Genet Congenital non-syndromal autosomal recessive ... · 3MedGenet 1995;32:336-343 Congenital non-syndromal autosomal recessive deafness in Bengkala, an isolated Balinese village

3 Med Genet 1995;32:336-343

Congenital non-syndromal autosomal recessivedeafness in Bengkala, an isolated Balinesevillage

Sunaryana Winata, I Nyoman Arhya, Sukarti Moeljopawiro, John T Hinnant,Yong Liang, Thomas B Friedman, James H Asher Jr

Department ofMicroscopic Anatomy,Faculty of Medicine,Udayana University,Denpasar, Bali,IndonesiaS Winata

Department ofBiochemistry, Facultyof Medicine, UdayanaUniversity, Denpasar,Bali, IndonesiaI N Arhya

Laboratory of Biology,Faculty of Biology andthe Inter UniversityCenter forBiotechnology, GadjahMada University,Yogyakarta, IndonesiaS Moeljopawiro

Department ofAnthropology,Michigan StateUniversity, EastLansing, Michigan48824, USAJ T Hinnant

Graduate Program inGenetics, MichiganState University, EastLansing, Michigan48824, USAY Liang

Graduate Program inGenetics, andDepartment ofZoology, 203 NaturalScience Building,Michigan StateUniversity, EastLansing, Michigan48824, USAT B FriedmanJ H Asher Jr

Correspondence to:Dr Asher.

Received 24 July 1994Revised version accepted forpublication 6 December1994

AbstractBengkala is an Indonesian village locatedon the north shore of Bali that has existedfor over 700 years. Currently, 2-2% of the2185 people in this village have profoundcongenital deafness. In response to thehigh incidence of deafness, the people ofBengkala have developed a village specificsign language which is used by many ofthe hearing and deaf people. Deafness inBengkala is congenital, sensorineural,non-syndromal, and caused by a fullypenetrant autosomal recessive mutation atthe DFNB3 locus. The frequency of theDFNB3 mutation is estimated to be 9-4%among hearing people who have a 17-2%chance of being heterozygous for DFNB3.

(J Med Genet 1995;32:336-343)

A medical and genetic analysis of deafness inBengkala, Bali began in 1990 with the iden-tification of deaf subjects and their re-

lationships. An anthropological investigation ofBengkala began in 1993 with studies of thecultural adaptations that were made in responseto the high incidence of deafness in the village.Approximately 2-2% of the inhabitants of thisBalinese village are profoundly deaf.We describe all ofthe families with congenital

hereditary deafness in Bengkala, examine pos-

sible clinical features that might have beenassociated with the deafness, analyse the med-ical and village histories to determine the modeof inheritance, and describe a method of es-

timating the frequency ofthe deafness mutationin Bengkala given the high level of assortativemating among the deaf in this village. Ouranalyses presented here indicate that deafnessin Bengkala is congenital, autosomal recessive,and non-syndromal.

Materials and methodsSTUDY POPULATIONIn Bengkala, Bali, as with other Indonesianvillages, officials maintain records of births,deaths, marriages, child development, the in-cidence of obvious common diseases, oc-

cupations, land ownership, and the number offruit bearing trees. The records for Bengkala,as ofJune 1993, indicated that there were 2185inhabitants. Historically, residents of Bengkalahave exhibited a high incidence of goitre as is

the case for many other villages in this part ofBali.1-3 Goitre has been the concern of In-donesian physicians for many years and hasbeen the subject of extensive regional iodidetherapy since 1980.2 In addition to goitre, 47of the residents of Bengkala are congenitallydeaf. Fig 1 shows the families of all of the deafpeople in Bengkala. It is suspected that thehistory of deafness in Bengkala spans manygenerations since virtually the entire populationof Bengkala uses a village specific sign languageas a means of integrating the deaf and hearinginto a community.

INFORMED CONSENTSubjects from Bengkala participated voluntarilyin this research project and were either hearingor deaf. Permission to conduct this study wasobtained from six Indonesian sources: (1) theRector of Udayana University, Denpasar, Bali;(2) the Governor of Bali (Gubernur KepalaDaerah Tingkatl); (3) the regional director(Bupatu) of the district containing Bengkala;(4) the subregional director (Camat); (5) theelected village head (Kepala desa) of Bengkala;and (6) each person participating in the project.The goals of the project and the possible risksof participating in the project were presentedto each participant. This often required the useof a signing interpreter.

FAMILY HISTORIES AND ASSESSMENT OF

DEAFNESSFamily histories and pedigrees were collectedover a three year period and were used as aguide to interviews needed to obtain multipleindependent confirmations of the hearing sta-tus, the time of onset of deafness, and theprogression of hearing impairment for eachdeaf person. The people of Bengkala know thattwo hearing parents can produce children whoare deaf at birth. Consequently, they have de-vised methods to test the hearing of youngchildren. The first technique uses a loud noiseto induce a startle response. Various objectsare struck to produce a loud noise, includinga gong from the Bengkala gamelan (orchestra).If a baby shows a startle response, it is assumedto be hearing while an absence of such a re-sponse indicates that the child is deaf. A secondtechnique used by two deaf parents is to placetheir child with a hearing family for a shorttime in order to determine whether the child

336

on 10 May 2019 by guest. P

rotected by copyright.http://jm

g.bmj.com

/J M

ed Genet: first published as 10.1136/jm

g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/

Congenital non-syndromal autosomal recessive deafness in Bengkala, an isolated Balinese village

141 140 138 139 136 134

KuKII21 243 242 4

233 232 252 246 44 45 70 87

IV I151 105 235 234 247 249 248 109 108 251 250

V C

239 237 238 46 65

144 143 26 142

8 43 148 28 147 25 24 23 14 23 23 181 107 145

166 165 164 163 162 161 83 151

196 211 212

K4,

29 106 34 301@1III

226 33 31 227 86 98

IV32 255 254

214 213

o 0215 217 2188 21 216

22 1 222219 75 220

2

225 223 224 256

K5227 226

53 230 2281 53

III6 2560 231

Figure 1 Six kindreds (Kl, K2, K3, K4, KS, and Kll) containing all of the deafpeople found in Bengkala. Filled symbols indicate congenitally deafsubjects. Those who became deaf after birth are designated with a C to the left of their symbol.

does or does not respond to the hunIf the child is determined to be deabegin teaching the child the villageguage.

-10 7B

0

10;la

-U04) A *

i a20

30 H

lo 1000 2000 4000 8000Audio frequency (Hz)

40J_

100500 1000 2000 4

Audio frequenc

nan voice. In order to determine the clinical nature ofIf, parents the deafness observed in Bengkala, audiologicalsign lan- examinations were conducted initially in 1990

using a vibrating F 512 Hz tuning fork placednext to the ear and then on the mastoid bonefor a quick qualitative survey of the presumedhearing and deaf subjects in Bengkala. Pre-sumed normal hearing, obligate hearingDFNB3 heterozygotes, and deaf people werethen given audiological examinations in a quietroom in Bengkala with a type D audiometer

[lb from Minaty Medical Science Co Ltd, Osaka,i Japan. Audiograms were obtained for 24 deaf

: people (11 males and 13 females) and 14 hear->

0 ing people (eight males and six females) knownto be a parent of a deaf child (figs 1 and 2).The maximum sound pressure level used toestablish non-responsiveness of the deafpeoplewas 90 dB at 500, 1000, 2000, 4000, and

L0 8000 Hz. The Romberg and tandem walk tests400O 8000 of 44 deaf people were performed to evaluate,y (Hz) vestibular disturbances.

Figure 2 Audiograms for six obligate heterozygotes for the DFNB3 mutation. These areparents of deafpeople in fig 1. (A) Obligate heterozygotes with normal hearing (K4-33A, K6-56 [l, and K3-92 0). (B) Obligate heterozygotes with borderline normal hearing(Kl-36 A, K(2-83 L, and Kl-37 0). Only subject 56 ofK6 is not represented in fig 1

since this kindred moved from Bengkala. Open symbols indicate left ears; filled symbolsindicate right ears. The deafpersons from Bengkala (fig 1) were unresponsive when testedat the maximum sound pressure level of 90 dB at the frequencies indicated.

POPULATION SAMPLINGIn the summer of 1992, blood samples from21 hearing and deaf villagers from Bengkalawere obtained and DNA was extracted. These

Ki

II

Ill

lV

v

K2,

1I

Ill

Iv

v

K3

II

III

Iv

A-10

0

10K 0

Eni 5

a

COCO)00)m

e)x

20 _-

30 H

40j_

100501

337



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/

Winata, Arhya, Moeljopawiro, Hinnant, Liang, Friedman, Asher

21 samples were typed for four linked di-nucleotide repeat marker loci (STRs),D19S179 (Mfdl76), D19S253 (Mfd239),D19S47 (Mfd9), and D19S200 (MS265),known to be highly informative in populationsof western European ancestry.4 From an ana-lysis of these four STR markers in the 21hearing and deaf people, the inhabitants ofBengkala appeared to have levels of geneticvariability similar to the levels found in westernpopulations indicating that STRs would beuseful in determining the map position ofDFNB3.5

In the summer of 1993, 109 additional vil-lagers were interviewed, gave their consent toparticipate, were photographed, and sub-sequently gave blood samples for DNA ex-traction. These villagers included 51 hearingmen and women who professed to have noknown first or second degree deaf relatives. Ofthese 51 volunteers, DNA samples from 48Representative Bengkala Individuals (RBIs)were used to determine marker allele fre-quencies in the village of Bengkala. The gene-alogies of the RBIs were also investigated inthe summer of 1993. Each RBI was asked tolook at the photographs taken when bloodsamples were drawn to determine if any werekinsmen or affines of each other or of the deafpeople in our sample. Queries were also madeto determine if any additional deaf people ex-isted in the village that were unknown to us.This process uncovered subject 177, K2 (fig1), a very elderly deaf woman, who was ill atthe time and shielded from visitors. Subject177 had eight hearing sibs, two of whom (179and 144) were carriers of the recessive deafnessmutation. Deaf villagers were also interviewedwith the aid of a hearing interpreter who knewthe Bengkala sign language, generally in a set-ting where hearing members of the deafpeople's families were present. They were alsoasked to look at the pictures of RBIs to identifypossible kinsmen. These queries uncovered afew links between members of small familiescontaining deaf people allowing us to group allthe deaf in Bengkala into the six kindreds (Kl,K2, K3, K4, K5, and Kll) as shown in fig 1.To ensure the correct identity ofeach person,

Polaroid photographs were made and attachedto an identity card that accompanied the personas an Indonesian phlebotomist, in the presenceof one of the authors, withdrew two 10 mlsodium heparinised blood samples. Two ad-ditional physicians from the local communityclinic (Puskusmas) near Bengkala were alsopresent at the time blood samples were drawnand examined each participant as well as anyother person in Bengkala wishing medical ad-vice. DNA was extracted using the Super Quik-Gene DNA isolation kit (Analytical GeneticTesting Center Inc, Denver, Colorado). DNAsamples were used initially to check paternityof one critically important person (No 69, K3,fig 1), the only hearing person presumablywith two deaf parents, and then to map thedeafness gene segregating in Bengkala(DFNB3).6

ANTHROPOLOGICAL INVESTIGATIONSBetween 1990 and 1993, 13 pedigrees of thedeafand their relatives had been identified. Theinitial anthropological investigations conductedin the summer of 1993 were aimed at collectingadditional information on kinship and marriagein Bengkala and obtaining a better un-derstanding of the history of Bengkala.The process of learning the social or-

ganisation and ritual life of the community wasthen begun through interviews and observation.Historical studies were initiated to uncoverrecords pertaining to the dates when Bengkalamight have experienced population bottlenecksand when congenital deafness might first havebecome apparent in Bengkala. The historicalstudy of Bengkala involved collecting personalhistories and traditional village origin stories aswell as examining the village archives held bythe Gedong Kertia Library in Singaraja. Prasasti,metal plates written in Sanskrit in the 12thcentury, were examined and found to describethe village charter granted to Bengkala by theKing of Bali.

ANALYSIS OF HISTORICAL RECOMBINANTSUsing 150 STRs located an average of 30 cMapart, DFNB3 was found to be linked to thepericentromeric region of chromosome 17.6A conventional linkage analysis involving 250CEPH family members6 was then used to con-struct a linkage map of 13 chromosome 17STRs that flank DFNB3. Using historical re-combinant haplotypes of these 13 DFNB3flanking marker loci on 17p among 13 deafpeople from Bengkala, the known re-combination fractions from the CEPH familiesand equation 1 below, the number of gen-erations since the appearance of the DFNB3gene in Bengkala was estimated to be:

g= - [ln(I -HR)]/O (1)

where 0 is the known frequency of re-combination between two linked markers, HRis the proportion of historical recombinanthaplotypes observed for these markers, and gis the number of generations assumed to havepassed since DFNB3 appeared in Bengkala.78This analysis assumes that the DFNB3 muta-tion first appeared in Bengkala in completeallelic disequilibrium with closely linked mark-ers. Over time, as a consequence of re-combination and possibly mutation of linkedmarker alleles, the DFNB3 mutation has be-come associated with different combinations ofclosely linked marker alleles, that is, differentrecombinant haplotypes.The number of generations since DFNB3

appeared in Bengkala was also estimated usinga different method of calculating the proportionof disequilibrium that is expected to remainsince the mutation entered the population givenby equation 2:

ln[((rln) -_p)/(p (2)

g= ln(l1-0)

338



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/


where r is the observed number ofhomozygousrecessive deaf people who are also homozygousfor the i"' linked marker allele following g gen-erations of linkage disequilibrium decay, pi isthe frequency of the linked i"' marker allele,p2n is the expected number of deaf also ho-mozygous for the i' marker allele providedHardy-Weinberg and allelic equilibrium hasbeen reached, and n is the number of ho-mozygous recessive deaf people examined.

ResultsBIOLOGICAL NATURE OF DEAFNESSBengkala has a long history of reported goitre.'Since 1980, the people in Bengkala haveparticipated in an extensive iodide sup-plementation programme to eliminate goitrefrom Bali. Deafness in Bengkala was initiallysuspected to be caused by Pendred syndrome(MIM 274600)1 which is inherited as anautosomal recessive trait in which subjects aredeaf with goitre. X2 analysis indicates that sex,goitre, and hearing status are randomly as-sociated among members of the kindreds pre-sented in fig 1 (X2=2*87, 3df, p=0412).Regardless of hearing status, sex and goitre arerandomly associated among members of thekindreds shown in fig 1 and the RBIs (x2 =4 30, 3 df, p=0-231). These observations ruleout the possibility that deafness in Bengkala iscaused by Pendred syndrome. However, thereis presently an unexplained association betweensex and goitre in Bengkala with women ex-hibiting a higher incidence of goitre than men(x2 = 10-95, 1 df, p=0000941).There are over 35 additional syndromes

which have deafness as one of the reportedfindings.' -'4 For example, subjects with syn-dromes exhibiting hearing impairments mayalso show facial dysmorphic features, for ex-ample, Waardenburg syndrome with dystopiacanthorum'5 and Noonan syndrome with hy-pertelorism.12 Facial measurements of innercanthal distance, inner pupillary distance, andouter canthal distance of the RBIs do not differsignificantly from measurements of deaf sub-jects (data not shown). Thus, deaf people donot exhibit facial dysmorphic features as-sociated with Waardenburg syndrome type 1.To the best of our knowledge, all deafmembersof Bengkala are without any additional stig-mata. Deafness in Bengkala appears to be non-syndromal.

HEARING TESTSAudiological examinations indicated that deaf-ness in Bengkala is bilateral, sensorineural, andthat 24 tested deaf people were uniformly non-responsive to pure tone audiological ex-aminations at sound pressure levels up to90 dB. Histories of each deaf person in Beng-kala indicate that hearing loss was congenitalin all but two profoundly deaf subjects (Nos46 and 48, designated C in fig 1). These lattertwo subjects were excluded from the se-gregation and population analyses given belowand the mapping analyses.6 Physical ex-aminations of 44 deaf people failed to show

any obvious neurological anomalies includingvestibular disturbances that would be indicatedby the tandem Romberg test.

Fourteen people known to have parented adeaf child and who are thus obligate het-erozygotes were examined and their rep-resentative pure tone audiograms appear infig 2. Seven of these obligate heterozygotesrepresented by fig 2A produced normal puretone audiograms while the remaining sevenof these heterozygotes represented by fig 2Bproduced borderline normal pure tone audio-grams. People were considered borderline nor-mal if they exhibited a 25 db deficit at aminimum of two frequencies from 500 Hz to8000 Hz. A more complete survey of the popu-lation of Bengkala is needed before we mayconclude that a high proportion of DFNB3heterozygotes have a borderline hearing im-pairment.

PEDIGREES AND THE SEGREGATION ANALYSISAll familial relationships shared by deaf peoplein Bengkala are indicated by the pedigreespresented in fig 1. In families studied to date (fig1), no indications of consanguineous marriageshave been discovered. Examination of the kin-dreds in fig 1 indicates that there are 21 deafmen and 26 deaf women living in Bengkala.This observation suggests that deafness inBengkala is not a sex linked or sex influencedtrait. In addition, 14 marriages involving twounrelated deaf parents produced 15 deaf andone hearing (No 69) offspring (fig 1). Thepresence of 15 deaf offspring from apparentlyunrelated parents who are both deaf stronglysupports the conclusion that deafness in Beng-kala is caused by a fully penetrant autosomalrecessive mutation at a single locus, DFNB3.The single exceptional subject (fig 1, No 69)was evaluated using 13 STR markers on chro-mosomes 2, 4, 7, 17, and 19. For six of these13 marker loci (D4S194 (Mfdl46), D7S559(Mfd265), D17S799 (AFM192yh2), GA-TA1OHO7, D17S261 (Mfd41), and D17S520(Mfdl5A) subject No 69 possesses alleles notpresent in the father of record (No 58). NewSTR alleles may arise with a mutation rate ashigh as 10-3.16 The probability that all six non-maternal alleles found in subject No 69 aroseby mutation is at most 10-18, a highly unlikelyevent. To date, children born to two deaf par-ents in Bengkala are always congenitally deaf.A total of 16 deaf people in Bengkala were

born to parents who are both hearing (fig 1).If deafness in Bengkala is caused by a fullypenetrant autosomal recessive mutation at asingle locus, the marriages producing these 16people must involve parents who are het-erozygous for a mutation at this locus. Suchmarriages should produce one deaf child forevery three hearing children. In Bengkala, weobserved 16 deaf to 18 hearing children, asignificant deviation from an expected men-delian 1:3 ratio (X2=8.824, 1 df, p=0 002984). Such a deviation, however, is con-sistent with complete truncate ascertainmentbias which is caused by an apparent deficiencyin the number of hearing children from het-

339



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/

Winata, Arhya, Moeljopawiro, Hinnant, Liang, F?iedman, Asher

Table 1 Marriages and the expected frequencies of deafchildren from these marriages assuming random marriagesamong hearing parents in Bengkala

F M Frequency of marriage Deaf progeny

p P+/+ +I+ x x 0 0

(2-P) (2-P)

2(1-P) P+/-+/+ x x 0 = 0

(2-p) (2 -P)P 2(1 - P)

+/+ +l- x x 0 = 0(2-p) (2 - P)

2(1 -P) 2(1 -P) (1- p)2x x 1/4 =

(2 - p) (2 -P) (2 - p)2

F= father; M = mother; + = normal allele and-= the mutatedDFNB3 allele; relative frequencies derived from equation 4.The mutation (-) segregating in these families is assumed tobe fully penetrant.

erozygous parents who, by chance, produce nodeaf children and are, therefore, not as-certained. Correcting for ascertainment bias,we expected 19-6 (SD 2-0) deaf offspring."7Since the expected (19-6) and observed (16)deaf progeny are not significantly different atp=0 05, we cannot reject the underlying hy-pothesis that deafness in Bengkala is caused bya fully penetrant autosomal recessive mutationat a single locus. As this mutation causes non-syndromal autosomal recessive deafness, it hasbeen designated by Dr P McAlpine as DFNB3in accord with theHUGO Nomenclature Com-mittee.6

POPULATION GENETICSBy a conventional Hardy-Weinberg analysis,the frequency of the mutant DFNB3 allelein Bengkala is 0-147 (14-7% HW) and thefrequency of hearing adults heterozygous forthe DFNB3 mutation is 0-252 (25-2% HW).These calculations assume random marriagesamong all members of the village. As the fre-

quency of the deaf in Bengkala is 2 2%, ifmarriages are random with respect to hearingstatus, we would expect to find no more

than one marriage between deaf partners

(0.0222 x 2185/2 = 0 529). At present, there are

14 couples in Bengkala in which both membersare deaf. Hence, marriages in this communityare not random with respect to the hearingstatus of a person. Deaf people in Bengkalatend to marry other deaf people and have alldeaf progeny (fig 1). Therefore, a standardHardy-Weinberg analysis provides an over-

estimate of the frequency of the DFNB3 allelein Bengkala.An alternate method to estimate the fre-

quency of the DFNB3 allele among hearingsubjects can be obtained by considering onlyhearing adults and their progeny. For the sakeof this analysis, we assume that there is randommarriage among hearing people in Bengkala.In this case:

P2( + / + hearing)+2P(1-P)(+//-hearing)=P(2-P) (3)

where (1 - P) indicates the frequency of theDFNB3 mutation and +/- indicates subjectscarrying the DFNB3 mutation. The normalisedproportions of the above frequencies are given

by:

P/(2-P)(+/+) +2(1 -P)/(2-P)(+/-) = 1(4)

Table 1 shows the types of marriages, theirexpected frequencies, and the expected popu-

lation frequency of deaf progeny from thesemarriages occurring in Bengkala derived fromequation 4. Using the total frequency of deafpeople born to parents who are both hearing[1 6/(2,185-47 + 16) = 0*0074], the frequencyof the DFNB3 mutation among hearing peopleinBengkala, 1 -P = 0094 or9 4% (14-7% HW)while the frequency of hearing adults who ac-

Table 2 The marker loci surrounding DFNB3, the number of deafpeople homozygous for a specific marker allele (r),the frequency of this allele (pi), the distance between markers in centimorgans (cM) associated with the recombinationfraction (0), the number of recombinants among 250 CEPH family members (CEPH Rec), historical recombinants(HR) observed among the 13 deaf, estimated number of generations (gi) since DFNB3 was introduced into Bengkalacalculated from equations 1 and 2, and the distance between markers in megabases ofDNA (Mb)

Marker Locus r pi cM 0 CEPH Rec HR g, g2 Mb

Mfdl44 D17S520 5 0-22700 0000 0 9 - - -

AFM225zcl D17S804 7 0-2155-4 0054 14 9 79 91 -

GATA1OHO7 6 0-2274-1 0-041 10 4 4-1 7-3 -

AFM192yh2 D17S799 8 0-28312-9 0-126 32 6 2-1 3-3 1-2

VAW409 D17S122 12 0 53300 0000 0 1 39-2 1136 0-1

Mfd4l D17S261 13 0-5113-2 0-032 8 0 - - 40

AFM234tal D17S805 13 0 4752-1 0-021 5 2 3-8 9-8 -

AFMO26vh7 D17S783 11 0-4231-7 0-017 4 1 2-3 12-2 -

GATA2CO7 10 0-7203-6 0-036 9 2 2-2 4-8 -

AFM179xgll D17S798 10 0-6143-5 0 035 9 4 4-8 10-1 -

GGAA7D 11 7 0-0764.5 0045 11 6 5-8 6-3 -

Mfdl5 D17S250 8 0-3082-6 0-026 6 3 4-7 17-6 -

AFM200zf4 D17S800 6 0-570

Allele frequencies (pi) estimated from 48 RBIs.

340



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/


tually carry the DFNB3 mutation is 0-172 or17-2% (25-1% HW). This second estimate ofthe frequency ofDFNB3 in Bengkala may alsobe an overestimate.

Bali contains seven paseks, title groups or

supravillage connections based on an ancientmigration, or wave of conquest, from one ofthe Hindu palace-temple complexes es-tablished following the Hindu conquest of Baliduring the 12th century. Many of the deaf inBengkala are members of pasek Gelgel. Thiswould suggest that marriages even among thehearing in Bengkala may not be random. How-ever, as the complete pattern of marriages inBengkala has yet to be determned and as a

direct method for identifying the presence ofthe DFNB3 mutation requires that DFNB3 becloned, it is not possible at this time to obtainan unbiased estimate of the frequency of theDFNB3 mutation in Bengkala.

WHEN WAS DFNB3 INTRODUCED INTO

BENGKALA?

DFNB3 has been mapped to the peri-centromeric region of chromosome 17 usinga new genome wide disequilibrium search, a

strategy described briefly elsewhere6 and indetail in a manuscript in preparation (Asher etat). The current map distances between mark-ers in the region to which DFNB3 maps are

presented in the column of table 2 labelledcM. The STRs D17S122 (VAW409) andD17S783 (AFMO26vh7) flank DFNB3 andare 5-3 cM apart.6 If double recombinants are

minimised, these two markers are separatedby three historical recombinations. Giventhe haplotypes present in the 13 deaf subjects(table 2), DFNB3 appears to be in completeallelic disequilibrium with respect to theSTRs D17S261 (Mfd4l) and D17S805(AFM234tal). As indicated by equation 1,'the decay of this disequilibrium is dependentupon the distance between markers as a re-

combination fraction (0) and the number ofgenerations (g) since DFNB3 appeared in thepopulation. The more distant the marker isfrom the DFNB3 mutation, the larger numberof historical recombinant (HR) haplotypes willbe expected and observed (table 2). The fre-quency of these recombinant haplotypes andequation 1 were used to estimate the numberofgenerations since DFNB3 appeared in Beng-kala. As an example of this calculation, con-

sider the markers D17S805 and D17S783 thatare 2- 1 cM apart. In this interval, we observedHR=2/26 historical recombinant haplotypes,0 = 0-021 recombination fraction in-dependently determined from CEPH families,and the calculated number of generations sinceDFNB3 appeared is g=-[ln(1-2/26)]!0-021 = 3-8. Considering all marker intervalsand the associated recombinant haplotypes,table 2 presents 10 estimates for the numberof generations since DFNB3 first appearedin Bengkala. In general, these values range

between 2-2 and 7-9 generations with a singleexception.From a conventional linkage analysis,

D17S122 and D17S261 exhibited no re-

combinants out of 250 CEPH family mem-bers.6 These two loci are approximately 100 kbapart618 which is roughly equivalent to 0 =0-001.1' Among the 13 deaf, there is a singlerecombinant haplotype for these two markers.Using this marker combination and equation1, DFNB3 may have been introduced intoBengkala approximately 39-2 generations ago.As a generation in Bengkala can be as short as15 years, the generation estimates from table2 indicate that DFNB3 may have appeared inBengkala as recently as 32 years ago or as longago as 566 years.The number ofgenerations since the DFNB3

mutation was introduced into Bengkala (g) wasalso estimated from equation 2 which uses thefrequency of the linked marker allele. For themarkers D17S805 and D17S783, g=9-8 gen-erations when r/n= 11/13, 0 = 0-021, and pi =0-423 (which is assumed not to have changedsince the introduction of the DFNB3 mutationinto Bengkala). The values of g estimated fromequation 2 are also presented in table 2. Thus,we have two sets of estimates of the numberofgenerations since DFNB3 appeared in Beng-kala. Including all 10 estimates, equation 1yields estimates of g= 7-7 (SD 11 -2) gen-erations while equation 2 yields estimates ofg= 19-4 (SD 33 3) generations. These two es-timates are not significantly different becauseof the large variation found in each estimateand suggest that on average DFNB3 may haveappeared in Bengkala between 10 and 20 gen-erations or 150 to 300 years ago.

DiscussionHuman childhood deafness occurs worldwidewith an incidence of 1:650 to 1:2000 livebirths.202' At least 50% of congenital deafnessis thought to be hereditary.22 Seventy-five per-cent of hereditary congenital deafness is theresult of autosomal recessive mutations.22325Many people with autosomal recessive deafnessshow no other obvious associated features andso would be classified as having non-syndromaldeafness. The total number ofunlinked mutantloci producing non-syndromal autosomal re-cessive deafness has been estimated to be be-tween 20 and 2000, based on surveys ofinterethnic marriages in which the majority ofmarriages involving two unrelated deaf parentsproduce all hearing offspring.2' In contrast, twodeaf parents in Bengkala always produce deafchildren (fig 1).

Recently, Friedman et al6 mapped theDFNB3 mutation segregating in Bengkalausing a new strategy, allele frequency de-pendent homozygosity mapping (AHM).9AHM relies on the existence of a dense geneticmap of highly polymorphic genetic markers,assumes that the population examined is poly-morphic for these markers, and uses the factthat all mutations arise in allelic disequilibriumwith linked markers. To apply AHM, a reliabledetermination of highly polymorphic markerallele frequencies must be made.9 This requiredthe availability of unaffected, unrelated, Rep-resentative Bengkala Individuals (RBIs). In

341



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/

3Winata, Arhya, Moeliopawiro, Hinnant, Liang, Friedman, Asher

the case of DFNB3, an examination of 13deaf people and 48 RBIs provided sufficientinformation to map DFNB3 to the peri-centromeric region of chromosome 17.69

Historical recombinant haplotypes (table 2)used to narrow the chromosomal location ofDFNB3 can also be used along with linkagedata from CEPH families6 and equation 1 and2 to estimate the number of generations sinceDFNB3 was introduced into Bengkala. Ourestimates using equation 1 range from 2- 1 gen-erations to a maximum of 39-2 generations agowhile the estimates from equation 2 range from3-2 to 113 6 generations ago. For both setsof estimates, there is a single major outlierinvolving the same genetic region, Dl 7S261 toD17S122, which shows no meiotic re-combinants among 250 CEPH family mem-bers. In the absence of this outlier, the averageestimate using equations 1 and 2 respectivelyare 4-2 (SD 1-9) and 8-9 (SD 4-3) generationsor 63-0 (SD 28-5) and 133-5 (SD 64 5) yearsfor a 15 year generation.

In order to obtain an independent estimateof the date when DFNB3 first appeared inBengkala, investigation of deafness in otherBalinese villages were begun. Villages within 2kilometres ofBengkala, Tamblang (4 deaf/4127inhabitants), Bila (7/2097), and Kubutambahan(5/8947) all have low frequencies of deafnesscompared to Bengkala (49/2185). On the otherhand, Anturan, a small village approximately30 km from Bengkala has nine subjects withnon-syndromal deafness (approximately 1% ofthe population). There appears to be an his-torical relationship between deafness in Beng-kala and in Anturan. Many of the deaf inBengkala and all of the deaf in Anturan aremembers of pasek Gelgel. The origin temple(kawitan) serves as the ultimate reference pointfor all members of a pasek, who periodicallyreturn to the temple for ceremonies. Membersof each pasek remember the moves their an-cestors made as they spread out across Balifrom a kawitan, and periodically perform ce-remonies at the temples of the various villageswhere their ancestors once lived. Because ofmigration of the various pasek during the con-quest of Bali, it is possible that the same muta-tion for congenital recessive deafness issegregating in Bengkala and Anturan and thatthe high frequency of the gene in these twovillages is a consequence of a reproductivebottleneck or founder effect that occurredsometime after the 12th century when theHindu conquest and migrations began but be-fore the time members ofpasek Gelgel moved toAnturan and Bengkala. Our genetic estimatessuggest that this bottleneck could have oc-

curred within the last 200 years, nearly 600years after the Hindu conquest and migration.The two procedures for estimating the time

since DFNB3 first appeared in Bengkala giveaverages of 63 and 134 years. If correct, thepeople of Bengkala recognised the existence ofa genetic problem, developed social customsto identify newborn deaf children, developed a

sign language that is taught to deaf and hearingchildren alike, and integrated the deaf intothe community in slightly less than 150 years.

These ages appear consistent with the pedigreesin fig 1 that indicate the presence of deafpeopleand a relatively high frequency of the DFNB3mutation at least four generations (60 years)ago.The portion of chromosome 17 presented in

table 2 has 12 linkage intervals with an averagegenetic distance from CEPH meiotic re-combinations of 3 6 cM and an average of 3 9historical recombinants per interval. In the in-terval between D17S261 and D17S805, theinterval suspected to contain DFNB3, 250CEPH family members exhibited eight meioticrecombinations for a value of 0= 0-032(3 2 cM). This region, however, did not exhibita single historical recombinant among the 13deaf people. A contingency x2 analysis com-paring meiotic and historical recombinationsindicates a significant difference between thesetwo sources of recombination (X2 = 32 32, 11 dfat p=0 00066). The major contributions tothis x2 come from the region between D 17S520and D17S840 which is over 20 cM away fromthe interval suspected to contain DFNB3.Eliminating this region from consideration, thecontingency x2 is not significant (X2= 11 45,10 df, at p = 0-32334 1). This suggests that mei-otic recombinants and historical recombinants,with the exception of a single map interval, areconsistent with each other, lending support tothe results obtained from our analysis of thehistorical recombinants.Of the mutations causing deafness, those

that cause non-syndromal autosomal recessivedeafness are probably the most challenging tomap because of the large number of genescausing this type of deafness and the difficultyof finding large consanguineous families.The first two, NSRD1 and NSRD2 (now des-ignated DFNB1 and DFNB2, respectively, bythe HUGO Nomenclature Committee), weremapped to chromosomes 13 and 11 using con-ventional linkage analyses of extended, con-sanguineous pedigrees.Y27 The DFNB 1mutation has variable expressivity and theDFNB2 mutation has a variable age of onsetand possibly incomplete penetrance. From datapresented here, the recessive deafness mutationsegregating in Bengkala differs phenotypicallyfrom DFNB1 and DFNB2 in that DFNB3 isa fully penetrant recessive mutation that causesprofound congenital deafness and maps tochromosome 17p.6

In summary, Bengkala is a Balinese villagethat is at least 700 years old with approximately2185 inhabitants, 2-2% of whom are deaf. Acomplete survey of all of the villagers identifiedsix kindreds with 49 deaf people. Physical andaudiological examinations and family his-tories indicate that DFNB3 is congenital, sen-sorineural, and non-syndromal. A segregationanalysis presented here indicated that deafnessis caused by a fully penetrant autosomal re-

cessive mutation of the DFNB3 gene. A popu-lation analysis indicates that 17 2% of hearingpeople in Bengkala are heterozygous for themutated DFNB3 gene. Using mapping datapublished elsewhere,6 historical recombinantsbetween STRs closely linked to DFNB3 in-dicate that the DFNB3 mutation could have

342



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/


arisen or arrived in Bengkala by a stochasticevent within the last 200 years.

We are grateful for the help and hospitality of the communityofBengkala. We thank the late Dr Soedarminto, Project Directorand Chairman, Department of Biochemistry, Faculty of Medi-cine, Udayana University, for his support and guidance of thisproject. We also thank Dr Joedoro Soedarsono, Director, IUCfor Biotechnology, Gadjah Mada University, Yogyakarta, forsupporting this research project in its infancy and the use ofthe facilities at the IUC for Biotechnology to perform the initialDNA extractions and PCR amplifications. We thankM Swinda,phlebotomist, and Dr Sudewa Djelantik, Director, Red CrossBlood Bank, Sangla Bali Hospital, for the use of his laboratoryto extract genomic DNA. We thank Dr W Sudana, Departmentof Otolaryngology, Dr Dwi Sutanegara, Department of IntemalMedicine, Dr W Kondra, Department of Neurology, and DrDarmasutapa, Department of Psychology, Udayana University,for their evaluations of hearing and deaf people from Bengkala.We thank the previous and present Kepala Desa of Bengkala,and Dr Wayan Artama, IUC for Biotechnology, Gadjah MadaUniversity, Yogyakarta for their assistance in conducting thisresearch. YL was supported jointly by the Michigan StateUniversity Graduate Program in Genetics and from a grant toJHA and TBF from NIDCD (RO1 DC01160-02). We ac-knowledge the Deafness Research Foundation for support ofthis project. Personal funds of SW, TBF, JHA, JTH, and INAsupported the majority of this project.

1 Noosten HH. Krop op Bali: Overgedrukt uit het genees-kundig tijdschrift voor Nederlandsch. Indie Apt 1934;16:1421-45.

2 Soedarminto, Sutanegara D, Arhya W, et al. Research on

endemic goiter in Bali: a cooperative study between the RegionalDevelopment Unit in Bali and Udayana University. Uni-versity of Udayana, 1991: 16 (in Indonesian).

3 Report from a jointWHO/UNICEF/ICCIDD Consultation.Indications for assessing iodine deficiency disorders andtheir control programs. WHOINUT 1993;1: 14.

4 Weber JL, Wang Z, Hansen K, et al. Evidence for humanmeiotic recombination interference obtained through con-

struction of a short tandem repeat polymorphism linkagemap of chromosome 19. Am Hum Genet 1993;53:1079-95.

5 Morell R, Liang Y, Asher JH Jr, et al. Analysis of shorttandem repeat (STR) allele frequency distributions in aBalinese population. Hum Mol Genet 1995;4:85-91.

6 Friedman, TB, Liang Y, Weber JL, et al. A gene responsiblefor congenital, recessive deafness DFNB 3 maps to thepericentromeric region of chromosome 17. Nature Genet1995;9:86-91.

7 Hastbacka J, de la Chapelle A, Kaitila I, et al. Linkagedisequilibrium mapping in isolated founder populations:diastrophic dysplasia in Finland. Nature Genet 1992;2:204-11.

8 Luria SE, Delbruck M. Mutations of bacteria from virussensitivity to virus resistance. Genetics 1943;28:491-511.

9 Asher JH Jr, Liang Y, Winata S, et al. Allele frequencydependent homozygosity mapping: an alternative strategyfor mapping diseases caused by fully penetrant autosomalrecessive mutations. In preparation.

10 McKusick VA. Mendelian inheritance in man. 10th ed. Bal-timore: Johns Hopkins University Press, 1992.

11 Warkany J. Congenital malformations: notes and comments.Chicago: Year Book Medical Publishers, 1971.

12 Jones LK. Smith's recognizable patterns ofhuman malformation.Philadelphia: W B Saunders, 1988:709.

13 Buyse ML, ed. Birth defects encyclopedia. Dover, Mass: Cen-ter for Birth Defects Information Services Inc, 1990.

14 Stevenson RE, Hall JG, Goodman RM, eds. Human mal-formations and related anomalies. Vol 1 and 2. New York:Oxford University Press, 1993.

15 Asher JH Jr, Morell R, Friedman TB. Waardenburg syn-drome (WS): the analysis of a single family with a WS1mutation showing linkage to RFLP markers on humanchromosome 2q. Am Jf Hum Genet 1991;48:43-52.

16 Weber JL, Wong C. Mutation of human short tandemrepeats. Hum Mol Genet 1993;2:1123-8.

17 Emery AEH. Methodology in medical genetics. In: Anintroduction to statistical methods. London: Churchill Liv-ingston, 1986:37-54.

18 Chevillard C, LePaslier D, Passage E, et al. Relationshipbetween Charcot-Marie-Tooth IA and Smith-Magenisregions. snU3 may be a candidate gene for the Smith-Magenis syndrome. Hum Mol Genet 1993;2:1235-43.

19 NIHICEPH Collaborative Mapping Group. A Com-prehensive genetic linkage map of the human genome.Science 1992;258:67-87.

20 Fraser GR. Studies on isolates. Jf GenetHum 1964;13:32-46.21 Brownstein Z, Friedlander Y, Peritz E, Cohen T. Estimated

number of loci for autosomal recessive severe nerve deaf-ness within the Israeli Jewish population, with implicationsfor genetic counseling. Am JMed Genet 1991;41:306-12.

22 Steele MW. Genetics of congenital deafness. Pediatr ClinNorth Am 1981;28:973-80.

23 Nance WE, McConnell FE. Status and prospects ofresearchin hereditary deafness. In: Harris H, Hirschhorn K, eds.Advances in human genetics. New York: Academic Press1973:175-250.

24 Nance WE, Sweeney A. Genetic factors in deafness of earlylife. Otolaryngol Clin North Am 1975;8:19-48.

25 Marres HAM, Cremers WRJ. Autosomal recessive non-syndromal profound childhood deafness in a large pedi-gree. Arch Otolaryngol Head Neck Surg 1989;115:591-5.

26 Guilford P, Ben Arab S, Blanchard S, et al. A non-syndromicform of neurosensory, recessive deafness maps to thepericentromeric region of chromosome 13q. Nature Genet1994;6:24-8.

27 Guilford P, Ayadi H, Blanchard S, et al. A human generesponsible for neurosensory, nonsyndromic recessivedeafness is a candidate homolog of the mouse sh-1 gene.Hum Mol Genet 1994;3:989-93.

343



g.bmj.com

/J M


g.32.5.336 on 1 May 1995. D

ownloaded from

http://jmg.bmj.com/

Documents

Genet Congenital non-syndromal autosomal recessive ... · 3MedGenet 1995;32:336-343 Congenital non-syndromal autosomal recessive deafness in Bengkala, an isolated Balinese village