [ 94]1 in 2000 1985 [ 95]

Netherlands 1 in 3600 1975 [ 96]N-Ireland 1 in 1700 1959 [Cited by Clarke, 97]

1 in 1900 1979 [ 98]Poland 1 in 900-2500 1970 [ 99]Sweden 1 in 7700 1962 [100]

1 in 4000 1970 [101]1 in 3500 1976 [102]1 in 2200-4500 1982 [103]

Switzerland 1 in 2900 1976 [104]Turkey 1 in 3000 1973 [105]

1.5 What causes the high frequency of cystic fibrosis ?

The high birth prevalence of cystic fibrosis of about one in 2500 in the Caucasian

population [1] has been discussed in many papers. A variety of mechanisms has

been suggested to account for this high frequency: genetic drift [106], high mutation

rate [107], meiotic drive and increased fertility of CF carriers [108], genetic

heterogeneity [109], heterozygote advantage [109-113] and segregation distortion of

CF alleles [108,114,115] . Heterozygote advantage is the most likely explanation. A

positive selection of approximately 2% [113] would be sufficient to maintain the CF

gene frequency at current levels.

The best understood example of heterozygote advantage remains sickle cell

haemoglobin, and the protection it confers against malaria. In case of CF,

abnormally functioning chloride channels could lead to increased protection against

bacterial toxin-mediated diarrhoea. Some of the diseases which have been suggested

as the balancing agent for CF are: influenza [116,117], syphilis [118], malaria

[119], bubonic plaque [112] and cholera [119-122]. Gabriel et al. [123] showed

reduced diarrhoea in the cystic fibrosis heterozygote mouse compared with the

normal mouse after exposure to cholera toxin. This effect, however, has not been

studied in human populations sofar.

19

1.6 Screening for cystic fibrosis carriers

The investigation of the validation of a mouthwash procedure and of the

determination of the carrier prevalence in the Netherlands (Chapters 2 and 3) can

partially be considered as preparatory studies for possible population-based

screening programmes. Immediately upon the cloning of the CF gene in 1989,

intensive discussions began regarding the feasibility and appropriateness of offering

carrier testing to the general population. Wilfond and Fost [124] have reviewed

many of the concerns in detail. The following arguments against population

screening have been raised. (1) It is too expensive and not cost-effective. (2) It is

not possible to identify all at-risk couples. (3) Families do not wish to avoid the

disease. (4) Prevention through abortion is morally objectionable. (5) A cure for the

disease is possible. (6) The personnel needed for genetic counselling are not

available. (7) Many couples are left at slightly increased risk. (8) There is a risk of

discrimination regarding insurance or employment. The following arguments in

favour of carrier testing have been raised. (1) Testing will become less expensive

and be cost-effective. (2) The majority of couples at-risk can be identified. (3)

Many couples wish to avoid the disease through various reproductive options. (4)

There is an obligation to inform the public about the availability of carrier testing.

Pilot studies for CF testing can address some of this issues.

Investigations in different groups of people revealed a positive attitude

toward carrier screening [125-128]. The American Society of Human Genetics and

the National Institutes of Health stated that carrier screening would be desirable

only if 90-95% of mutations can be detected. Present data suggest that the

sensitivity can be well above 90% in some Northern European populations, for

instance 98% in Belgium [129], 98% in a Celtic population in Brittany [130] and

99% in Wales [131]. The∆F508 mutation can be detected by simple gel

electrophoresis in material like blood samples, dried blood spots, mouthwashes,

mouthbrushes or fetal cells. The application and validation of the mouthwash

procedure for the DNA isolation out of buccal cells is described in Chapter 2. Most

laboratories now test routinely for from 10 to over 30 common CF mutations using

forward or reverse ASO methods [132] and the Amplification Refractory Mutation

System [133]. Solid-phase methods based on ligation detection, single base

sequencing and other tests offer great promise for the development of highly

20

automated procedures for the economical analysis of dozens of mutations

simultaneously [134]. An economic two-step laboratory approach can be used for

mutation detection [135]. In the first step only a few mutations will be determined

and only the partners of identified carriers will be tested for additional mutations.

The development of economic highly automated procedures for mutation analysis

and the positive attitude toward carrier testing will strongly increase the demand for

population based screening.

1.7 Identity By Descend (IBD) analysis

Haplotype sharing analysis is based on linkage disequilibrium of polymorphic loci

with a disease gene. In previous investigations extensive linkage studies have been

applied for the mapping of rare disease alleles. Although these studies provided

valuable contributions toward the mapping of mutated genes, a few disadvantages

can be observed using the traditional gene localization methods. (1) It is difficult to

obtain adequate family material and large pedigrees. (2) After the selection of

families all members should be examined concerning clinical features. (3) Hundreds

of markers should be involved in a complete genome search. (4) Costs of material

and manpower are high.

Identity by descent (IBD) analysis is an alternative association study method,

in which unrelated patients in a founder population can be selected for gene

mapping. Patients with identical genetic diseases are likely to share the disease gene

and an area around that gene from a common ancestor. The size of the expected

shared area depends on the number of meioses connecting the different persons,

because in time the size of the region will be reduced by crossovers. The larger the

shared area the closer the connection with a common predecessor from whom the

gene originates [136]. The size of the shared regions allows IBD to be used in the

genetic mapping of genes. Gruis et al. [137] described haplotype sharing in

presumedly unrelated patients from a founder population. This sharing was

determined by identity by descent of an area surrounding the disease gene. Houwen

et al. [138] showed that mapping within 2 cM was possible using only three

affected persons.

There are some points of importance which must be taken into consideration

21

when using IBD mapping. (1) The patients samples have to be collected in an

established founder population. (2) For the assignment of haplotypes it is necessary

to analyze also DNA of the parents. In case the phase is unknown it is possible to

assign haplotypes according to their frequency. This, however, will reduce the

power of the test drastically. (3) The allele frequency of the various markers used

with IBD mapping has to be established in a control group within the same founder

population.

A two step IBD approach can be used for gene mapping. In the first step a

few not too far related patients with an identical genetic disorder can be haplotyped

for a rough mapping of the disease gene. Markers can be selected with intervals of

about 20 cM. This means that 100-200 markers will cover the whole genome. Fine

mapping can be performed on the shared regions in a group of seemingly unrelated

patients within a restricted area. These regions will be haplotyped then with highly

polymorphic markers with relative distances of about 2 cM. This method allows fast

and relative cheap gene mapping. A pooling strategy has been described [139] in

which DNA from affected individuals were pooled and used as a PCR template for

microsatellite primers. The amplification product of pooled DNA samples reflects

all alleles represented in the pooled population. The relative frequency of the alleles

correlates with the intensity of each allelic band in the gel after electrophoresis. In

case of linkage disequilibrium of the microsatellite with a disease genotype there

will be a shift in allele frequencies towards a homozygous allele pattern in the

affected DNA pool. The use of the pooling strategy facilitates the identification of

the disease loci by reducing the number of amplification products and gel lanes.

Differences between control and affected pools can easily be examined, without the

need of assigning individual genotypes.

IBD analysis can also be used for determination of the origins of certain

mutations or the relationships of patients with identical genetic diseases. In case of

cystic fibrosis, mutations have been reported which occur predominantly in certain

countries [46]. The mutations may have originated in these countries not very long

ago. An example is the A455E CF mutation, which is mainly detected in Southern-

Holland and in Canada among certain French Canadians. It is likely that patients

with this mutation have a common not too far predecessor. Chapter 6 describes the

IBD analysis in Dutch and French Canadian patients.

22

24

References

1. Welsh MJ, Tsui L-C, Boat TF, Beaudet AL (1994) Cystic fibrosis In: Scriver CR, Beaudet AL, Sly WS et al.(eds) The Metabolic and Molecular Basis of Inherited Disease. New York, McGraw-Hill, pp 3799

2. Fanconi G, Uehlinger E, Knauer C (1936) The coeliac syndrome with congenital cystic pancreatic fibromatosis and bronchiectasis. Wien Med Wochenschr 86 , 753

3. Andersen DH (1938) Cystic fibrosis of the pancreas and its relation to celiac disease: a clinical and pathological study. Am J Dis Child 56 , 344

4. Farber S (1945) Some organic digestive disturbances in early life. J Michigan Med Soc 44 , 406

5. Anderson DH, Hodges RG (1946) Celiac syndrome. V. Genetics of cystic fibrosis of the pancreas with a consideration of etiology. Am J Dis Child 72 , 62

6. Di Sant'Agnese PA, Darling RC, Perera GA, Shea E (1953) Abnormal electrolytic composition of sweat in cystic fibrosis of the pancreas. Clinical significance and relationship of the disease. Pediatrics 12 , 549

7. Gibson LE, Cooke RE (1959) A test for concentration of electrolytes in sweat in cystic fibrosis of the pancreas utilizing pilocarpine by iontophoresis. Pediatrics 23 , 545

8. Knowles M, Gatzy J, Boucher R (1981) Increased bioelectric potential difference across respiratory epithelia in cystic fibrosis. N Engl J Med 305 , 1489 9. Quinton, PM (1983) Chloride impermeability in cystic fibrosis. Nature 301 , 421

10. Tsui L-C, Buchwald M, Barker D, Braman JC, Knowlton R, Schumm JW, Eiberg H, Mohr J, Kennedy D, Plavsic N, Zsiga M, Markiewicz D, Akots G, Brown V, Helms C, Gravius T, Parker C, Rediker K

and Donis-Keller H (1985) Cystic fibrosis locus defined by a genetically linked polymorphic DNA marker. Science 230 , 1054 11. Estivill X, Scambler PJ, Wainwright BJ, Hawley K, Frederick P, Schwartz M, Baiget M, Kere J,

Williamson R, Farrall M (1987) Patterns of polymorphism and linkage disequilibrium for cystic fibrosis. Genomics 1, 257

12. Rommens JM, Iannuzi MC, Kerem B, Drumm ML, Meimer G, Dean M, Rozmahel R, Cole JL, Kennedy D, Hidaka N, Zsiga M, Buchwald M, Riordan JR, Tsui L-C, Collins FS (1989). Identification

of the cystic fibrosis gene: chromosome walking and jumping. Science 245 , 1059

13. Kerem B, Rommens JM, Buchanan JA, Markiewicz D, Cox TK, Chakravarti A, Buchwald M, Tsui L-C (1989) Identification of the cystic fibrosis gene: genetic analysis. Science 245 , 1073

14. Zielenski J, Rozmahel R, Bozon D, Kerem B-S, Grzelczak Z, Riordan JR, Rommens J, Chee L-C (1991). Genomic DNA sequence of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR)

Gene. Genomics 10 , 214

15. Cystic Fibrosis Genetic Analysis Consortium (December 16, 1994a) Cystic fibrosis mutation data (written communication)

16. Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R, Grzelczak Z, Zielenski J, Lok S, Plavsic N, Chou J-L, Drumm ML, Iannuzzi MC, Collins FS, Tsui L-C (1989) Identification of the cystic fibrosis

gene: cloning and characterization of complementary DNA. Science 245 , 1066

25

17. Van der Hout AH, Van der Veen AY, Aten JA, Buys CHCM (1988) Localization of DNA probes with tight linkage to the cystic fibrosis locus by in situ hybridization using fibroblasts with a 7q22 deletion. Hum Genet 80 ,161

18. Anderson MP, Rich DP, Gregory RJ, Smith AE, Welsh MN (1991a) Generation of cAMP-activated chloride currents by expression of CFTR. Science 251 , 679

19. Anderson MP, Gregory RJ, Thompson S, Souza DW, Paul S, Mulligan RC, Smith AE, Welsh MN (1991b) Demonstration that CFTR is a chloride channel by alteration of its anion selectivity. Science

253 , 202

20. Anderson MP, Berger HA, Rich DP, Gregory RJ, Smith AE (1991c) Nucleotide triphosphates are required to open the CFTR chloride. Cell 67 , 775

21. Bear CE, Duguay F, Naismith AL, Kartner N, Hanrahan JW, Riordan JR (1991) Cl channel activity in -

Xenopus oocytes expressing the cystic fibrosis gene. J Biol Chem 266 , 19142

22. Berger HA, Anderson MP, Gregory RJ, Thompson S, Howard PW (1991) Identification and regulation of the cystic fibrosis transmembrane conductance regulator. J Clin Invest 88 , 1422

23. Dalemans W, Barbry P, Champigny G, Jallat S, Dott K, Dreyer D, Crystal RG, Pavirani A, Lecocq JP, Lazdunski M (1991) Altered chloride ion channel kinetics associated with the )F508 cystic fibrosis mutation. Nature 354 , 526

24. Drumm ML, Wilkinson DJ, Smit LS, Worrell RT, Strong TV, Frizzel RA, Dawson DC, Collins FS (1991) Chloride conductance expressed by delta F508 and other mutant CFTRs in Xenopus oocytes.

Science 254 , 1797

25. Kartner N, Hanrahan JW, Jensen TJ, Naismith AL, Sun S, Ackerley CA, Reyes EF, Tsui L-C, Rommens JM, Bear CE, Riordan JR (1991) Expression of the cystic fibrosis gene in non-epithelial invertebrate cells produces a regulated anion conductance. Cell 64 , 681

26. Rommens JM, Dho S, Bear CE, Kartner N, Kennedy D, Riordan JR, Tsui L-C, Foskett JK (1991) cAMP-inducible chloride conductance in mouse fibroblast lines stably expressing the human cystic

fibrosis transmembrane conductance regulator. Proc Natl Acad Sci USA 88 , 7500

27. Tabcharani JA, Chang X-B, Riordan JR, Hanrahan JW (1991) Phosphorylation-regulated Cl channel in -

CHO cells stably expressing the cystic fibrosis gene. Nature 352 , 628

28. Cliff WH, Schoumacher RA, Frizzell RA (1992) cAMP-activated Cl Channels in CFTR- transvected cystic fibrosis pancreatic epithelial cells. Am J Physiol 262 , C1154

29. Cunningham SA, Worrell RT, Benos DJ, Frizzell RA (1992) cAMP-stimulated ion currents in Xenopus oocytes expressing CFTR cRNA. Am J Physiol 262 , C783

30. Bijman J, Dalemans W, Kansen M, Keulemans J, Verbeek E, Hoogeveen A, De-Jong H, Wilke M, Dreyer D, Lecocq JP (1993) Low conductance chloride channels in IEC6 and CF nasal cells expressing CFTR. Am J Physiol 264 , L229

31. Bijman J, Englert HC, Lang HJ, Greger R, Fromter E (1987) Characterisation of human sweat duct chloride conductance by chloride channel blockers. Pflugers Archiv Eur J Physiol 408 , 511

32. Gregory RJ, Cheng SH, Rich DP, Marshall J, Paul S, Hehir K, Ostedgaard L, Klinger KW, Welsh MJ, Smith AE (1990) Expression and characterization of the cystic fibrosis transmembrane conductance

regulator. Nature 347 , 382

26

33. Halley DJJ, Bijman J, Jonge HR de, Sinaasappel M, Neyens HJ, Niermeyer MF (1990) The cystic fibrosis defect approached from different angles. New perspectives on the gene, the chloride channel,

diagnosis and therapy. Eur J Pediatr 149 , 670

34. Rich DP, Anderson MP, Gregory RJ, Cheng SH, Paul S, Jefferson DM, McCann JD, Klinger KW, Smith AE, Welsh MJ (1990) Expression of cystic fibrosis transmembrane conductance regulator corrects

Defective chloride channel regulation in cystic fibrosis airway epithelial cells. Nature 347 , 358

35. Tsui L-C (1992) Mutations and sequence variations detected in the cystic fibrosis conductance regulator (CFTR) gene: a report from the Cystic Fibrosis Genetic Analysis Consortium. Hum Mut 1, 197

36. Welsh MJ, Smith AE (1993) Molecular mechanisms of CFTR chloride channel dysfunction in cystic fibrosis. Cell 73 , 1251

37. Bear CE, Li C, Kartner N, Bridges RJ, Jensen TJ, Ramjeesingh M, Riordan JR (1992) Purification and functional reconstruction of the cystic fibrosis trans membrane conductance regulator (CFTR). Cell 68 , 809

38. Welsh MJ, Anderson MP, Rich DP, Berger HA, Denning GM, Ostedgaard LS, Sheppard DN, Cheng SH, Gregory RJ, Smith AE (1992) Neuron 8, 821

39. Collins FS (1992) Cystic fibrosis: molecular biology and therapeutic implications. Science 256 , 774 40. Quinton PM (1990) Cystic fibrosis: a disease in electrolyte transport. FASEB J 4, 2709

41. Tsui L-C, Buchwald M (1991) In Advances Hum Genet, H. Harris and K. Hirschhorn, eds (New York:Plenum Press), pp. 153

42. Gill DR, Hyde SC, Higgins CF, Valverde MA, Mintenig GM, Sepulveda FV (1992) Separation of drug transport and chloride channel functions of the human multidrug resistance P-glycoproteine. Cell 71 ,

23

43. Valverde MA, Diaz M, Sepulveda FV, Gill DR, Hyde SC, Higgins CF (1992) Volume-regulated chloride channels associated with the human multidrug-resistance P-glycoprotein. Nature 355 , 830

44. Wine JL, Silverstein SC (1992) ATP and chloride conductance. Nature 360 , 18

45. Gabriel SE, Clarke LL, Boucher RC, Stutts MJ (1993) CFTR and outward rectifying chloride channels are distinct proteins with a regulatory relationship. Nature 263 , 263

46. Cystic fibrosis Genetic Analysis Consortium (1994b) Population variation of common cystic fibrosis mutations. Hum Mut 4, 167

47. Delaney SJ, Rich DP, Thomason SA, Hargrave MR, Welsh MJ, Wainwright BJ (1993) CFTR splice variants are not conserved and fail to produce Cl channels. Nat Genet 4, 426-

48. Strong TV, Wilkinson DJ, Mansoura MK, Devor DC, Henze K, Yang Y, Wilson JM, Cohn JA, Dawson DC, Frizzel RA, Collins FS (1993) Expression of an abundant alternatively spliced form of the cystic fibrosis transmembrane conductance regulator (CFTR) gene is not associated with a cAMP-activated chloride conductance. Hum Mol Genet 2, 335

49. Cheng SH, Gregory RJ, Marshall J, Paul S, Souza DW, White GA, O'Riordan CR, Smith AE (1990) Defective intracellular transport and processing of CFTR is the molecular basis of most cystic fibrosis.

Cell 63 , 827

27

50. Thomas PJ, Shenbagamurthi P, Ysern X, Pedersen PL (1991) Cystic fibrosis transmembrane conductance regulator: nucleotide binding to a synthetic peptide. Science 251 , 555

51. Sheppard DN, Rich DP, Ostedgaard LO, Gregory RJ, Smith AE, Welsh MJ (1993) Mutations in CFTR associated with mild-disease-form Cl channels with altered pore properties. Nature 362 , 160-

52. Serre JL, Simon-Bouy B, Mornet E, Jaume-Roig B, Balassopoulou A, Schwartz M, Taillandier A, Boué J, Boué A (1990) Studies of RFLP closely linked to the cystic fibrosi locus throughout Europe lead to new considerations in populations genetics. Hum Genet 84 , 449

53. Casals T, Vazquez C, Lazaro C, Girbau E, Gimenez FJ, Estivill X (1992) Cystic fibrosis in the Basque country: high frequency of mutation delF508 in patients of Basque origin. Am J Hum Genet 50 , 404 54. Casals T, Nunes V, Palacio A, Giménez J, Gaona A, Ibáñez N, Morral N, Estivill X (1993) Cystic

fibrosis in Spain: high frequency of mutation G542X in the Mediterranean coastal area. Hum Genet 91 , 66

55. Morral N, Bertranpetit J, Estivill X, Nunes V, Casals T, Giménez J, Reis A, Varon-Mateeva R, Macek Jr M, Kalaydjieva L, Angelicheva D, Dancheva R, Romeo G, Russo MP, Garnerone S, Restagno G, Ferrari M, Magnani C, Claustres M, Desgeorges M, Schwartz M, Schwartz M, Novelli G, Ferec C, Arce

M de, Nemeti M, Kere J, Anvret M, Dahl N, Kadasi L (1994) The origin of the major cystic fibrosis mutation ()F508) in European populations. Nature Genet 7, 169

56. Bertranpetit J, Cavalli-Sforza LL (1991) A genetic history of the Iberian Peninsula. Ann Hum Genet 55 , 51

57. Tata F, Stanier P, Wicking C, Halford S, Kruyer H, Lench NJ, Scambler PJ, Hansen C, Braman JC, Williamson R, Wainwright BJ (1991) Cloning the mouse homolog of the human cystic fibrosis transmembrane conductance regulator gene. Genomics 10 , 301

58. Snouwaert JN, Brigman KK, Latour AM, Malouf NN, Boucher RC, Smithies O, Koller BH (1992) An animal model for cystic fibrosis made by gene targeting. Science 257 , 1083

59. Clarke LL, Grubb BR, Gabriel SE, Smithies O, Koller BH, Boucher RC (1992) Defective epithelial chloride transport in a gene-targeted mouse model of cystic fibrosis. Science 257 , 1125

60. Dorin JR, Dickinson P, Alton EWFW, Smith SN, Geddes DM, Stevenson BJ, Kimber WL, Fleming S, Clarke AR, Hooper ML, Anderson L, Beddington RSP, Porteous DJ (1992) cystic fibrosis in the mouse by targeted insertional mutagenesis. Nature 359 , 211

61. Colledge WH, Ratclif R, Foster D, Williamson R, Evans MJ (1992) Cystic fibrosis mouse with intestinal obstruction. Lancet 340 , 680

62. Ratcliff R, Evans MJ, Cuthbert AW, MacVinish LJ, Foster D, Anderson JR, Colledge WH (1993) Production of a severe cystic fibrosis mutation in mice by gene targeting. Nature Genet 4, 35

63. O'Neal WK, Hasty P, McCray PB Jr, Casey B, Rivera-Pérez J, Welsh MJ, Beaudet AL, Bradley A (1993) A severe phenotype in mice with a duplication of exon 3 in the cystic fibrosis locus. Hum Mol

Genet 2, 1561

64. O'Neal WK, Beaudet AL (1994) Somatic gene therapy for cystic fibrosis. Hum Mol Genet 3, 1497

65. Johnson LG, Olsen JC, Sakardi B, Moore KL, Swanstrom R, Boucher RC (1992) Nature Genet 2, 21

28

66. Zabner J, Couture LA, Smith AE, Welsh MJ (1994) Correction of cAMP-stimulated fluid secretion in cystic fibrosis airway epithelia: Efficiency of adenovirus-mediated gene transfer in vitro. Hum Gene Ther 5, 585

67. Chu CS, Trapnell BC, Curristin SM, Cutting GR, Crystal RG (1992) Extensive posttranscriptional deletion of the coding sequences for part of nucleotide-binding fold 1 in respiratory epithelial mRna transcripts of the cystic fibrosis transmembrane conductance regulator gene is not associated with the clinical manifestations of cystic fibrosis. J Clin Invest 90 , 785

68. Mulligan RC (1993) The basic science of gene therapy. Science 260 , 926

69. Morgan RA, Anderson WF (1993) Human gene therapy. Annu Rev Biochem 62 , 191

70. Houšt�k J, Vávrová V (1962) On the incidence of cystic fibrosis of the pancreas in Czechoslovakia. Ñ Pediatr 17 , 445

71. Simanková N, Vávrová V, Houšt�k J (1964) Etude génétique de la mucuviscidose. Ann Génét 7, 84

72. Houšt�k J, Vávrová V (1967) Notre experience à propos de la mucoviccidose. Rev Méd Liège 22 , 421

73. Brunecky Z, Laxova R, Kamaryt J (1970a) Incidence and diagnosis of cystic fibrosis (mucoviscidosis): experiences with night sweat collection. Kinderärtzl Prax 38 , 514

74. Brunecki Z, Simanková N, Vávrová V (1970b) Incidence and genetics of cystic fibrosis. Ñ Pediatr 25 , 74

75. Brunecky Z (1972) The incidence and genetics of cystic fibrosis. J Med Genetics 9, 33

76. Nielsen EL (1972) Cystic fibrosis: incidence in Danmark. Acta paediatr Scan 61 , 377 77. Nielsen OH (1988) Cystic fibrosis in Denmark 1945 to 1985. Acta Paediatr Scan 77 , 836

78. Carter CO, Watson GI (1967) General aspects of cystic fibrosis of the pancreas. Mod Probl Pediat 10 , 372

79. Pugh RJ, Pickup JD (1967) Cystic fibrosis in the Leeds region: incidence and life expectancy. Arch Dis Child 42 , 544

80. Hall BD, Simpkiss MJ (1968) Incidence of fibrocystic disease in Wessex. J Med Genet 5, 262

81. Cain ARR, Deall AM, Noble TC (1972) Screening for cystic fibrosis by testing meconium for albumin. Arch Dis Childh 47 , 131

82. British Paediatric Association Working party on cystic fibrosis (1988) Cystic fibrosis in the United Kingdom 1977-1985: an improving picture. Brit Med J 297 , 1599

83. Nevanlinna HR (1972) The finnish population structure, a genetic and genealogical study. Hereditas71 , 195

84. Bernheim M, Monnet P, Jeune M, Robert JM, Comby J (1961) La malade fibro-kystique des parenchymes glandulaires: Etude génétique de 41 families. Pediatrie 16 , 17

85. Gilly R, Hermier M,Robert JM,Challamel MJ,Gery G (1971) Genetic study of mucoviscidosis:statistics of 273 families. Arch franç Pédiat 28 , 49

29

86. Gilly R (1973) Mucoviscidosis. Pathogenesis. Ann Pediat 20 , 5

87. Feingold J, Hennequet A, Jehanne M, Feigelson J, Toudic L, Quinious O, Briard ML (1974) Incidence of pancreatic cystic fibrosis in France. Ann Génét 17 , 257

88. Bois E, Feingold J, Demenais F, Runavot Y, Jehanne M, Toudic L (1978) Cluster of cystic fibrosis cases in a limited area of Brittany (France). Clin Genet 14 , 73

89. Vivell VO, Jacobi H, Münchbach K (1963) Zur Mucoviscidosis im Kindesalter. Monatsschr Kinderheilkd 111 : 62

90. Koch G, Schwanitz G (1968) Neuere human-genetische aspekte der mucoviscidose beim kind, in: Windorfer, A and Stephan, U (eds.), Mucoviscidose, 2. deutches symposion, Georg Thieme Verlag,

Stuttgart.

91. Dietzsch HS, Gottschalk B (1975) CF screening: results of a pilot study on 10,000 newborns in the German Democratic Republic. European working group for cystic fibrosis. Dublin.

92. O'Reilly D, Murphy J, McLaughlin J, et al (1974). The prevalence of coeliac disease and cystic fibrosis in Ireland, Scotland and Wales. Int J Epidemiology 3, 247

93. Migliori V (1973) Mucoviscidosis. Minerva Med 64 , 4906

94. Righetti ABB, Mighavacca M, Prampolini L, Guinta A (1976) Extensive neonatal screening of CF. In Proc VII Int Cystic Fibrosis Congress. Ed. Hennequet A, pp 153

95. Romeo G, Bianco M, Devoto M, Menozzi P, Mastella G, Giunta AM, Micalizzi C, Antonelli M, Battistini A, Santamaria F, Castello D, Marianelli A, Marchi AG, Manca A, Miamo A (1985) Incidence

in Italy, genetic heterogeneity and segregation analysis of cystic fibrosis. Am J Hum Genet 37 , 338

96. Ten Kate LP (1977) Cystic fibrosis in the Netherlands. Int J Epidem 6, 23

97. Clarke CA (1969) The relative fitness of human mutant genotypes, in: Roberts DF, Harrison GA (eds). Natural selection in human populations. Pergamon Press, London.

98. Nevin GB, Nevin NC, Redmond AD (1979) Cystic fibrosis in Northern Ireland. J Med Genet 16 , 122

99. Bozkowa K, Golebiowska H, Rutkowski J, Nowakowska A, Holzer J (1970) Incidence of cystic fibrosis in Poland. 1st ann Meeting Eur Working Group for Cystic Fibrosis. Stockholm, Sweden, August 24-25

100.Selander P (1962) The frequency of cystic fibrosis of the pancreas in Sweden. Acta paediatr 51 , 65

101. Kollberg H (1970) Cystic fibrosis in Sweden, 1st Ann Meeting Eur Working Group for Cystic Fibrosis.

Stockholm, Sweden.

102. Kollberg H (1976) Screening for cystic fibrosis (CF) by analysis of albumin in meconium. VII. int.

Cystic Fibrosis congress, May 31-June 1-3. Paris, France, pp. 32

103. Kollberg H (1982) Incidence and survival curves of cystic fibrosis in Sweden. Acta Paediatr Scan 71 ,

197

104.Kaiser D, Morer HP, Hammershlag P (1976) Experience of 3 years meconium screening in

Switzerland. In Proc VII Int Cystic Fibrosis Congress. Ed. Hennequet A, pp 148

30

105. Gürson CT, Sertel H, Gürkan M, Pala S (1973) Newborn screening for cystic fibrosis with the chloride

electrode and neutron activation analysis. Helvetica paediatrica acta 28 , 165

106.Wright SW, Norton NE (1968) Genetic studies on cystic fibrosis in Hawaii. Am J Hum Genet 20 , 157

107. Goodman HO, Reed SC (1952) Heredity of fibrosis of the pancreas: possible mutation rate of the gene.

Am J Hum Genet 4, 59

108.Kitzis A et al. (1988) Unusual segregation of cystic fibrosis alleles. Nature 336 , 316

109. Romeo G, Devoto M, Galietta LJV (1989) Why is the cystic fibrosis gene so frequent? Hum Genet 84 ,

1

110. Danks DM, Allan J, Anderson CM (1965) A genetic study of fibrocystic disease of the pancreas. Ann

Hum Genet London 28 , 323

111.Super M, Shalkwyk DJ van (1979) Heterozygote advantage in cystic fibrosis: mosquito tests. Clin

Genet 16 , 65

112.Cassano WF (1985) Cystic fibrosis and the plague. Med Hypoth 18 , 51

113. Jorde LB, Lathrop GM (1988) A test of the heterozygote advantage hypothesis in cystic fibrosis

carriers. Am J Hum Genet 42 , 808

114. Knudson AG, Wayne L, Hallet WY (1967) On the selective advantage of cystic fibrosis heterozygotes.

Am J Hum Genet 19 , 388

115. Gloria-Bottini F, Antonelli M, Quattrucci S et al. (1980) Is there a relationship between sex of cystic

fibrosis carriers and sex ratio of their offspring? Hum Genet 54 , 79

116. Shier WT (1979) Increased resistance to influenza as a possible source of heterozygote advantage in

Cystic fibrosis. Med Hypoth 5, 661

117.Meindl, RS (1987) A selective advantage for cystic fibrosis heterozygotes. Am J Phys Anthropol 74 , 39

118. Hollander DJ (1982) Etiogenesis of the european cystic fibrosis polymorphism:heterozygote advantage

against venereal syphilis? Med Hypoth 8, 191

119. Baxter PS, Goldhill J, Hardcastle J, Hardcastle PT, Taylor CJ (1988) Accounting for cystic fibrosis.

Nature 336 , 211

120. Rodman DM, Zamudio S (1991) The cystic fibrosis heterozygote advantage surviving cholera? Med

Hypoth 36 , 253

121.Bijman J, De Jonge H, Wine J (1988) Cystic fibrosis advantage. Nature 336 , 430

122.Hansson GC (1988) Cystic fibrosis and chloride-secreting diarrhoea. Nature 333 , 711

123. Gabriel SE, Brigman KN, Koller BH, Boucher RC, Stutts MJ (1994) Cystic fibrosis heterozygote

resistance to cholera toxin in the cystic fibrosis mouse model. Science 266 , 107

124. Wilfond BS, Fost N (1990) The cystic fibrosis gene: Medical and social implications for heterozygote

detection. JAMA 263 , 2777

125.Ten Kate LP, Tijmstra TJ (1989) Carrier screening for cystic fibrosis. Lancet, 973

31

126. Kaplan F, Clow C, Scriver CR (1991) Cystic fibrosis carrier screening by DNA analysis: A pilot study

of attitudes among participants. Am J Hum Genet 49 , 220

127. Watson EK, Marchant J, Bush A, Williamson B (1992) Attitudes towards prenatal diagnosis and carrier

screening for cystic fibrosis among the parents of patients in a paediatric cystic fibrosis clinic. J Med Genet 29 , 490

128.Williamson R (1993) Universal community carrier screening for cystic fibrosis ? Nature Genet 51 , 943

129. Cuppens H, Marymen P, DeBoeck C, Cassiman JJ (1993) Detection of 98.5 % of the mutations in 200

Belgian cystic fibrosis alleles by reverse dot-blot and sequencing on the complete coding region and exon/intron junctions of the CFTR gene. Genomics 18 , 693

130. Férec C, Audrézet MP, Mercier B, Guillermit H, Moullier P, Quéré I, Verlinque C (1992) Systematic

Screening for mutations in the cystic fibrosis gene: New implications for carrier detection. Nature Genet 1, 188

131. Cheadle JP, Goodchild MC, Meredith AL (1993) Direct sequencing of the complete CFTR gene: The

molecular characterzation of 99.5 % of CF chromosomes in Wales. Hum Mol Genet 2, 1551

132. Shuber AP, Skoletsky J, Stern R, Handelin BL (1993) Efficient 12-mutation testing in the CFTR gene:

A general model for complex mutation analysis. Hum Mol Genet 2, 153

133. Ferrie RM, Schwartz MJ, Robertson NH, Vaudin S, Super M, Malone G, Little S (1992) Development,

multiplexing, and application of ARMS test for common mutations in the CFTR gene. Am J Hum Genet 51 , 251

134. Syvänen A-C, Landegren U (1994) Detection of point mutations by solid-phase methods. Hum Mutat 3,

172

135. Beaudet AL, O'Brien WE (1992) Advantages of a two-step laboratory approach for cystic fibrosis

carrier screening. Am J Hum Genet 50 , 439

136. Te Meerman GJ, Meulen MA van der, Sandkuijl LA (1995) Perspectives of identity by descent (IBD)

mapping in founder populations. (Submitted)

137. Gruis N, Sandkuijl, LA, Bergman W, Frants RR (1994) Common 9p Haplotype in Dutch Fammm

families, PhD thesis N. Gruis, The genetics of the familial atypical multiple melanoma syndrome, Leiden

138. Houwen RMJ, Baharbo S, Blankenship K, Raeymaekers P, Juyn J, Sandkuyl LA, Freimer NB (1994)

Genome screening by searching for shared segments: mapping a gene for benign recurrent intrahepatic cholestasis. Nature Genet 8, 380

139. Carmi R, Rokhlina T, Kwitek-Black AE, Elbedour K, Nishimura D, Stone EM, Sheffield VC (1995)

Use of a DNA pooling strategy to identify a human obesity syndrome locus on chromosome 15. Hum Mol Genet 4, 9

Chapter 2

Validation of determination of ∆F508 mutations of the cystic

fibrosis gene in over 11,000 mouthwashes.

Hendrik G. de Vries1, J. Margriet Collée2, Marjan H.R. van Veldhuizen2, Leny

Achterhof 3, Cees Th. Smit Sibinga3, Hans Scheffer1, Charles H.C.M. Buys1, Leo

P. ten Kate2.

1 Department of Medical Genetics, University of Groningen, The Netherlands2 Department of Human Genetics, Free University of Amsterdam, The Netherlands3 Blood bank Groningen-Drenthe, Groningen, The Netherlands

Hum. Genet (in press).

Abstract

Mouthwashes can be used as DNA resource for mutation detection. Because

collection and DNA isolation is simple and cheap, it could especially be used for

large numbers of samples. To determine the failure rate (the proportion of

mouthsamples in which no PCR product was obtained) and the specificity of buccal

epithelial cell mutation detection in large numbers of samples, we collected mouth-

washes and blood samples from 11,413 blood donors and tested the mouthwashes

for the ∆F508 mutation, which has an estimated frequency of 75% among cystic

fibrosis chromosomes in The Netherlands. Blood samples were tested for the∆F508

mutations only if the mutation was identified in the mouthwash or in case of a

failure to obtain PCR products. The sensitivity of the test was determined in

mouthwashes of 75∆F508 carriers known from earlier family studies. These

samples were offered blindly between the mouthwashes of the blood donors. Both

specificity and sensitivity of the mouthwash procedure were 100%. The overall

failure rate was 5.6 %. This large figure was caused mainly by insufficient rinsing

of the mouth in one particular blood bank. Exclusion of the results of this blood

33

bank reduced the failure rate to 1.8%. Our results confirm that also for large

number of samples the mouthwash procedure is suitable for mutation detection and

with proper instructions can be used in community screening.

Introduction

White blood cells are the main source of DNA used for gene analysis (Miller et

al.1980). Blood sampling and DNA isolation from great numbers of such samples,

however, have a few disadvantages: medical supervision is necessary for blood

collection, people may shrink from venopuncture, there is a risk of exposure to

blood pathogens for the donor as well as for the investigator, and costs of DNA

isolation are high. An alternative is sampling of mouthwashes followed by DNA

isolation from buccal cells (Lench et al.1988). This procedure is much simpler and

cheaper than existing methods, especially with large numbers of samples. There is

no need for medical supervision of sample collection, and the risk of infections is

eliminated.

To determine the failure rate (the proportion of mouthwash samples in which

no PCR product was obtained), the specificity and the sensitivity of the mouthwash

procedure, we studied the presence of the∆F508 mutation of the cystic fibrosis

(CF) gene in a large number of mouthwash samples with known or unknown

carrierstatus.

Materials and Methods

Recruitment and sampling

11,413 blood donors from 15 blood banks were approached in a study to determine

the prevalence of carriers of the∆F508 mutation in the CFTR gene in this country

and the failure rate and specificity of the mouthwash procedure. The design was

approved by the ethical committee of the Groningen University Hospital. Blood

bank officials were instructed how to recruit volunteers among the blood donors.

After informed consent, the blood donors were asked to rinse their mouth with 15

34

ml of 0.9% sterile saline for 10 seconds. The samples were collected in sputum

containers (Emergo, Landsmeer, The Netherlands). Matched blood samples of the

blood donors were collected in 10 ml heparin tubes. Samples were stored at 4° C

until shipment to either one of the two participating laboratories. Here all

mouthwash samples were analyzed. In order to determine the proportion of false

positives mutations detected in mouthwash DNA were checked in the matched

blood samples. Sensitivity was determined in mouthwash samples of 75 known

∆F508 carriers detected in earlier studies of CF families. The samples of these

carriers were offered blindly to the investigator among the mouthwashes of the

blood donors.

DNA isolation and (mutation) analysis

DNA isolation out of buccal samples was performed according to a procedure

modified from Lench et al.(1988). Buccal cells were pelleted by centrifugation at

1500 g for 10 minutes. The cell pellet was resuspended in 100 µl 50 mM sodiumhy-

droxide and boiled for 10 minutes. Samples were neutralised with 14 µl 1 M TRIS

(pH 7.5) and 5 µl was taken for PCR analysis. Genomic DNA was extracted from

heparin blood according to Miller et al.(1980).

The ∆F508 mutation was determined in the following way. PCR was

performed with primers according to Scheffer et al.(1989) in a Pharmacia LKB-

Gene ATAQ controller with denaturation temperatures decreasing from 94°C to

88°C in three steps of 10 cycles. The∆F508 mutation was detected directly on a 13

% polyacrylamide gel (50 samples per gel). Mutations detected in mouthwashes

were checked in the matched blood samples.

Results

Out of 11,413 mouthwashes we were able to analyse PCR products in 10,772

samples. A total number of 254∆F508 carriers were detected in this group. All 254

positive results could be confirmed by DNA analysis in matched blood samples.

The specificity of the mouthwash procedure was 10501/10501 or 100% (95% CI=

99.96-100%, assuming a sensitivity of 100%, see below).

35

DNA isolation failures occurred in 641 cases, due to insufficient mouthwas-

hings, leaking containers and food rest contamination. The overall failure rate of

mouthwashes therefore was 5.6%. In the matched blood samples of these cases an

additional 17∆F508 carriers were detected. The difference in the carrier frequencies

detected in mouthwashes and blood samples was not statistically significant.

The samples (492) collected from one particular blood bank yielded few

successful DNA isolations, with a failure rate of 90 percent. This probably has been

caused by insufficient rinsing of the mouth. Exclusion of these data resulted in a

total yield of 98.2 % successful DNA analyses (failure rate 1.8%).

Of the failures remaining after excluding those samples of the above

mentioned blood bank, almost 40% was caused by containers leaking during

transport of the mouthwash samples to the laboratories. Another 40% was caused by

insufficient rinsing and 20% by foodrest contamination.

More then 1,000 samples were sent from the blood banks to the laboratories

after storage for 14 days at 4°C. DNA isolation from these mouthwashes has the

same success rate of over 98%.

We were able to obtain PCR products from 74 out of 75 mouthwash samples

of known ∆F508 carriers. In all 74 samples we confirmed carriership of this

mutation. This gives a sensitivity of 100% (95 % CI= 95.1-100%). In one case we

were unable to isolate DNA due to contamination with food remains. The

percentage of successfully analysed mouthwashes of known carriers, therefore, was

98.2% (failure rate 1.8%).

Discussion

We detected a total of 271∆F508 carriers out of 11,413 unrelated individuals. Since

∆F508 represents 75% of the CF mutations in The Netherlands (Scheffer et al.

submitted), this is in keeping with a prevalence of the CF carrierstatus (all

mutations) of about 1 in 30 in The Netherlands (Ten Kate 1977). The specificity

and sensitivity of the mouthwash

procedure were both 100%.

The validation of DNA isolation out of buccal cells

(mouthwashes/mouthbrushes) has been described before (Richard et al. 1993;

36

Gilfillan et al. 1994;). These studies applied multiplex PCR amplifications to detect

the most frequent CFTR mutations. The success rates of these studies were

comparable with our results. The present study is the first to present data on the

sensitivity of the mouthwash procedure as well as effects of storage and transport.

Determination of the failure rate and the specificity of the procedure are based upon

a substantially larger number of samples than in the previous studies.

Our analysis of 11,413 mouthwashes showed a PCR success rate of 94.4%.

After exclusion of the samples from one particular blood bank this percentage

increased to 98.2%. Because mutation detection was comparable among the other

14 blood banks, the failures to obtain PCR-products in the samples of one blood

bank seem to represent an isolated case.

We did not collect mouthwashes under optimum conditions, because blood

donors generally are advised to eat before donating blood. Excluding the one

particular blood bank, this resulted in 30 PCR failures out of 10,921 (0.27%).

For the application of the mouthwash procedure one should be aware of the

above mentioned problems. The need for giving good instructions should not be

underestimated, even with simple collection procedures. If this condition is fulfilled,

mouthwashes can easily be used for single mutation detections with a large number

of samples, like in carrier screening programmes.

Acknowledgements

The authors want to thank the following persons and blood banks for making

mouthwashes and bloodsamples available: Th.M. Brouwers, director of blood bank

Apeldoorn-Deventer-Zutphen; G.C. van der Plas, director of blood bank Arnhem

e.o.; H.J.C. de Wit, director of blood bank Friesland; Mrs J.H. Barmentloo, director

of blood bank Gooi en Eemland; Dr. J.A. van der Does, director of blood bank ’s-

Gravenhage e.o.; R.C. Schaap, director of blood bank Kennemerland; Mrs dr. A.

Brand, director of blood bank Leiden e.o.; Dr. L.H. Siegenbeek van Heukelom,

director of blood bank Noord-Holland-Noord; Dr. H.J.A. Sonderkamp, director of

blood bank Noord-Limburg; F.H.M. Welle, director of blood bank Noord-Oost

Brabant; H.P. Olthuis, director of blood bank Nijmegen; Dr. A.C.J.M. Holdrinet,

director of blood bank West-Brabant; Dr. J.Ph.H.B. Sybesma, director of blood

37

bank Zuid-West Nederland and Dr. M. Bins, director of blood bank Zwolle. This

study was made possible by a grant from The Netherlands Organization for

Scientific Research (NWO).

References

Gilfillan A, Axton R, Brock DJH (1994) Mass screening for cystic fibrosis heterozygotes: Two assaysystems compared. Clin Chem 40:197-199

Lench N, Stanier P, Wiliamson R (1988) Simple non-invasive method to obtain DNA for gene analysis.Lancet:1356-1358

Miller SA, Dykes DD, Polesky HF (1980) A simple salting out procedure for extracting DNA from humannucleated cells. Nucleic Acids Res 10:1845

Richards B, Skoletsky J, Shuber AP, Balfour R, Stern RC, Dorkin HL, Parad RB, Witt D, Klinger KW(1993) Multiplex PCR amplification from the CFTR gene using DNA prepared from buccal brushes/swabs.Hum Mol Genet 2:159-163

Scheffer H, Verlind E, Penninga D, Te Meerman G, Ten Kate LP, Buys CHCM (1989) Rapid screening for∆F508 deletion in cystic fibrosis. Lancet:1345-1346

Scheffer H, Mol B, Dijkstra D-J W, Buys CHCM. Allele-specific tests for the CFTR-missense mutationsA455E and A1251N flanking a similar conserved motif in the two nucleotide binding folds with differentclinical phenotypes and occuring relatively frequent in The Netherlands (submitted)

Ten Kate LP (1977) Cystic fibrosis in The Netherlands. Int J Epidem 6:23-35

38

Chapter 3

Prevalence of∆F508 cystic fibrosis carriers in The Netherlands:

region of residence, age, and number of offspring as explanatory

variables

H.G. de Vriesa, J.M. Colléeb, H.E.K. de Wallea, M.H.R. van Veldhuizenb, C.Th.

Smit Sibingac, H. Scheffera, C.H.C.M. Buysa, L.P. ten Kateb.

a Department of Medical Genetics, University of Groningen, The Netherlandsb Department of Human Genetics, Free University of Amsterdam, The Netherlandsc Blood bank Groningen-Drenthe, Groningen, The Netherlands

(submitted)

Abstract

An average cystic fibrosis (CF) carrier frequency of 1 in 25 in Europe is cited in

numerous reports, although a great variability in estimated prevalences has been

found in different European populations. The estimates of these frequencies were

based on numbers of CF patients before identification of the gene in 1989. Here we

report the results of a study to determine the carrier frequency of the∆F508

mutation in The Netherlands by analyzing mouthwashes and matched bloodsamples

from 11,654 blood donors all over the country. In informed consent forms the blood

donors were asked to fill in their age, sex, and number of children. Twenty six

percent of the blood donors did not know that we were looking for a CF mutation.

We analyzed possible relationships between a number of theoretically

explanatory variables and the∆F508 carrier frequency by means of univariate and

multivariate logistic regression. These variables were: 1. Distance of the blood

banks to the northeastern part of the country (distance). This variable was chosen

because a positive correlation between distance and prevalence of CF at birth had

been reported before in this country. 2. Whether the blood donors knew that we

were looking for a CF mutation (CFinfo). 3. Sex of the donor. 4. Age of the donor.

41

5. Number of children of the donor (family size).

We detected a∆F508 carrier frequency of 1 in 42 (95% CI 1/37-1/47) in The

Netherlands. This gives an estimated overall CF carrier frequency of 1 in 31 (95%

CI 1/27-1/35), significantly less than 1 in 25.

The univariate logistic regression analysis of the effects of the explanatory

variables on the carrier frequency revealed no significant relationships, except for an

increase in carrier frequency with increasing distance to the northeastern region. In

the multivariate analysis with all 5 independent variables distance, age and family

size were significantly related to the carrier frequency, but sex and CFinfo were not.

There was a significant interaction between age and family size. In our final model

distance, age and family size were positively related to the carrier frequency, while

the interaction of age with family size showed a negative relation. These results

confirm that there is a gradient in gene frequency with low frequencies in the

northeastern part of the country and high frequencies in the southern part. They also

suggest a relation of age and family size with carrier frequency. This relation,

however, is too complex to be explained by heterozygote advantage.

Introduction

Cystic fibrosis (CF) is the most common fatal autosomal recessive disease in

caucasoid populations. The clinical expression of the disease is variable, but most

patients show chronic pulmonary disease and pancreatic insufficiency. The median

life expectancy for newborns with CF is between 25 and 40 years at the moment

[1].

Generally, in persons of European descent, a prevalence at birth of 1 in 2500

is quoted, which conforms to a carrier frequency of 1 in 25. In reality, however,

previous studies in European populations and populations of European descent

elsewhere have shown a range of birth prevalences from 1 in 1700 to 1 in 6500 [2-

4]. Finland forms an exception with a birthprevalence of 1 in 40,000. Differences

were not only observed between countries but also between regions within countries

[3,5,6]. In The Netherlands for instance a birth prevalence of 1 in 3600 was found

in the seventies, conforming to a carrier frequency of 1 in 30 [3]. Within the

country a gradient was observed with lower frequencies in the northeastern part of

42

the country. It is not clear whether the differences found in previous studies

between countries and between regions are real or were caused by variations in

diagnosis and reporting.

Almost all epidemiologic studies on the prevalence of CF were performed

before the identification of the CFTR-gene in 1989 [7-9]. The present availability of

PCR-based mutation detection methods offers an opportunity for an accurate

determination of carrier frequencies and a reevaluation of the old data.

We here report on a search for∆F508 carriers among 11,654 blood donors in

The Netherlands. We also analyzed the relationship between the∆F508 carrier

frequency and a number of possible explanatory variables: distance of the blood

banks to the northeastern part of the country, whether the blood donors knew that

we were looking for a CF mutation, sex and age of the blood donors and their

number of children. The study also produced data on the validity of testing large

numbers of mouthwash samples. The results of the validity study have been

reported elsewhere [10].

Material and Methods

recruitment and sampling

We approached directors of all 22 blood banks in the country to obtain mouthwash

samples and matched blood samples. Each blood bank serves a geographically

defined region without overlap between them. Out of the 22 blood banks 15 were

willing to co-operate. Blood bank officials were instructed how to recruit volunteers

among the blood donors. 13 of the 15 blood banks adhered to the following

protocol. Informed consent forms consisted of two sticked papers. On the frontpage

readers were asked to participate in a study on the validity of a method to detect a

"genetic trait" in mouthwashes and on the frequency of the trait. Readers then were

given the options of obtaining more information on the trait sought after first or to

join the study anonymously without being informed further and without getting the

test result. In case they wanted further information the sticked paperborders of the

consent form had to be stripped off and data concerning cystic fibrosis, carriership

and genetic counselling were presented at the reverse side of the frontpage. Readers

43

coming this far then had the following options: 1) to participate in the test either

with or without being informed about the results of the test, or 2) not to participate

at all. In all cases the blood donors were asked to fill in their age, sex and number

of children. The design was approved by the ethical committee of the Groningen

University Hospital. This design allowed us to see whether knowledge of the

identity of the trait sought after would influence the participation. For instance if

family members of CF-patients would be more likely to participate, this would bias

our results.

Out of the 15 blood banks two followed a different protocol, approved by

their own ethical committee. Blood donors from these 2 blood banks were not

informed about the identity of the trait which was searched for, and participated

anonymously. In addition blood spots were obtained from the CLB Blood bank.

This blood bank functions as a central facility and receives blood donations from all

regional blood banks. At the time of the sampling only blood spots originating in

one region could be obtained.

We did not give any publicity in the media or elsewhere about this

investigation, first to avoid overrepresentation of donors who have a family member

with CF, and secondly to prevent that people, who already joined this study but

opted not to read the information on the identity of the trait, would obtain

information against their wish after all.

Subjects participating in the mouthwash study were asked to rinse their

mouth with 15 ml of 0.9% sterile saline for 10 seconds. The material then was

collected in sputum containers (Emergo, Landsmeer, The Netherlands). Matched

blood samples were collected in 10 ml heparin tubes.

DNA isolation and (mutation) analysis

Mutation detection was performed on mouthwash DNA. In order to determine the

proportion of false positives, the mutations detected in mouthwash DNAs were

checked for confirmation in the matched blood samples. In case no PCR product

was obtained in mouthwash DNA or no mouthwash DNA was available, mutation

detection was performed on blood samples only. The specificity and sensitivity of

the mouthwash procedure were both 100% (95% CI specificity= 99.96-100%; 95%

44

CI sensitivity= 95.1-100%) [10].

Statistics

The number of participants in each blood bank was not proportional to the

size of the population in the region served by the blood bank. Regional frequencies,

therefore, were adjusted for the size of the population in each region in order to

obtain the average∆F508 carrier frequency in the country. The formula used was as

follows:

Cfreq=∑(Wreg*Creg/Nreg)

with Cfreq = mean carrier frequency of the country, Wreg = regional weight:

proportion of the population in the region with respect to the national population

size, Creg = number of carriers in the region, and Nreg = number of participants in

the region.

An estimate of the overall CF carrier frequency is obtained by dividing the

∆F508 carrier frequency by its relative frequency among CF chromosomes. The

estimated relative frequency of the∆F508 mutation, detected in a total of 532 CF

chromosomes, is 73.3% in The Netherlands (personal communication Hans

Scheffer). The equation used for the calculation of the confidence interval, when

some carriers are not detected, is described by Parker and Phillips [11].

In order to represent possible differences in∆F508 carrier frequencies

between regions, the regional frequencies were represented as standard deviation

scores. The formula of the standard deviation score is:

SDreg = (Creg-Ctot*N reg/Ntot)/(Ctot*N reg/Ntot)1/2

with SDreg = regional standard deviation score, Creg = regional number of carriers,

Ctot = total number of carriers found in this study, Nreg = regional number of

participants, and Ntot = total number of participants in The Netherlands. Areas with

SDreg scores <-2 or >2 accommodate significantly less or more CF carriers

respectively compared with the national average.

45

In view of the number of variables and because the dependent variable was

binary (carrier yes/no) we used the logistic regression model to describe the

relationship between this outcome and a set of independent or explanatory variables.

The independent variables in our model were: 1. Distance of each blood bank to the

northeastern region in km/25 (calculated from latitude and longitude). 2. Whether

the blood donors knew that we were looking for a CF mutation (yes/no). 3. Sex of

the blood donors (male/female). 4. Number of children of blood donors (0,1,2...). 5.

Age of the blood donors (in years). To examine these possible relations we tested

the variables first in a univariate way and after that we used multivariate logistic

regression. The formula of the multiple logistic regression model to calculate the

chance of being a∆F508 carrier is as follows:

p = e[b0+b1x1+b2x2+...+bnxn]/(1+e[b0+b1x1+b2x2+...+bnxn])

p = chance to be a∆F508 carrier

b0= constant regression coefficient

b1 to bn= regression coefficients of the related variables

x1 to xn= (coded) values of the related variables

The statistical package we used was Statistix 4.0.

Results

A total of 11,654 subjects was examined for∆F508 carriership in the various blood

bank regions in The Netherlands (table 1). We were able to obtain and analyze PCR

products from 10,772 out of 11,413 mouthwashes. The failure rate of the

mouthwash procedure is described elsewhere [10]. A total of 254∆F508 carriers

was detected in this group. The remaining 882 analyses were performed on blood

samples only ( 244 from the CLB blood bank). We detected 22∆F508 carriers in

this group. So 276∆F508 carriers were detected among 11,654 subjects.

The overall∆F508 carrier frequency in the country, after adjustment for the

size of the regional population, was 1 in 42 (95% CI 1/37-1/47). The total CF

carrier frequency, including other mutations, therefore, can be estimated at 1 in 31

46

(95% CI 1/27-1/35), based on an estimated relative∆F508 frequency of 73.3%.

The regional distribution of carrier frequencies is shown in figure 1. The

lightshaded areas, which are predominantly located in the northeastern part of the

country harbour less∆F508 carriers than the darkshaded regions in the southern

parts of The Netherlands.

On average 74% of the participants of the 13 blood banks who sticked to the

original protocol read the information on CF (range of 52% to 95%). So 26%

participated without being informed about the identity of the trait that was searched

for. The age of the blood donors ranged from 17 to 69 years, with an average of

40.5 years. The ratio male/female blood donors was 1.5, and the mean number of

their children was 1.6.

The univariate logistic regression analysis of the effects of the explanatory

variables on the carrier frequency revealed a significant relationship for the distance

to the northeastern region (table 2).

Table 1. Number of subjects investigated and number of∆F508 carriers found per blood bank region in TheNetherlands and proportion of the regional population with respect to the national population size.

blood bank regionsa subjects ∆F508 proportioncarriers of population

1. Groningen-Drenthe 2013 34 0.1042. Friesland 770 18 0.0613. Zwolle e.o. 617 9 0.0614. Apeldoorn-Deventer-Zutphen 407 8 0.0455. Noord-Holland-Noord 490 9 0.0566. Arnhem e.o. 705 23 0.0827. Gooi en Eemland 591 12 0.0618. Utrecht 144 3 0.0739. Kennemerland 397 10 0.040

10. Nijmegen 412 13 0.04311. Leiden e.o. 409 9 0.04712. Noord-Oost Brabant 492 14 0.05313. ’s-Gravenhage 603 16 0.09614. Zuid-West Nederland 2024 53 0.07415. Noord-Limburg 424 11 0.04016. West-Brabant 1156 34 0.064

Total 11654 276 1.000

a Blood banks ordered according to distance to the northeastern region

47

Before running the multivariate model we checked the continuous variables if they

were linear in the logit and they were. The results of the multivariate model with all

5 independent variables are shown in table 3a. This full model shows a positive

relation of distance and age with the carrier frequency, while the family size

shows a negative relation. The variables sex and CFinfo were not statistically

significant, so they were omitted in a reduced model (table 3b). The likelihood ratio

test comparing the two models indicated that the reduced model was as good as the

full model. In neither model we made allowance for possible interactions between

variables. There was, however, a significant interaction between age and family size

with respect to being a carrier. Table 3c shows our final model which contains the

variables distance, family size, age and the interaction between age and family size

(age in years * number of children). The goodness-of-fit, calculated with the

Hosmer-Lemeshow statistic [13], of this model revealed a p-value of 0.56. This

indicates that the model seems to fit quite well since the difference between the

expected and the observed results is not significant. Distance, age and family size

all showed a positive affect on carrier frequency, while the interaction showed a

negative effect. The variables age and family size cannot be analyzed separately.

Because of the interaction the odds ratio of family size is not constant over age.

The odds ratios of age, family size and interaction are therefore not shown.

Table 2. Univariate logistic regression analysis relating the∆F508 carrier frequency to the distance to thenortheastern region (distance), whether the blood donors knew that we were looking for a CF mutation(CFinfo), sex of the blood donors (sex), number of their children (family size) and age of the blood donors(age).

variables number ofsubjectsanalyzed

regressioncoefficient (b)

standard error

Se(b)

p odds ratio 95% CIodds ratio

distanceCFinfosexfamily sizeage

11654116549259

1051010501

0.0546-0.1592-0.0230-0.0748

0.0080

0.02050.12650.13510.05000.0058

0.00760.20830.86470.13520.1702

1.060.850.980.931.01

1.01 - 1.100.67 - 1.090.75 - 1.270.84 - 1.021.00 - 1.02

The relation of age and family size with the carrier frequency is also shown

in table 4. For the non-carriers there is a positive correlation between age and

number of children. The relation is different in the carrier group. In this group the

49

family size decreases in the last two age categories, while the carrier frequency

increases with age. These results give an impression of the different associations

between family size and carrier frequency with respect to age.

Table 3. Multivariate logistic regression analysis relating the∆F508 carrier frequency to the distance to thenortheastern region (distance), whether the blood donors knew that we were looking for a CF mutation(CFinfo), sex of the blood donors (sex), number of their children (family size) and age of the blood donors(age).

(a) Full model

variablesa regressioncoefficient (b)

standard errorSe(b)

p odds ratio 95% CIodds ratio

constantdistanceCFinfosexfamily sizeage

-4.54690.0566

-0.12600.0787

-0.12450.0182

0.40410.02070.15040.14050.06010.0069

0.00630.40200.57550.03840.0085

1.060.881.080.881.02

1.02 - 1.100.66 - 1.180.82 - 1.420.78 - 0.991.00 - 1.03

a Analyzed in 9247 subjects

(b) Omitting CFinfo and sex

variablesa regressioncoefficient (b)

standard errorSe(b)

p odds ratio 95% CIodds ratio

constantdistancefamily sizeage

-4.41010.0517

-0.12750.0145

0.27620.02040.05820.0065

0.01130.02860.0258

1.050.881.01

1.01 - 1.100.79 - 0.991.00 - 1.03

a Analyzed in 10488 subjects

(c) Final model

variablesa regressioncoefficient (b)

standard errorSe(b)

p odds ratio 95% CIodds ratio

constantdistance (x1)family size (x2)age (x3)age*family size (x4)

-5.19510.05060.61750.0353

-0.0176

0.36500.02040.22150.00870.0051

0.01300.00530.00010.0006

1.05 1.01 - 1.09

a Analyzed in 10488 subjects

50

Table 4. The relation of the family size and carrier frequency with the age in 10 years categories.

non-carriers ∆F508 carriers

age subjects family size subjects family size carrier frequency

=<2021-3031-4041-5051-60>60

170 0.011799 0.303098 1.623222 1.991487 2.28462 2.62

3 0 1.7%37 0.49 2.0%77 1.59 2.4%87 1.80 2.6%31 1.68 2.0%15 1.60 3.1%

Discussion

A total of 11,654 blood donors was examined for∆F508 carriership in The

Netherlands. Seventy four percent of the donors knew that we were looking for

carriers of the∆F508 mutation. We did not find a difference in carrier frequency

between the groups of informed and non-informed blood donors. This indicates that

we did not introduce a selection bias by giving the blood donors information about

the trait that was searched for.

We have found a∆F508 carrier frequency of 1 in 42 in the Dutch

population. This corresponds with an estimated total CF carrier prevalence of 1 in

31, based on an estimated relative∆F508 frequency of 73.3%. The estimated overall

CF carrier frequency of 1 in 31 is significantly different from the generally quoted

carrier frequency of 1 in 25 (p<0.005). This supports the results of an earlier study

[3] in which an estimated birth prevalence of one CF patient in 3600 newborns was

obtained, conforming to a carrier frequency of one in 30. Because of the observed

biological variation it is incorrect to refer to a common carrier frequency of 1 in 25

for all European populations. The carrier frequencies have to be established

separately for each population. Our study also shows that results of old but properly

performed studies on the birth prevalence of CF may be dependable.

The estimate of the relative frequency of the∆F508 mutation on CF

chromosomes (73.3%) is mainly based on prevalent cases in stead of incident cases.

The DNA of Dutch patients with the clinical manifestations of CF has been

collected for the last 6-10 years. Because of the high infant mortality rate of CF

patients in the past, which was possibly predominantly correlated with the presence

51

of a ∆F508 mutation, samples of these patients may be underrepresented.

Chromosomes with non-∆F508 mutations, often associated with less severe

phenotypes, therefore may have accumulated in the surviving patients. This would

lead to an underestimation of the∆F508 mutation in the present patient population.

There is, however, also a possibility of underdiagnosis of CF patients with non-

∆F508 mutations associated with a mild phenotype. These arguments indicate that

there may be some variance in the estimate of the relative∆F508 frequency on CF

chromosomes.

In our study the variable distance shows a significant positive relation with

the carrier frequency, in both the univariate and multivariate logistic regression

analyses. This geographical gradient in the∆F508 carrier frequency with high

prevalences in the southern part and low prevalences in the northeastern part also

confirms a previous finding of a lower birth prevalence of CF in the north-east of

the country [3].

It is not possible as yet to give a detailed description of the geographic

distribution of the total CF carrier frequency, including non-∆F508 mutations,

because not all non-∆F508 mutations are randomly distributed in the Dutch

population; especially the A455E mutation is very concentrated in the southern parts

of The Netherlands (unpublished results), and the identity of 15% of all CF

mutations is unknown in this country at the moment.

In the multivariate logistic regression analysis we found a positive effect of

age and family size on carrier frequency, while there was an interaction term

involving the same variables which was negatively correlated with carrier

frequency. A possible explanation for a positive effect of age and family size on

carrier frequency is heterozygote advantage either at present or in the past [14]. If

heterozygote advantage is still effective today, this would result in a higher life

expectancy and a larger number of children of carriers and therefore in their

overrepresentation in older people and people with more children (selection effect).

Alternatively, if heterozygote advantage was only effective until recently but no

more today (e.g. by disappearance of an environmental threat) carrier frequency

would decrease in time, resulting in a lower carrier frequency among younger

people than among older ones (cohort effect). In a cross-sectional study it is not

possible to differentiate between these two alternatives.

It is generally assumed that a positive selection of heterozygotes of about 2%

52

would be sufficient to maintain the CF carrier frequency at present levels.

Relaxation of selection also would result in relatively small changes in carrier

frequency. One would not expect to be able to demonstrate such small effects in a

study of this scale. In both cases, however, existing differences in carrier

frequencies between age groups might be exaggerated by studying blood donors,

because they are selected for their health.

There is another phenomenon in our data which we cannot explain by

heterozygote advantage, viz. the negative effect of the interaction between age and

number of children. This effect is so strong that carrier frequency increases with age

only for persons with less then 3 children. Likewise it increases with family size

only for persons below 40 years of age. So, for people of 40 years and over or with

3 or more children the negative effect of the interaction is stronger than the positive

effect of age and family size, resulting in a decrease in carrier frequency. This

phenomenon does not fit in with expectations from the heterozygote advantage

hypothesis and we have not been able to advance a coherent explanation. The

findings are unexpected and should first be confirmed in independent studies before

we can draw firm conclusions.

Acknowledgements

The authors want to thank the following persons and blood banks for making

mouthwashes and blood samples available: Th.M. Brouwers, director of blood bank

Apeldoorn-Deventer-Zutphen; G.C. van der Plas, director of blood bank Arnhem

e.o.; H.J.C. de Wit, director of blood bank Friesland; Mrs J.H. Barmentloo, director

of blood bank Gooi en Eemland; Dr. J.A. van der Does, director of blood bank ’s-

Gravenhage e.o.; R.C. Schaap, director of blood bank Kennemerland; Mrs dr. A.

Brand, director of blood bank Leiden e.o.; Dr. L.H. Siegenbeek van Heukelom,

director of blood bank Noord-Holland-Noord; Dr. H.J.A. Sonderkamp, director of

blood bank Noord-Limburg; F.H.M. Welle, director of blood bank Noord-Oost

Brabant; H.P. Olthuis, director of blood bank Nijmegen; Dr. A.C.J.M. Holdrinet,

director of blood bank West-Brabant; Dr. J.Ph.H.B. Sybesma, director of blood

bank Zuid-West Nederland; Dr. M. Bins, director of blood bank Zwolle and Dr.

L.P. de Waal, director of the CLB Blood bank, Amsterdam. The authors are also

53

very grateful to the blood donors for their spontaneous co-operaration.

This study was made possible by a grant from The Netherlands Organization for

Scientific Research (NWO).

References

1. Welsh MJ, Tsui L-C, Boat TF, Beaudet AL: Cystic fibrosis; in Scriver CR, Beaudet AL, Sly WS,Valle MD (eds): The Metabolic and Molecular Basis of Inherited Disease. New York, McGraw-Hill,1994,vol 3,pp 3799-3876.

2. Wood RE, Boat TF, Doershuk CF: Cystic fibrosis. Am Rev Respir Dis 1976;113:833-878.

3. Ten Kate LP: Cystic fibrosis in The Netherlands. Int J Epidem 1977;6:23-35.

4. Warwick WJ: The incidence of cystic fibrosis in Caucasian populations. Helv Paediat Acta 1978;33:117-125.

5. Brunecky Z: The incidence and genetics of cystic fibrosis. J Med Genetics 1972;9:33-37.

6. Cystic fibrosis Genetic Analysis Consortium: Population variation of common cystic fibrosis mutations.Hum Mut 1994;4:167-177.

7. Rommens JM, Iannuzi MC, Kerem B, Drumm ML, Meimer G, Dean M, Rozmahel R, Cole JL, KennedyD, Hidaka N, Zsiga M, Buchwald M, Riordan JR, Tsui L-C, Collins FS: Identification of the cysticfibrosis gene: chromosome walking and jumping. Science 1989;245:1059-1065.

8. Riordan JR, Rommens JM, Kerem B, Alon N, Rozmahel R, Grzelczak Z, Zielenski J, Lok S, Plavsic N,Chou J-L, Drumm ML, Iannuzzi MC, Collins FS, Tsui L-C: Identification of the cystic fibrosis gene:cloning and characterization of complementary DNA. Science 1989;245:1066-1073.

9. Kerem B, Rommens JM, Buchanan JA, Markiewicz D, Cox TK, Chakravarti A, Buchwald M, Tsui L-C:Identification of the cystic fibrosis gene: genetic analysis. Science 1989;245:1073-1080.

10. De Vries HG, Collée JM, Veldhuizen MHR, Scheffer H, Ten Kate LP: Validation of determination of∆F508 mutations of the cystic fibrosis gene in almost 11,500 mouthwashes. Hum Genet (in press).

11. Parker RA, Phillips III JA: Population screening for carrier status: Effects of test limitations on precisionof carrier prevalence rates. Am J Hum Genet 1994;49:317-322.

12. Micky J, Greenland S: A study of the impact of confounder selection criteria on effect estimation. Am Jepidem 1989;129:125-137.

13. Hosmer DW, Lemeshow S: Applied logistic regression. A Wiley-Interscience

14. Jorde LB, Lathrop GM: A test of the heterozygote advantage hypothesis in cystic fibrosis carriers. Am JHum Genet 1988;42:808-815.

54

Chapter 4

Relative frequencies of CFTR mutations in The Netherlands:

Regional variation is present even in a small country

J. Margriet Collée1, Hendrik G. de Vries2, Hans Scheffer2, Dicky J.J. Halley3, Leo

P. ten Kate1

1 Department of Human Genetics, Free University of Amsterdam, The Netherlands2 Department of Medical Genetics, University of Groningen, The Netherlands3 Department of Clinical Genetics, Erasmus University, Rotterdam, The Netherlands

(submitted)

Abstract

Cystic fibrosis (CF) is one of the most common autosomal recessive disorders in

white populations. Significant regional differences of CF mutations among affected

individuals have been reported.

We studied the geographic distribution of the relative frequencies of the three

most common Dutch CF mutations,∆F508, A455E, and G542X, by analyzing data

on residences of CF patients.

Significantly higher relative frequencies of the A455E mutation and the G542X

mutation were observed in the South-west and the South-east, respectively. For the

∆F508 mutation a rather uniform distribution of relative frequencies was found. The

results of our study show that even in a small country like The Netherlands certain

CF mutations may be more common in one region than in another.

Introduction

Cystic fibrosis (CF) is one of the most common autosomal recessive disorders in

white populations. The classic clinical picture is characterised by chronic pulmonary

57

disease, pancreatic insufficiency and elevated concentrations of sweat electrolytes

[1]. In The Netherlands the birth prevalence of CF has been estimated to be 1 in

3600, corresponding to a carrier frequency of 1 in 30 [2].

Since the identification of the CF gene and its major mutation∆F508 in 1989

[3,4,5], over 500 mutations have been described. In a recent study we found a

∆F508 carrier frequency of one in 42 in the Dutch population. This corresponds

with an estimated total CF carrier prevalence of 1 in 31, based on an estimated

relative ∆F508 frequency of 73.3% [6]. The second and third most common

mutations, A455E and G542X, represent approximately 3 and 2%, respectively [7].

All other identified mutations were detected in less than two percent of all CF

chromosomes each.

As significant regional differences in the relative frequencies of CF mutations

have been reported -with France as an example [8,9]- and migration from one to

another part of the country used to be uncommon in The Netherlands until only a

few decades ago, we studied regional differences in relative frequencies of Dutch

cystic fibrosis mutations. In view of the small numbers of other mutations only

∆F508, A455E and G542X were treated individually.

Patients and methods

In The Netherlands DNA testing for CF mutations is performed in two centres for

Clinical Genetics, one located in the west (Erasmus University Rotterdam) and one

in the North (University of Groningen). Both laboratories receive blood samples

from patients throughout the country. In the summer of 1994 samples from 634

independent Dutch CF patients (i.e. 1268 alleles) had been analyzed for diagnostic

purposes. Of the patients studied we collected data on the results of CF mutation

detection, on the residential area (residence and province), and on the age of the

patient at receipt of the blood specimen. Patients were excluded if no information

on their residence was available.

DNA testing

The detection of the∆F508, A455E, and G542X mutations in Rotterdam was

58

performed as described by Kerem et al. [10]. In Groningen the∆F508 mutation

detection was carried out according to Scheffer et al. [11] and the analysis for the

G542X mutation as described by Ferrie et al. [12]. The A455E mutation was

detected with a specific amplification refractory mutation system (ARMS)(HS,

unpublished).

Data analysis

The Netherlands consists of 12 provinces. First, the distribution of the∆F508

mutation over the 12 provinces was compared with the distribution of the non-

∆F508 mutations. Departure from a random distribution was tested for by means of

the chi-square test (with 11 degrees of freedom). Next, in view of the small

expected numbers of the A455E and G542X mutations in most provinces, provinces

were combined in three groups. Group 1 consisted of the Northern and Eastern

provinces (1-6 in figure 1), group 2 comprised the provinces called Noord-Holland,

Zuid-Holland and Utrecht (7-9 in figure 1) which contain the major cities of the

country, and group 3 included the remaining Southern provinces (10-12 in figure 1).

This grouping may seem rather arbitrary but takes account of some natural barriers:

the former Zuider Zee between group 1 and group 2, and the major rivers of Rhine

and Meuse between group 3 and the rest of the country. The distribution of the

∆F508, A455E, G542X, other known mutations and unknown mutations over the

three groups of provinces was then analyzed (chi-square test with 8 degrees of

freedom). To study the distribution of the∆F508, A455E and G542X in more

detail, we calculated standard deviation (SD) scores for the proportion of the

particular mutation in each of the 12 Dutch provinces. The expression used for the

SD score was as follows:

Dx = (Ax-At*nx/nt)/(At*nx/nt)½

Dx = standard deviation score per province x

Ax = number of CF chromosomes with a particular mutation found in province x

At = number of CF chromosomes with a particular mutation found in The

Netherlands

nx = number of CF chromosomes studied in province x

59

nt = number of CF chromosomes studied in The Netherlands

In provinces with standard deviation scores of <-2 or >2, the mutation under

study forms a significantly lesser or greater proportion of CF chromosomes in that

area when compared to the national average.

The major CF clinics, which have integrated DNA diagnostics for CF more

routinely in patient care (unpublished results, Dutch CF registry), are not randomly

distributed throughout The Netherlands. We, therefore, did not calculate frequencies

relative to the size of the population of the region.

Results

Of the 634 CF patients we were able to retrieve information on places of

residence in all but 36 (5.7%) cases. This resulted in a total of 1196 CF

chromosomes that were included in the analysis. In this study population the∆F508

mutation was present in 75% of the CF chromosomes. The relative frequencies of

the A455E and G542X mutations were 3.7% and 2.2% respectively. Other identified

mutations accounted for 7.4% of CF chromosomes. In 11.5% of the CF

chromosomes no known mutation was found.

The relative frequencies of∆F508, A455E, G542X, other identified CF

mutations, and unknown CF mutations by region are shown in table 1. When

comparing the distribution of∆F508 over the 12 provinces to the distribution of

non-∆F508 mutations (all together), no significant differences were seen indicating

that its relative frequency is rather uniformly distributed. When comparing the

distributions of the∆F508, A455E, G542X, other known mutations, and unknown

mutations over the three groups of provinces a significant deviation from a random

distribution was observed (chi2 = 22.1, p<0.01). This high score is caused mainly by

an abnormal distribution of the A455E and G542X mutations, while the∆F508 and

the other mutations contributed much less to the chi-square. SD scores of the

relative frequencies of∆F508, A455E and G542X were calculated by province, and

are shown in figure 1. The∆F508 mutation was uniformly distributed without SD

values more or less than 2. However, significantly more CF chromosomes in the

South-West of The Netherlands carried the A455E mutation: in the provinces Zuid-

60

Holland and Zeeland this mutation was identified in 6.2% (SD= 2.45) and 11.1%

(SD=2.84) of the chromosomes, respectively. The only other province in which the

A455E mutation was relatively frequent was Utrecht (SD=1.38), a central region

that flanks Zuid-Holland.

For the G542X mutation a significantly larger proportion when compared

with the national average was observed in Limburg (8.6%, SD=3.33), in the South-

east of the country. In the adjacent province Noord-Brabant this mutation was also

relatively frequent (SD=1.27).

When looking at the ages of CF patients at the time of DNA analysis a

remarkable dissimilarity was observed between patients with different mutations.

The mean age for patients with a double∆F508 mutation or a∆F508 mutation and

a G542X mutation (only three homozygotes for G542X were found) was 11.5 years

and 8.9 years respectively. The mean age for patients with a∆F508 and an A455E

mutation (no homozygotes for A455E were found) was 23.3 years. No patients were

identified who were compound heterozygotes of an A455E and a G542X mutation.

Table 1. Numbers of∆F508, A455E, G542X, and other mutations by province in Dutch CF patients

Provincea ∆F508 A455E G542X Otherknownmutations

Unknownmutations

Total numberof CFchromosomes

Group 1

Groningen (1)Friesland (2)Drenthe (3)Overijssel (4)Flevoland (5)Gelderland (6)

Group 2

Utrecht (7)Noord-Holland (8)Zuid-Holland (9)

Group 3

Zeeland (10)Noord-Brabant (11)Limburg (12)

282025551368

63142251

3615445

100100

53

22

651

000001

067

075

403328

31624

6154

34431

11

32952

6193

362432621688

74196356

5420058

The Netherlands 900 44 26 43 138 1196

61

Discussion

The results of our study show that in The Netherlands certain CF mutations among

CF patients are more common in one area than in another. We observed this for the

A455E and the G542X mutation, but not for the∆F508 mutation.

With respect to the∆F508 mutation we earlier performed a study on the

frequency of∆F508 carriers among blood donors which showed a lower prevalence

of carriers in the North of the country [6]. As in the present study no differences

were observed in relative frequencies for the∆F508 mutation between the Dutch

provinces, this suggests a lower prevalence of CF and CF carriership (all mutations

together) in the North. In fact this phenomenon has been suggested before in other

studies on the prevalence of CF in the Netherlands [2,13].

The A455E mutation has mainly been described in The Netherlands and in

Canada [7]. In the latter it has been introduced during the period 1650-1900 [14,15).

Some of the Canadian and Dutch patients share a common haplotype over distances

up to 25 cM surrounding the A455E mutation [16], which suggests that this

mutation has been derived from a not too distant common predecessor. Furthermore,

it is associated with a less severe CF phenotype [17], which results in a higher

prevalence among adult CF patients. In this study this is reflected by the higher age

of patients with a A455E mutation when compared to those with a∆F508 mutation.

As one of the major clinics for adult CF patients is located in the province Zuid-

Holland, in the west of the country, this confronted us with a possible confounding

factor. In order to distinguish between a truly high prevalence of the A455E

mutation in the South-West and a confounder effect, we repeated the data analysis

by calculating standard deviation scores for CF patients under 16 years of age (706

CF chromosomes, A455E in 1.4%). This resulted again in significantly higher

frequencies and deviation scores for Zuid-Holland (3.5%, SD=2.28) and Zeeland

(9.4%, SD= 3.78), while the relative frequencies of the A455E mutation in other

provinces were not different from the results in the total group, thus confirming our

first observation. Similarly analysis of the distribution of the∆F508 and the G542X

mutations in patients under 16 years of age confirmed the results in the total group.

The G542X mutation has been detected in various populations [7] with a

tendency to be more frequent in South European countries. While the A455E

mutation gives rise to mild disease, the G542X mutation is associated with more

63

severe symptoms and pancreatic insufficiency [18]. In this study we found a higher

frequency of this mutation in the South, predominantly in the province of Limburg.

In conclusion, the results of this study show that even in a relatively small

country like The Netherlands regional differences of CF mutations among affected

individuals can be observed.

Acknowledgements

This study was made possible by a grant from The Netherlands Organization for

Scientific Research (NWO).

References

1. Welsh MJ, Tsui L-C, Boat TF, Beaudet AL: Cystic fibrosis; in Scriver CR, BeaudetAL, Sly WS, ValleMD (eds): The Metabolic and Molecular Basis of Inherited Disease. New York, McGraw-Hill,1994,vol3,pp 3799-3876.

2. Ten Kate LP. Cystic fibrosis in The Netherlands. Int J Epidem 1977;6:23-35.

3. Rommens JM, Iannuzzi MC, Kerem B, et al. Identification of the cystic fibrosis gene: chromosomewalking and jumping. Science 1989;245:1059-1065.

4. Riordan JR, Rommens JM, Kerem B, et al. Identification of the cystic fibrosis gene: cloning andcharacterization of complementary DNA. Science 1989;245:1066-1073.

5. Kerem B, Rommens JM, Buchanan JA, et al. Identification of the cystic fibrosis gene: genetic analysis.Science 1989;245:1073-1080.

6. De Vries HG, Collée JM, Van Veldhuizen MHR, Smit Sibinga CTh, Scheffer H, Buys CHCM, Ten KateLP. Prevalence of∆F508 cystic fibrosis carriers in The Netherlands: region of residence, age, and numberof offspring as explanatory variables (submitted)

7. Cystic Fibrosis Genetic Analysis Consortium. Population variation of common cystic fibrosis mutations.Hum Mut 1994;4:167-177.

8. Férec C, Audrezet MP, Mercier B, Guillermit H, Moullier P, Quere I, Verlingue C. Detection of over 98%cystic fibrosis mutations in a Celtic population. Nature Genet 1992;1:188-191.

9. Claustres M, Laussel M, Desgeorges M, Giansily M, Culard JF, Razakatsara G, Demaille J. Analysis ofthe 27 exons and flanking regions of the cystic fibrosis gene: 40 different mutations account for 91.2% ofthe mutant alleles in Southern France. Hum Mol Genet 1993;2:1209-1213.

10. Kerem B, Zielenski J, Markiewicz D, Bozon D, Gazit E, Yahav J, Kennedy D, Riordan JR, Collins FS,Rommens JM, Tsui L-C. Identification of mutations in regions corresponding to the two putative

64

nucleotide (ATP)-binding folds of the cystic fibrosis gene. Proc Natl Sci USA 1990;87:8447-8451.

11. Scheffer H, Verlind E, Penninga D, Te Meerman G, Ten Kate LP, Buys CHCM. Rapid screening for the∆F508 deletion in cystic fibrosis. Lancet 1989;1345-1346.

12. Ferrie RM, Schwartz MJ, Robertson NH, Vaudin S, Super M, Malone G, Little S. Development,multiplexing, and application of ARMS tests for common mutations in the CFTR gene. Am J HumGenet 1992;51:251-262.

13. Cornel MC, Te Meerman GJ, Ten Kate LP. Jaarlijkse inventarisatie van kystische fibrose in Nederland.T Soc Gezondheidsz 1985;63:266-269.

14. Rozen R, Schwartz RH, Hilman BC, Stanislovitis P, Horn GT, Klinger K, Daigneault J, De BraekeleerM, Kerem B-S, Tsui L-C, Fujiwara TM, Morgan K. Cystic fibrosis mutations in North Americanpopulations of French ancestry: analysis of Quebec French Canadian and Louisiana Acadian families.Am J Hum Genet 1990;47:606-610.

15. Rozen R, De Braekeleer M, Daigneault J, Ferreira-Rajabi L, Gerdes M, Lamoureux L, Aubin G, SimardF, Fujiwara TM, Morgan K. Cystic fibrosis mutations in French Canadians: Three mutations arerelatively frequent in a Quebec population with an elevated incidence of cystic fibrosis. Am J Med Genet1992;42:360-364.

16. De Vries HG, Van der Meulen MA, Rozen R, Halley DJJ, Scheffer H, Ten Kate LP, Buys CHCM, TeMeerman GJ. Haplotype identity between individuals who share a CFTR mutation allele Identical ByDescent: demonstration of the usefulness of the haplotype sharing concept for gene mapping in realpopulations (submitted)

17. Gan KH, Veeze HJ, Van den Ouweland AMW, Halley DJJ, Scheffer H, Van der Hout A, Overbeek SE,De Jongste JC, Bakker W, Heijerman HGM. Evidence for a cystic fibrosis mutation associated with mildlung disease. N Engl J Med 1995:333:95-99.

18. Kristidis P, Bozon D, Corey M, Markiewitcz D, Rommens J, Tsui L-C, Durie P. Genetic determinationof exocrine pancreatic function in cystic fibrosis. Am J Hum Genet 1992;50:1178-1184.

65

Chapter 5

Number and sex of offspring of ∆F508 carriers outside CF

families

HG de Vries1, JM Collée2, WP Meeuwsen1, H Scheffer1, LP ten Kate2

1 Department of Medical Genetics, University of Groningen, The Netherlands.2 Department of Human Genetics, Free University, Amsterdam, The Netherlands.

Hum. Genet. (1995) 95:575-576

Abstract

The number and sex of offspring was determined in a group of 7,841 random

selected blood donors who were screened for the∆F508 mutation. We did not find

evidence for differences in number or sex ratio of offspring between∆F508 carriers

and non-carriers.

In previous studies there have been indications of increased fertility of cystic

fibrosis (CF) carriers (Danks et al. 1965) and sex ratio distortion in their offspring

(Gloria-Bottini et al. 1980, Pritchard et al. 1983 ), which could be causally related

to the high birth prevalence of CF. Part of these observations may however have

resulted from ascertainment biasses (Ten Kate 1977; Jorde and Lathrop 1988).

The characterisation of the gene in 1989 (Kerem et al. 1990) enables a new

approach to be used in offspring studies by collecting data of subjects without a

family history of CF. To illustrate this possibility we determined number and sex of

offspring in blood donors who were screened for the∆F508 mutation (77% of CF

mutations in The Netherlands) by mouthwash Polymerase chain reaction (PCR) and

compared this data in carriers and non-carriers. Mouthwashes and data on offspring

were sampled simultaneously. The presence of∆F508 in mouthwashes was

confirmed by analysis in matched blood-samples.

So far 186∆F508 carriers have been identified in a group of 7,841 blood

67

donors. To determine sibship-size we here show data of the 4,116 blood donors

who were 40 years old and over. Most of them will have completed their families.

Data on the sex of children was determined in the total sample.

As shown in table 1, the number of offspring of CF carriers is 1.83 against

2.12 for non-carriers. The average number of children of female carriers is 1.86

against 1.82 for male carriers. In both comparisons the differences are not signifi-

cant. The sex ratio (male/female) in the offspring of carriers and in children of non-

carriers is 1.01. The sex ratio in offspring of male carriers does not differ signifi-

cantly from sexratio in children of female carriers (1.00 and 1.05 respectively).

Table 1. Number and sex of offspring of∆F508 carriers

Subject Number of Sex ofchildren children

Type Number Total Mean Number Male Female Ratio

Carrier

Male 771,2 140 1.82 1103 93 93 1.00

Female 291,2 54 1.86 673 46 44 1.05

Total 1062 194 1.83 1863 139 137 1.01

non-carrier

total 4,010 8,490 2.12 7,655 6,493 6,441 1.01

1 Male/female ratio in all 7,841 blood donors was 2.3/12 Blood donors 40 years and over3 Sex unknown in 9 blood donors

So far we have no evidence for differences in number or sex ratio of

offspring of ∆F508 carriers and non-carriers. As the project continues data will

accumulate until approximately 14,000 blood donors will have been screened

(expected number of carriers 325).

68

References

Danks DM, Allan J, Anderson CM. A genetic study of fibrocystic disease of the pancreas. Ann hum Genet1965;28:323-356.

Gloria-Bottini F, Antonelli M, Quattrucci S. et al. Is there a relationship between sex of cystic fibrosiscarriers and sex ratio of their offspring? Hum Genet 1980;54:79-82.

Jorde LB, Lathrop GM. A test of heterozygote-advantage hypothesis in cystic fibrosis carriers. Am J HumGenet 1988;42:808-815.

Kerem B-S, Rommens JM, Buchanan JA et al. Identification of the cystic fibrosis gene: genetic analysis.Science 1989;245:1073-1080

Pritchard DJ, Hickman GR, Nelson R. Sex ratio and heterozygote advantage in cystic fibrosis families. ArchDis Child 1983;58:290-293.

Ten Kate LP. A method for analysing fertility of heterozygotes for autosomal recessive disorders, withspecial reference to cystic fibrosis, Tay-Sachs disease and phenylketonuria. Ann hum Genet 1977;40:287-297.

69

Chapter 6

Haplotype identity between individuals who share a CFTR

mutation allele Identical By Descent: demonstration of the

usefulness of the haplotype sharing concept for gene mapping in

real populations

Hendrik G. de Vries1, Martin A. van der Meulen1, Rima Rozen2, Dickie J.J. Halley3,

Hans Scheffer1, Leo P. ten Kate4, Charles H.C.M. Buys1 and Gerard J. te Meerman1

1 Department of Medical Genetics, University of Groningen, The Netherlands2 Departments of Human Genetics, Pediatrics and Biology, Montreal, Canada3 Department of Clinical Genetics, Erasmus University, Rotterdam, The Netherlands4 Department of Human Genetics, Free University of Amsterdam, The Netherlands

(submitted)

Abstract

Cystic fibrosis (CF) patients with the A455E mutation, both in the French Canadian

and in the Dutch population, share a common haplotype over distances up to 25

cM. French Canadian patients with the 621+1G→T mutation share a common

haplotype of more than 14 cM. In contrast, haplotypes containing the∆F508

mutation show haplotype identity over a much shorter genomic distance within and

between populations, probably due to multiple introduction of this most common

mutation.

Haplotype analysis for specific mutations in CF or in other recessive diseases

is a model for studying the occurrence of genetic drift conditional on gene

frequencies. Moreover, from our results it can be inferred that analysis of shared

haplotypes is a suitable method for genetic mapping in general.

71

Introduction

In recent times it has been repeatedly observed that haplotypes surrounding rare

alleles of a gene are quite large [1-9]. Sharing of large genomic areas can be used

as a method to map disease genes: Identity By Descent (IBD) Mapping [4,10]. An

empirical question is whether haplotype sharing can be observed in real populations

to an extent where IBD mapping using haplotype sharing is feasible.

As an empirical model for high and low frequency disease alleles weakly

associated with a disease or with dominance and low penetrance, we have chosen to

study mutations leading to the recessive disease cystic fibrosis (CF). Only a very

small fraction of the disease alleles present in the population can be observed in

diseased individuals. Thusfar more than 500 presumed mutations have been

identified in the CFTR gene. The most common mutation (∆F508) has a high

frequency in Caucasian populations (up to 1.5%). A less frequent mutation is the

A455E missense mutation. According to data from the cystic fibrosis consortium

[11] this mutation is mainly detected among French Canadian and Dutch CF

patients. The A455E mutation comprises 8% of all CF mutations in the French

Canadian population of the Saguenay-Lac St. Jean region of Quebec [12] and 3% in

The Netherlands [unpublished data]. The overall CF carrier frequency is 1 in 15 in

the particular French Canadian population [12] and 1 in 30 in the Netherlands [13].

This results in allele frequencies of 1/400 (8% of 1/30) and 1/2000 (3% of 1/60)

respectively. The concentrated geographical distribution indicates that this mutation

has been introduced not very long ago.

In models for multifactorial disease the mutated gene concept is not directly

applicable since alleles act as risk factors in combination with other factors.

However, if gene/gene or gene/environment interactions are modelled in founder

populations, where the multifactorial background is more in common than in mixed

populations, specific alleles of genes may be involved, leading to association at the

population level. The implication is that such alleles show increased frequencies in

affected individuals. It cannot be ruled out that multifactorial diseases are

determined by only very few genes, which would make an increased frequency of

specific alleles in affected individuals even more likely [14]. By studying haplotype

sharing in two populations, one with a very late origin (French Canadians) and one

with a longer history (Southern part of The Netherlands) we demonstrate the

72

usefulness of the haplotype sharing concept for gene mapping in real populations.

Materials and methods

Recruitment of CF patients

The A455E mutation, which is associated with a less severe CF phenotype, has

mainly been detected in Canada and in The Netherlands [11]. In Canada this

mutation has been introduced by French immigrants during the period 1650-1900

[8,12]. The Dutch patients with an A455E mutation all come from the southern

parts of The Netherlands (unpublished results). Bloodsamples from 15 independent

Dutch CF patients with the A455E mutation were collected in Groningen and

Rotterdam. Samples from 10 French Canadian CF patients with the A455E muta-

tion, from a subpopulation in north-eastern Quebec-the Saguenay-Lac St.Jean region

[8,12], were collected in Montreal. The father of proband 3 and the mother of

proband 6, which are both carriers of the 621+1G→T mutation, were sibs. Because

all patients were compound heterozygotes, haplotype sharing could also be

determined around two other CF mutations (∆F508, 621+1G→T). In all patients

clinical diagnosis was confirmed by demonstrating the∆F508 [15], A455E [16] and

621+1G→T [17] mutations. In all cases DNA from one of the parents was used for

phase determination.

Microsatellite analysis

Three intragenic microsatellites, IVS8CA (intron 8), IVS17BTA, and IVS17BCA

(intron 17b) [18] and 7 extragenic microsatellites, D7S518, D7S501, D7S523,

D7S486, D7S480, D7S490, D7S635 [19], were analyzed by single amplification.

The primer sequences are shown in Table 1. PCR was performed with 400 ng of

DNA, 50 mM KCl, 10 mM Tris (pH 9.0), 1.5 mM MgCl2, 0.01% gelatin, 0.1%

Triton X-100, 0.5 µM of each microsatellite primer (biotinylated) and 0.25 U of

Super Taq (SphearoQ), in a total volume of 50 µl. Part of the PCR product was

diluted to 1:20 and mixed with formamide solution. Absolute PCR product sizes

73

Table 1. PCR primers of microsatellites at and flanking the CF locus. Data have been derived from Zielenskiet al [18] and Gyapay et al [19].

Marker Primersequences 5’→3’ Heterozygosity Distance toCFTR gene (cM)

D7S518 CAGTAGGCAGGGGTGG 0.87 15GGGTGTGTCTGTGTGACAAC

D7S501 CACCGTTGTGATGGCAGAG 0.81 7ATTTCTTACCAGGCAGACTGCT

D7S523 CTGATTCATAGCAGCACTTG 0.80 1AAAACATTTCCATTACCACTG

D7S486 AAAGGCCAATGGTATATCCC 0.81 0.2GCCAGGTGATTGATAGTGC

IVS8BTA TCTATCTCATGTTAATGCTG 0.41 0GTTTCTAGAGGACATGATC

IVS17BTA GACAATCTGTGTGCATCG 0.89 0GCTGCATTCTATAGGTTATC

IVS17BCA AAACTTACCGACAAGAGGA 0.38 0TGTCACCTCTTCATACTCAT

D7S480 CTTGGGGACTGAACCATCTT 0.86 2AGCTACCATAGGGCTGGAGG

D7S490 CCTTGGGCCAATAAGGTAAG 0.78 5AGCTACTTGCAGTGTAACAGCATTT

D7S635 CCAGGCCATGTGGAAC 0.81 10AGTTCTTGGCTTGCGTCAGT

were determined by adding standard size markers (generated from Pharmacia

m13mp18) to each sample. The reaction products were detected in a 6% denaturing

polyacrylamide gel with an Automated Laser Fluorescent system (Pharmacia LKB).

In total a region of 25 cM flanking the CF gene was genotyped. For certain

microsatellite alleles it was not possible to establish phase. In these cases, haploty-

pes have been assigned by comparing alleles of flanking microsatellites between the

CF patients.

74

Results

Table 2 shows the results of the haplotype analysis of the Canadian and Dutch CF

patients with a A455E mutation. The haplotype of the intragenic markers on the

A455E chromosome of the 10 French Canadian patients was identical in all cases,

namely 22-35-13, in which the numbers represent the numbers of repeats of the

IVS8CA, IVS17BTA and IVS17BCA respectively. The extragenic alleles are

numbered according to the relative length of the PCR product.

Table 2. Microsatellite haplotypes for independent A455E chromosomes from a French Canadian and aDutch population.

5 3 2 Intragenic 1 6 8 (cM)^ ^ ^ / \ ^ ^ ^

D7S-635 490 480 8CA 17BTA 17BCA 486 523 501 518

French Canadianchromosomes

1 7! 1 3 22! 35! 13 8 4 5 102 7 1 3 22 35! 13 8 4! 5 103 7 1 3 22 35 13 8 4 5! 104 3! 4 3 22! 35 13 8 4 5 105 3 4 3 22! 35 13 8 4 5 106 3 4 3! 22 35 13 8 4 5! 107 1! 1 2 22! 35 13 8 4 5 28 1 1 2 22 35 13 8 4! 4 109 4! 2 2 22 35 13 8 4 5 10

10 7 1! 3 22 35 13 8 3 3 2

Dutchchromosomes

1 7 1 3 22 35! 13 8 4 3 22 1 1 3 22! 35 13 8 4 1 123 1 4 6 22 35 13 8 4 1! 4!4 6 1 3 22 35 13 8 5 7 25 6 1 3 22 35 13 8 5 7 96 6 1 3 22! 35 13 8 5 5 27 6 1! 3 22 35 13 10 5 5 28 1 1 3! 22 35 13 8 5! 3 119 4 1 3 22 37 13 8 5 3 10

10 3 1 3 22! 35 13 8 3 3 12!11 5 6 3 22 35 13 8 5 6 212 8 2 3 22! 35 13 8 5 7 213 5! 6 3 22 35 13 8 3 7! 1014 10 2 3 22 35 13 8! 2 3 315 1 4 5 22 35! 13 8 5 7 3

! phase unknown

75

Of the 10 Canadian patients, 9 share a region of at least 5 cM, surrounding the

A455E mutation (Figure 1). The patients 1, 2 and 3 share a region of more than 25

cM. With a different haplotype this is also the case for patients 4, 5 and 6. The

haplotype of the intragenic markers of 14 of the 15 Dutch patients are identical to

the Canadian patients. They share a region of more than 4 cM surrounding A455E

(Figure 2). The remaining patient showed a different intragenic haplotype (22-37-

13). The haplotypes of the non-A455E chromosomes of the Canadian and Dutch

patients are shown in Table 3.

Table 3. Microsatellite haplotypes for independent∆F508 and 621+1G→T chromosomes from a FrenchCanadian and a Dutch population.

5 3 2 Intragenic 1 6 8 (cM)^ ^ ^ / \ ^ ^ ^

D7S-635 490 480 8CA 17BTA 17BCA 486 523 501 518

French Canadianchromosomes

2 621+1G→T 1 4 5 21 31! 13 8 5! 7 33 621+1G→T 1 4 5 21 31 13 8 5 1! 66 621+1G→T 1 4 5! 21 31 13 8 5 7! 59 621+1G→T 1! 4 5 21 31 13 8 5 5 21 DF508 9! 3 1 17! 32! 13 8 7 1 37 DF508 9! 3 1 17! 32 13 8 1 5 24 DF508 7! 4 1 17! 32 13 8 7 1 45 DF508 9 6 7 17! 32 13 5 5 7 58 DF508 2 4 3 17 35 17 3 5! 8 7

10 DF508 8 4! 2 23 31 13 11 5 3 4

Dutchchromosomes

5 DF508 6 1 6 17 31 13 10 4 3 27 DF508 1 6! 8 17 31 13 10 6 5 98 DF508 4 2 7! 17 31 13 10 3! 6 11

15 DF508 1 6 5 17 31! 13 10 2 7 41 DF508 4 1 3 17 31! 13 3 3 4 29 DF508 9 6 3 17 31 13 8 4 1 76 DF508 7 3 3 17! 32 13 8 2 1 2

11 DF508 1 4 7 17 32 13 8 6 5 914 DF508 1 2 1 17 32 13 5! 3 7 74 DF508 3 3 2 17 35 17 10 5 4 2

12 DF508 8 1 2 17! 35 17 10 2 7 413 DF508 4! 1 5 17 47 13 10 4 8! 33 DF508 1 2 1 23 31 13 8 5 5! 10!

! phase unknown

76

Dystrophia and rickettsia genes has been demonstrated [20]. The only aberrant

dutch haplotype (22-37-13) can best be explained by independent introduction,

although there is also a possibility of a mutation in the original haplotype.

The ∆F508 mutation is a very old mutation, which has been introduced at

least 52,000 years ago in Europe [21]. Much variation is observed in intragenic

markers, which is probably due to repeat length mutations subsequent to the dF508

mutation. This most frequent CF mutation has been widely distributed in especially

Caucasian populations over a long period of time. In the Saquenay-Lac St.Jean

region we detected three different intragenic haplotypes in only six patients with a

∆F508 mutation. This means that there were most likely multiple independent

introductions of the∆F508 mutations in this population. Sharing of DNA regions

surrounding∆F508 mutations is limited to individuals with the same haplotype of

intragenic markers. The observed intragenic haplotypes are also the most common

haplotypes in Europe [21]. To detect haplotype sharing surrounding∆F508

mutations, more observations with identical intragenic haplotypes are necessary.

If the size of a shared haplotype is large, the common predecessor must be

quite recent. The difference in the extent of sharing found for the∆F508 mutation

and for the other mutations, shows that haplotype sharing can be expected to occur

only if there are not too many independent introductions of a gene in the

population. Haplotype sharing and other association studies can therefore be

expected to be successful in founder populations of a proper size relative to the

gene frequency. Independent introduction of about 10-20 copies will because of

genetic drift lead to perhaps 1-4 copies remaining in widely varying numbers of

individuals after 8 or more generations. This is in agreement with the theoretical

expectation according to Fisher [22], who computed that the probability of survival

of a single gene after 10 generations in a stable population is about 10%. It appears

from the French Canadian genealogical data, that gene flow cannot be traced back

to unique predecessors, even if all genealogical data are complete. The informativity

of genealogical data seems therefore limited in comparison to what is being learned

from direct haplotype comparison.

The similarity in intragenic ∆F508 haplotypes in the two populations

suggests multiple introduction of the same intragenic haplotypes, of which some

have survived. In Finland genetic drift has by this process led to a low CF

frequency and to a distribution of mutations different from that of the rest of

79

Europe [23].

The suitability of a population for haplotype sharing studies will increase if

there is a larger population with which the founder population is connected. Once a

genomic region has been identified which possibly contains an allele predisposing

to a disease, it is easy to perform a genome screening in affected individuals at the

1 cM level. This situation seems to exist both for the French Canadians and in The

Netherlands, where more isolated areas are connected with larger populations. At

the genomic distances of 1 cM association and linkage disequilibrium can be

detected in populations much larger than the one originally investigated. At the

same time, a large number of meioses is being observed implicitly, thus narrowing

the region where the gene must be located. This is well illustrated from the

consistent overlap shown in and between Figures 1 and 2. In The Netherlands, the

A455E data come from an unselected population, which shows that for rare genes,

such as the A455E mutation, founder effects are present at the population level. A

complication in haplotype comparison is marker allele mutation. In this dataset we

probably observe two such mutations (the Dutch patients number 7 and 9 show an

allele mutation in the microsatellites D7S486 and IVS17BTA, respectively) where

the haplotype for surrounding markers seems to be conserved.

The present study shows that the suitability of populations for gene mapping

through direct haplotype comparison can be investigated using some of the well

known recessive disease mutations that can easily be detected at the level of

carriers. These mutations have generally a quite high frequency, as would be

expected for alleles of genes involved in multifactorial diseases. By relating the

gene frequency to the size of the shared haplotypes, an impression can be obtained

whether haplotype sharing analysis can be used to find the map location of other

genes.

Acknowledgements

The authors thank Dr. J.P. Heyerman (Leyenborg Hospital, The Hague) for sending

DNA of the parents of Dutch CF patients. This study was made possible by a grant

from The Netherlands Organization for Scientific Research (NWO).

80

References

1. Cannon-Albright LA, Goldgar DE, Gruis NA, Neuhasen S, Anderson DE, Lewis CM, Jost M, Tran TDand Nyguen K: Localization of the 9p melanoma susceptibility locus to a 2cM region between D9S736and D9S171, Genomics 1994;23:265.

2. Gruis N, Sandkuijl LA, Bergman W and Frants RR. Common 9p haplotype in Dutch FAMMM families,Genetics of the familial Atypical Multiple Mole-Melanoma Syndrome, in PhD thesis, Leiden, 1994.

3. Nystrom-Lahti M, Sistonen P, Mecklin J-P, Pylkkanen L, Aaltonen LA, Jarvinen H, Weissenbach J, Dela Chapelle A and Peltomaki P: Close linkage to chromosome 3p and conservation of ancestral foundinghaplotypes in hereditary nonpoloyposis colorectal cancer families. Proc Natl Acad Sci USA1994;91:6054-6058.

4. Houwen RHJ, Baharloo S, Blankenship K, Raeymaekers P, Juyn J, Sandkuijl LA and Freimer NB:Genome screening by searching for shared segments: mapping a gene for benign recurrent intrahepaticcholestasis. Nature Genet 1994;8:380-386.

5. Hastbacka J, De la Chapelle A, Mahtani MM, Clines G, Reeve- Daly MP, Daly M, Hamilton BA,Kusumi K, Triveldi B, Weaver A, Coloma A, Lovett M, Buckler A, Kaitila I and Lander ES: Thediastrophic dysplasia gene encodes a novel sulphate transporter: positional cloning by fine-structurelinkage disequilibrium mapping. Cell 1994;78:1073-1087.

6. Sulisalo T, Francomano CA, Sistonen P, Maher JF, McKusik VA, De la Chapelle A, and Kaitila I: High-resolution genetic mapping of the cartilage-hair hypoplasia (CHH) gene in Amish and Finnish families.Genomics 1994;20:347-353.

7. Meyers DA, Postma DS, Panhuysen CIM, Xu J, Amelung PJ, Levitt RC and Bleecker ER: Evidence fora locus regulating total serum IgE levels mapping to chromosome 5. Genomics 1994;23:464.

8. Rozen R, Schwartz RH, Hilman BC, Stanislovitis P, Horn GT, Klinger K, Daigneault J, De BraekeleerM, Kerem B-S, Tsui L-C, Fujiwara TM and Morgan K: Cystic fibrosis mutations in North Americanpopulations of French ancestry: Analysis of Quebec French Canadian and Louisiana Acadian Families.Am J Hum Genet 1990;47:606-610.

9. Heyer E and Tremblay M: Variability of the genetic contribution of Quebec population foundersassociated to some deleterious genes. Am J Hum Genet 1995;56:970-978.

10. Te Meerman GJ, Van der Meulen MA and Sandkuijl LA: Perspectives of Identity By Descent (IBD)mapping in founder populations. Clin Exp Allergy 1995;in press.

11. Cystic fibrosis Genetic Analysis Consortium: Population variation of common cystic fibrosis mutations.Hum Mut 1994;4:167-177.

12. Rozen R, De Braekeleer M, Daigneault J, Ferreira-Rajabi L, Gerdes M, Lamoureux L, Aubin G, SimardF, Fujiwara TM and Morgan K: Cystic fibrosis mutations in French Canadians: Three CFTR mutationsare relatively frequent in a Quebec population with an elevated incidence of cystic fibrosis. Am J MedGenet 1992;42:360-364.

13. Ten Kate LP: Cystic fibrosis in The Netherlands. Int J Epidem 1977;6:23-35.

14. Van Ommen GJ: A foundation for limb-girdle muscular dystrophy. Nature Medicine 1995;1:412-414.

15. Scheffer H, Verlind E, Penninga D, Te Meerman G, Ten Kate LP, Buys CHCM: Rapid screening for∆F508 deletion in cystic fibrosis. Lancet 1989:1345-1346.

81

16. Kerem B, Zielenski J, Markiewicz D, Bozon D, Gazit E, Yahav J, Kennedy D, Riordan JR, Collins FS,Rommens JM, Tsui L-C: Identification of mutations in regions corresponding to the two putativenucleotide (ATP)-binding folds of the cystic fibrosis gene. Proc Natl Acad Aci USA 1990;87:8447-8451.

17. Zielenski J, Bozon D, Kerem B, Markiewicz D, Durie P, Rommens JM, Tsui L-C: Identification ofmutations in exons 1 through 8 of the cystic fibrosis transmembrane conductance regulator (CFTR) gene.Genomics 1991;10:229-235.

18. Zielenski J, Markiewicz D, Rininsland F, Rommens JM and Tsui L-C: A cluster of highly polymorphicdinucleotide repeats in intron 17b of the CFTR gene. Am J Hum Genet 1991;49:1256-1262.

19. Gyapay G, Morissette J, Vignal A, Dib C, Fizames C, Millasseau P, Marc S, Bernardi G, Lathrop M andWeissenbach J: The 1993-1994 Généthon human genetic linkage map. Nature Genet 1994;7:246.

20. Bétard, C, Raeymaekers P, Ouellette G, Jomphe M, Labuda M, Glorieux F, Mathieu J, Laberge C,Cassiman J-J, Sandkuijl LA and Gauvreau D: The validation of a novel linkage disequilibrium mappingtechnique on Steinert myotonic dystrophy and pseudovitamin D deficient rickets in a founder population.Med Genetik 1995;2:238 (abstract).

21. Morral N, Bertranpetit J, Estivill X, Nunes V, Casals T, Giménez J, Reis A, Varon-Mateeva R, Macek JrM, Kalaydjieva L, Angelicheva D, Dancheva R, Romeo G, Russo MP, Garnerone S, Restagno G,Ferrari M, Magnani C, Claustres M, Desgeorges M, Schwartz M, Schwartz M, Novelli G, Ferec C, DeArce M, Nemeti M, Kere J, Anvret M, Dahl N and Kadasi L: The origin of the major cystic fibrosismutation (∆F508) in European populations. Nature Genet 1994;7:169-175.

22. Vogel F and Motulski AG: Human genetics 3rd ed. Springer- Verlag, Berlin, 1986.

23. De la Chapelle A: Disease gene mapping in isolated human populations: the example of Finland. J MedGenet 1993;30:857-865.

82

Summary

This thesis describes the application of molecular techniques in population genetic

studies of cystic fibrosis (CF), the most common serious autosomal recessive

disorder affecting Caucasian populations. Birth prevalences in European countries

range from 1 in 1700 to 1 in 40,000. The disease is characterized mainly by chronic

obstruction and infection of the respiratory tract, meconium ileus, pancreatic

insufficiency, and elevated levels of sweat electrolytes (Chapter 1). The gene

responsible for CF has been identified in 1989. It has been named cystic fibrosis

transmembrane conductance regulator (CFTR) gene, is located on the long arm of

chromosome 7, is approximately 230,000 kb in size, and contains 27 exons. Thus

far more then 500 mutations have been identified in the CFTR gene. A variety of

mechanisms has been suggested to account for the high CF carrier frequency of 1 in

21 to 1 in 100 in the Caucasian population: genetic heterogeneity, high mutation

rate, genetic drift, meiotic drive, increased fertility of CF carriers, heterozygote

advantage and segregation distortion of CF alleles.

Approximately 75% of the CF chromosomes in The Netherlands harbour a

three basepair deletion in CFTR, which removes the phenylalanine residue at amino

acid position 508 of the CFTR polypeptide (∆F508). This mutation can easily be

detected by the polymerase chain reaction (PCR) - amplification of the flanking

DNA-sequence - and subsequent gel electrophoresis of the amplfied product. In

most laboratories blood is used as the source for DNA isolation. Blood sampling

and DNA isolation in great numbers have, however, a few disadvantages: costs of

DNA isolation are high, (para)medical personnel is necessary for blood collection,

people may shrink from venapuncture, and there is a risk of exposure to blood

pathogens. An alternative is sampling of mouthwashes followed by DNA isolation

from buccal cells. This procedure is much simpler and cheaper than existing

methods, especially with large numbers of samples. There is no need for medical

supervision of sample collection, and the risk of infections is eliminated.

For the validation of the mouthwash procedure, we collected a total of over

11,000 mouthwashes and matched blood samples from 15 blood banks and also 75

mouthwashes of known CF carriers (Chapter 2). Both the specificity and the

sensitivity of the mouthwash procedure were 100%. The overall failure rate was 1.8

85

% (excluding the results of one particular blood bank with a markedly high failure

rate). These results indicate that for large numbers of samples the mouthwash

procedure is suitable for mutation detection and can be used in community

screening.

We used the samples above to determine the prevalence of carriers of the

∆F508 mutation. The∆F508 carrier frequency in The Netherlands turned out to be

1 in 42, which gives an estimated overall CF carrier frequency of 1 in 31. This CF

carrier frequency is significantly less than the generally quoted frequency for

Europe of 1 in 25. Logistic regression analysis revealed possitive relationships

between the∆F508 carrier frequency and the region of residence, age of the blood

donors, and number of children of the blood donors. The distribution of the carrier

frequency displayed a geographical gradient in the∆F508 carrier frequency with

high prevalences in the southern part and low prevalences in the north-east of the

country. The positive relation of age and family size with the carrier frequency may

be explained as a consequence of hetrozugote advantage, but should be confirmed

in further studies.

Chapter 4 describes the frequency and distribution of the non-∆F508

mutations in The Netherlands. Thusfar, 17 less frequent mutations have been

detected. The A455E mutation with a relative frequency of about 2.8 % has

predominantly been detected in the south-western parts of The Netherlands.

A possible explanation for the high prevalence of cystic fibrosis in the

Caucasian population is a high fertility of CF carriers. We collected data from more

than 7000 blood donors to determine the sex ratio and number of offspring of

∆F508 carriers and non-carriers (Chapter 5). We were unable to detect significant

differences between both groups. This means that we cannot support the hypothesis

of increased fertility of CF carriers.

The A455E mutation, which is associated with a less severe CF phenotype,

has mainly been detected in Canada (in a subpopulation in northeastern Quebec -the

Saguenay-Lac St.Jean region) and in The Netherlands (south-western parts). The

concentration of the mutation indicates that it has been introduced not very long

ago. The A455E mutation therefore, may be suitable for a haplotype sharing

analysis between ’unrelated’ CF patients. DNA of 15 independent Dutch and 10

French Canadian CF patients with the A455E mutation were collected for haplotype

analysis. Each group of CF patients with the A455E mutation, shared a common

86

haplotype both intragenic and extragenic over distances up to 25 cM. In contrast,

the haplotypes of the∆F508 mutation showed much shorter haplotype identity

within and between these populations, probably due to multiple introduction of this

very old, but common mutation. The method of haplotype comparison for CF is a

model for studying the occurrence of genetic drift conditional on gene frequencies,

and indirectly demonstrates the suitability of empirical populations for genetic

mapping by haplotype sharing analysis.

87

Samenvatting

In elke menselijke cel met een kern bevinden zich 23 paar chromosomen.

Chromosomen bestaan voor een groot gedeelte uit DNA. Maar een klein gedeelte

(3%) van dit DNA komt in eiwitten tot expressie. Het DNA is een keten van

bouwstenen. Er zijn vier verschillende bouwstenen en de volgorde hiervan in het

DNA is bepalend voor de eigenschappen van ons lichaam. Elk chromosomenpaar

bevat een chromosoom van de moeder en één van de vader. De aanleg voor elke

eigenschap op de autosomen is daardoor in tweevoud aanwezig. Een stuk DNA met

de informatie voor een erfelijke eigenschap noemen we een gen. Bij de mens

komen 50.000 - 100.000 genen voor. In bepaalde gevallen kunnen er veranderingen

(mutaties) optreden in de DNA volgorde van de genen. Als deze mutaties

voorkomen in de voortplantingscellen, dan kunnen ze doorgegeven worden aan de

volgende generatie.

Cystic fibrosis (CF), ook wel taaislijmziekte genoemd, is een erfelijke

aandoening die zich meestal in de eerste levensjaren openbaart. De ziekte gaat ge-

paard met herhaalde longontstekingen, spijsverteringsproblemen en een te zout

zweet. Deze aandoening komt voornamelijk voor bij blanken (de Caucasische

volkeren). CF is een ernstige en nog niet te genezen ziekte. Door een zorgvuldige

behandeling kunnen de symptomen van CF steeds beter bestreden worden. Het

aantal patiënten in de volwassen leeftijd neemt dan ook toe, hoewel nog steeds pati-

enten op jeugdige leeftijd overlijden. In Nederland komt CF voor bij ongeveer 1 op

de 3600 kinderen. CF is een recessieve aandoening. Dit wil zeggen dat de ziekte

zich pas openbaart als zowel van de vader als van de moeder een chromosoom met

een mutatie in het CF-gen wordt verkregen. Er wordt gesproken van dragerschap

voor CF wanneer een CF mutatie in enkelvoud aanwezig is (ook deze is weer af-

komstig van de vader óf van de moeder). Dragers hebben naast de afwijkende

aanleg een normale aanleg die sterker is dan de afwijkende. Dragerschap voor CF

heeft dan ook geen gevolgen voor de eigen gezondheid. In Nederland is ongeveer 1

op de 30 mensen drager van CF.

In 1989 hebben onderzoekers de bouwsteen volgorde van het CF gen kunnen

bepalen. Hierdoor is het mogelijk geworden om mutaties in dit gen aan te tonen.

Dit is van groot belang voor het opsporen van CF patiënten en dragers. Het CF gen

is een relatief groot gen. Tot nu toe zijn er al meer dan 500 verschillende mutaties

89

in dit gen gevonden. De meest voorkomende mutatie is de∆F508 (spreek uit delta

F508) mutatie die in Nederland bij ongeveer 75% van de afwijkende CF-genen

wordt gevonden. De verschillende mutaties, maar ook andere invloeden, kunnen

voor een van patiënt tot patiënt wisselend ziektebeeld zorgen. Hierdoor is het voor

artsen soms zeer lastig om louter op grond van symptomen vast te kunnen stellen of

iemand CF heeft. Met behulp van DNA technieken is het veelal mogelijk om in

twijfelgevallen de diagnose CF te bevestigen of ontkrachten. Tevens kan door

middel van DNA onderzoek dragerschap vastgesteld worden. Er kan dan bekeken

worden of een (echt)paar dat kinderen wil, een risico (25%) op CF bij hun

kind(eren) heeft. In geval van een risico-zwangerschap is het mogelijk om door

middel van een vruchtwater punctie of een vlokkentest te onderzoeken of het kind

wel of niet CF zal krijgen.

In een in dit proefschrift beschreven onderzoek is gekeken naar het vó-

órkomen van CF dragers en patiënten in Nederland. Het onderzoek kan opgesplitst

worden in vijf delen. 1) Bepaling van de betrouwbaarheid van DNA isolatie uit

mondspoelsels. 2) Vaststellen van het aantal dragers van de∆F508 mutatie in

verschillende regio’s in Nederland. 3) Onderzoek naar de verspreiding van niet-

∆F508 mutaties. 4) Vergelijking van het aantal en het geslacht van kinderen van

∆F508 dragers en niet dragers. 5) Vergelijking van stukken DNA van Nederlandse

en Frans Canadese CF patiënten met een A455E mutatie.

1) DNA wordt gewoonlijk uit bloed geïsoleerd. Er zijn echter enige nadelen

verbonden aan het nemen van bloedmonsters voor het onderzoeken van grote

aantallen personen: de kosten voor de DNA isolatie zijn relatief hoog, de

bloedafname moet plaatsvinden door (para) medisch personeel, mensen zien er

tegenop om geprikt te worden, en er is een klein gevaar van besmetting voor zowel

de donor als de onderzoeker. Een alternatief voor bloedmonsters is de

mondspoelselprocedure. Door middel van het spoelen van de mond met zout water

worden wangslijmvlies cellen verkregen, waaruit DNA geïsoleerd kan worden. In

dit onderzoek werden meer dan 11.000 bloeddonoren bereid gevonden een

mondspoelsel af te staan. In 94.4% van alle gevallen was het mogelijk DNA hieruit

te isoleren. Het relatief grote aantal mislukte DNA isolaties was voor het merendeel

het gevolg van het onvoldoende spoelen bij één bepaalde bloedbank. Na het

weglaten van deze resultaten van die bloedbank steeg het percentage succesvolle

DNA isolaties naar 98.2%. In die gevallen waarin de∆F508 mutatie werd

90

aangetoond in mondspoelsel DNA, werd de uitslag gecontroleerd door DNA isolatie

uit een bijbehorend bloedmonster. In al deze gevallen was het mogelijk om de

dragerschap te bevestigen. Tevens werden 75 mondspoelsels van bekende∆F508

dragers onderzocht om te bepalen of in alle gevallen deze mutatie ook aan te tonen

was. In 73 mondspoelsels werd dragerschap bevestigd. Uit twee mondspoelsel kon

geen DNA worden geïsoleerd. Hieruit blijkt dat de mondspoelsel procedure een

betrouwbare en geschikte methode is voor DNA bepalingen bij grote aantallen

personen (hoofdstuk 2).

2) Met behulp van deze mondspoelsels was het mogelijk om voor het hele

land en per provincie het aantal dragers voor de∆F508 mutatie te bepalen. In heel

Nederland bleek 1 op de 42 personen deze mutatie te hebben. Als dit omgerekend

wordt naar de totale dragerschapsfrequentie, dus ook de dragers meerekenend die

een andere CF mutatie hebben, dan komt dit neer op 1 op 31. Met behulp van een

statistisch reken programma werd aangetoond dat de dragerschapsfrequentie in het

zuiden hoger was dan in het noorden van Nederland. Tevens werd gevonden dat

leeftijd en aantal kinderen positief gerelateerd was met de dragerschapsfrequentie.

Dit wil zeggen dat oudere personen met een gezin een grotere kans hebben om

drager te zijn van de∆F508 mutatie dan jongere mensen zonder kinderen. Deze

resultaten zouden kunnen wijzen op een heterozygoot voordeel. Heterozygoot

voordeel wil zeggen dat dragers van de mutatie beter beschermd zijn tegen

bijvoorbeeld het (ver)krijgen van bepaalde besmettelijke ziektes. Voordat er echter

voorbarige conclusies getrokken worden uit deze resultaten is het van belang om ze

eerst te bevestigen in onafhankelijk onderzoeken (hoofdstuk 3).

3) Hierboven werd al vermeld dat in eerder onderzoek in Nederland in 75 %

van alle CF mutaties de∆F508 werd aangetoond. Tot nu toe zijn 17 andere mutaties

in Nederland aangetoond met elk een relatieve frequentie van minder dan 3%. Om

de verspreiding van deze mutaties in Nederland te onderzoeken, werden gegevens

over de woonplaatsen van de CF patiënten verzameld (hoofdstuk 4). Deze gegevens

vormden tevens een aanvulling op de verspreiding van CF dragers in Nederland. Uit

dit onderzoek bleek dat ook andere niet-∆F508 mutaties niet willekeurig over

Nederland verspreid waren. Met name de A455E mutatie, met een frequentie van

2.8% werd voornamelijk in Zeeland en Zuid-Holland aangetoond. De geringe ver-

spreiding leverde een vervolg onderzoek op naar de oorsprong van deze mutatie, dat

beschreven wordt in hoofdstuk 6.

91

4) Er is veel onderzoek uitgevoerd om een verklaring te vinden voor de hoge

frequentie van CF bij de Caucasische bevolking. Twee mogelijke theorieën zijn: a)

een heterozygoot voordeel en b) een verhoogde fertiliteit van dragers. Een bekend

voorbeeld van heterozygoot voordeel is dragerschap voor sikkelcel-anemie, wat

bescherming biedt tegen malaria. In geval van CF zouden de dragers mogelijk

minder kans hebben op het krijgen van cholera. Dit voordeel is echter bij de mens

nog niet aangetoond. Verhoogde fertiliteit (vruchtbaarheid) van dragers betekent dat

CF dragers meer kinderen zouden krijgen dan niet-dragers. In hoofdstuk 5 is een

onderzoek beschreven waarin het geslacht en het aantal kinderen van∆F508 dragers

en niet-dragers met elkaar vergeleken worden. Er werd geen verschil aangetoond

tussen beide groepen. Dit betekent dat de tweede theorie met dit onderzoek niet

ondersteund kan worden. Voor het verwerpen ervan zijn echter grotere aantallen

nodig.

5) Uit een onderzoek waarbij gegevens werden verzameld van CF patiënten

uit heel Europa kwam naar voren dat de∆F508 mutatie waarschijnlijk minstens

52,000 jaar oud was. De mutatie zou zijn geïntroduceerd in Europa vanuit Klein

Azië. De minder frequente mutaties zijn waarschijnlijk veel minder oud. Dit is te

zien aan de geringere verspreiding van deze mutaties in verschillende landen. De

A455E mutatie is bijvoorbeeld alleen veelvuldig aangetoond in het zuiden van

Nederland en in een Frans-Canadese populatie in Canada. Om uit te zoeken of

patiënten met deze mutatie in de verre verte familie van elkaar zijn, werd het DNA

in de regio rondom het CF gen bij deze patiënten onderling vergeleken. Personen

die nauw met elkaar verwant zijn zullen namelijk een groter stuk DNA

overeenkomstig hebben dan personen waarbij dit niet het geval is. De methode

waarbij DNA regio’s vergeleken worden bij patiënten met een identieke erfelijke

aandoening wordt IBD mapping genoemd. IBD staat voor "Identity By Descent" en

betekent "overeenkomst door afstamming". Deze methode kan ook gebruikt worden

voor het opsporen van nog onbekende genen. Hiervoor is DNA nodig van patiënten

met een één en dezelfde erfelijke aandoening, die afkomstig zijn uit eenzelfde

gebied. Alle chromosomen worden dan onderzocht om vergelijkbare DNA regio’s te

vinden. In deze regio’s kan het verantwoordelijke gen zich bevinden. Omdat de

A455E mutatie bijna uitsluitend in de bovengenoemde populaties is gevonden, zijn

patiënten met deze mutatie uit die populaties zeer geschikt om de IBD methode te

testen. Uit het onderzoek bij 15 Nederlandse en 10 Frans-Canadese patiënten bleek

92

dat een klein stukje DNA rondom het CF gen in alle gevallen identiek was. Bij een

aantal Frans Canadese patiënten werden zelfs vrij grote overeenkomstige stukken

DNA aangetoond. Bij de Nederlandse patiënten waren deze DNA regio’s iets korter,

maar er was wel een duidelijke verwantschap met de Canadese patiënten. Mogelijk

hebben deze patiënten dan ook dezelfde voorouder en is de mutatie vanuit Frankrijk

in Nederland en in Canada terecht gekomen. Uit de resultaten blijkt dat de IBD

methode in founderpopulaties, dat wil zeggen geïsoleerde kolonies die ontstaan zijn

uit fracties van grotere oorspronkelijke populaties, goed toepasbaar is (hoofdstuk 6).

93

Curriculum vitae

Name: Hendrik Gerardus de Vries

Date of birth: April 14, 1965

Place of birth: Roosendaal

1978-1984 HAVO, Leeuwarden

1984-1988 Hogere Laboratorium Opleiding, Leeuwarden

Specialization: medical microbiology

1988-1991 Medical biology, University of Groningen

1991-1995 PhD student on a project funded by the Netherlands

Organization for Scientific Research (NWO). Validity and yield

of a screeningtest for the detection of carriers of the cystic

fibrosis gene. Department of Medical Genetics, University of

Groningen, The Netherlands.

94

Dankwoord

In de afgelopen vier jaar hebben een groot aantal personen er voor gezorgd dat mijn

onderzoek geresulteerd heeft in dit proefschrift. Al deze personen wil ik hartelijk

bedanken voor hun aandeel. Een aantal van hen wil ik bij name noemen. In de

eerste plaats Leo ten Kate. Hoewel jij in Amsterdam niet altijd even goed te

bereiken was, heb je wel de tijd gevonden voor de evaluaties van dit project en de

kritische beschouwingen van de manuscripten. Tot mijn grote spijt ben ik er niet

aan toe gekomen om oude botten te onderzoeken. Charles Buys, nadat jij positief

had gereageerd op het verzoek om als co-promotor op te treden, was jij bereid om

mij met raad terzijde te staan. Hans Scheffer wil ik bedanken voor de praktische

begeleiding van dit project. Hierdoor was ik in staat om mijn onderzoek in

Groningen voort te zetten.

Dit project was geheel afhankelijk van de spontane medewerking van alle

bloedbanken, het CLB en alle bloeddonoren. In het b zonder wil ik Dr C.Th. Smit

Sibinga, directeur van de bloedbank Noord-Nederland, bedanken voor zijn bijdrage

aan de opzet van deze studie.

Zonder de hulp van een aantal analisten was ik nooit in staat geweest om dit

project tot een goede einde te brengen. Bart Mol, wiens muziek keuze toch niet de

mijne was; Marjan van Veldhuizen, die als echte Friezin in het verre Amsterdam

haar experimenten uitvoerde; Wilma Meeuwsen, die vol enthousiasme over haar

kinderen kon vertellen en Jolanda Kliphuis, met wie ik veel gesprekken heb gehad

op ons labzaaltje. Hoewel dit project praktisch niet altijd even interessant was,

hebben jullie mij een hoop werk uit handen genomen. Hiervoor mijn hartelijke

dank.

Margriet Collée, ik ben jou zeer erkentelijk voor jouw aandeel in dit project.

Je hebt heel wat afgereisd om alle bloedbanken te benaderen en her en der praatjes

te verzorgen. Jij hebt mij doen inzien dat er met artsen toch een goede

samenwerking mogelijk is. Gerard en Martin kwamen met het idee om IBD toe te

passen op CF patiënten en dragers. Deze samenwerking resulteerde in een inte-

ressant artikel. Rima Rozen en Dicky Halley waren bereid om DNA’s van CF

patiënten naar mij op te sturen. Edwin wil ik bedanken voor de ALF resultaten.

Mentje en Tinke W ben ik zeer erkentelijk voor het feit dat ze niet geërgerd waren

(of dat althans niet lieten merken) wanneer ik zo’n tien maal per dag hun kamer

95

binnen stormde. Ze waren ook altijd spontaan bereid om een onwetende OIO te

helpen met zijn problemen. Menke wil ik bedanken voor het in no time klaarmaken

van mooie dia’s. Hermien was bereid om in de laatste fase "multiple logistische

regressie" los te laten op mijn resultaten. De term alleen al misstaat niet in een

artikel. Ik wil je hierbij van harte bedanken voor deze bijdrage. Eva Pliester bedank

ik voor het invoeren van de onderzoeksgegevens in de computer.

Gerrit, ik vond het bijzonder leuk om "de meest populaire collega" als

kamergenoot te hebben. Dank zij jou heb ik niet alleen meer inzicht gekregen in

"the never ending story" van SMA, maar mijn bridge spel is nu ook sterk verbeterd.

Verder wil ik alle collega’s van de vakgroep Medische Genetica bedanken

voor de prettige tijd die ik hier heb doorgebracht.

De leden van de promotie commissie Prof J.J. Cassiman, Prof A. Hofman en

Prof H.S.A. Heymans wil ik bedanken voor het lezen van de thesis.

Mijn ouders hebben mij altijd gestimuleerd om te gaan studeren. Zonder die

steun had ik die stap waarschijnlijk niet genomen. Ook voor hun betrokkenheid

tijdens mijn promotie project ben ik hen bijzonder dankbaar.

Regina (Ient), je weet het.

96