sign in sign up
<<
Home , Hyperpigmentation, Hypopigmentation, Melanin, Ocular albinism, Piebaldism, Waardenburg syndrome

6565 JMed Genet 1997;34:656-665 Syndrome of the month J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from

Waardenburg syndrome

Andrew P Read, Valerie E Newton

Abstract coincidences were multiplying, Waarden- Auditory-pigmentary syndromes are burg was prompted to undertake a system- caused by physical absence of melano- atic search among 1050 inmates of five cytes from the , , eyes, or the stria Dutch institutions for the deaf. vascularis of the . Dominantly Waardenburg's results, published in a huge inherited examples with patchy depig- paper2 in the American Journal of mentation are usually labelled Waarden- Genetics in 1951, defined the syndrome now burg syndrome (WS). Type I WS, named type I (WS1). characterised by dystopia canthorum, is He characterised the syndrome as autosomal caused by loss offunction mutations in the dominant with very high penetrance (159/161) PAX3 gene. Type III WS (Klein- of dystopia but reduced penetrance of all other Waardenburg syndrome, with abnormali- features. He appears to have taken little interest ties of the arms) is an extreme in the patients who did not have dystopia can- presentation of type I; some but not all thorum (he noted seven people with hetero- patients are homozygotes. Type IV WS chromia or isochromic hypoplastic irides, but (Shah-Waardenburg syndrome with Hir- without dystopia, whom he did not investigate schsprung ) can be caused by further). It was not until 1971 that Arias' drew mutations in the genes for -3 or attention to the existence of a separate division one of its receptors, EDNRB. Type II WS of the syndrome, which he named type II is a heterogeneous group, about 15% of Waardenburg syndrome (WS2). WS2 has whom are heterozygous for mutations in identical auditory and pigmentary features to http://jmg.bmj.com/ the MITF (microphthalmia associated WS1 but lacks dystopia canthorum. Two of ) gene. All these forms Waardenburg's original families had this vari- show marked variability even within fami- ant, but both were so small that Waardenburg lies, and at present it is not possible to had overlooked the familial "non-penetrance" predict the severity, even when a mutation of dystopia. is detected. Characterising the genes is Klein's patient was very different from those helping to unravel important develop- in Waardenburg's families. As well as showing on September 30, 2021 by guest. Protected copyright. mental pathways in the and the usual features of Waardenburg syndrome its derivatives. type I to an exceptional degree, she also (YMed Genet 1997;34:656-665) suffered from a severe musculoskeletal syn- Keywords: Waardenburg syndrome; auditory- drome resembling amyoplasia. At first, pigmentary syndromes Waardenburg speculated that perhaps she was homozygous for the gene mutated in his History syndrome. Later he evidently came to feel that On 14 December 1947 the Dutch ophthal- Klein's patient represented some different mologist and geneticist Petrus J Waardenburg clinical entity, a view that Klein did not presented a deaf-mute man with "dystopia appreciate.4 Over the years a small number of Department of other have been described who show , St punctorum lacrimarum, and patients Mary's Hospital, partial " to a meeting of the Dutch similar features to Klein's patient, but in milder Hathersage Road, Ophthalmological Society.' The patient had form. WS with musculoskeletal abnormalities Manchester M13 OJH, eyes but was bald, and Waardenburg did has been called type III WS, Klein- UK not at the time make the connection between Waardenburg syndrome, or WS3. Already in A P Read , white forelock, unusual eye col- 1983 Klein' suggested that this was a variant of WS 1. Centre for Audiology, our, and dystopia canthorum. He mentioned presentation University of a report of twins with the same eye In 1981, Shah et ar described 12 babies with Manchester, abnormality who were "coincidentally" also Hirschsprung's disease, white forelocks, and Manchester M13 9PL, deaf-mute. The following year, while he was white eyelashes, born to five families in UK visiting Geneva, David Klein showed him a Bombay. What, if any, relation these babies had V E Newton 10 year old girl with a remarkably severe to Waardenburg syndrome is not clear. Dysto- was not because Correspondence to: auditory-pigmentary syndrome who also pia was absent, hearing tested Professor Read. had dystopia canthorum. Realising that all babies died in the neonatal period, and the Waardenburg syndrome 657

Table 1 Thefour types of Waardenburg syndrome

Type MIM Inheritance Distinguishing feature Comments J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from I 193500 AD Dystopia canthorum W>1.95 Nearly all have PAX3 mutations II 193510 AD No dystopia Heterogeneous; 15% have MITF mutations III (Klein-Waardenburg) 148820 AD (most cases Hypoplasia of limb muscles; Variant presentation of WS 1; mostly sporadic) of elbows, fingers PAX3 heterozygotes; some may be homozygotes IV (Shah-Waardenburg) 277580 Mostly AR Hirschsprung's disease Heterogeneous; includes homozygotes for EDN3 or EDNRB mutations

pigmentary disorder of the irides was reported marked interfamilial and intrafamilial variabil- as "isochromia irides, light irides with ity. Table 3 shows the penetrance ofthe various mosaic pattern...a common inherited condi- features in a series of WS patients reported by tion in our population". However, the pre- Liu et al,6 from our own observations, and from sumed recessive combination of pigmentary published reports. The same pigmentation disturbance and Hirschsprung's disease has defects appear in WS1 and WS2, but with dif- been called Shah-Waardenburg syndrome, type ferent frequency.6 The higher reported inci- IV Waardenburg syndrome, or WS4. Recent dence of hearing loss in WS2 compared to demonstration of mutations in endothelin-3 WS1 is probably mainly a consequence of and its receptor (see below) have allowed a new diagnostic requirements: without dystopia can- and more concrete definition of WS4. thorum as a guide, patients need to show more auditory-pigmentary features to be classified as Clinical features ofWaardenburg affected. Fig 1 shows typical facial appearances ofWSl and WS2. syndrome types I and II Table 1 summarises the four clinical types of Waardenburg syndrome and formal diagnostic INCIDENCE AND PREVALENCE criteria are shown in table 2. Whereas WS1 Waardenburg' estimated the prevalence of his defines a specific genetic entity, the clinical syndrome to be 1/42 000 ofthe population and definition of WS2 is arbitrary, covering any 1.43% ofthe congenitally deaf. Fraser7 found a auditory-pigmentary syndrome that does not prevalence of 2.12-3.01/100 000 in his school clearly belong somewhere else. The label WS2 study of 2355 deaf children, and estimated a undoubtedly covers a heterogeneous collection prevalence of 1.44-2.05/100 000 in the general of defects. Apart from dystopia population. WS1 and WS2, as defined by the canthorum, all features ofWS 1 and WS2 show diagnostic criteria in table 2, are about equally common. In addition there are many families

Table 2 Diagnostic criteria for Waardenburg syndrome ascertained through a proband with hearing http://jmg.bmj.com/ types I and II loss, in which one or more relatives have single features of WS2, without anybody satisfying Diagnostic criteria for WS1 have been proposed by the Waardenburg Consortium.'6 In brief, to be counted as affected the full diagnostic criteria. Without molecular a person must have two major or one major plus two minor data it is impossible to say whether or not these criteria, from the following list: families should be added to the figures for Major criteria WS2. The mutation rate has been estimated as Congenital sensorineural hearing loss 0.4/100 000 gametes' or 0.39/100 000 on September 30, 2021 by guest. Protected copyright. Pigmentary disturbances of iris (a) complete heterochromia gametes.7 iridium: two eyes of different colour; (b) partial or segmental heterochromia: segments of blue or brown pigmentation in one or both eyes; or (c) hypoplastic blue eyes: characteristic DYSTOPIA CANTHORUM brilliant blue in both eyes Dystopia canthorum is the most penetrant fea- ture of WS1, being present in 99% of those Hair : white forelock affected.8 Dystopia presents with the appear- Dystopia canthorum: W> 1.95 averaged over affected family ance of blepharophimosis and with fusion of members (note that this was modified from the original the inner eyelids medially leading to a reduc- proposal of W>2.07 in the light of experience) tion in the medial sclerae (fig 1). The inferior Affected first degree relative lachrymal ducts are displaced laterally, with the punctae opposite the cornea. Not only the Minor criteria inner canthal, but also the interpupillary and Congenital leucoderma: several areas of hypopigmented skin outer canthal distances are greater than nor- mal, indicating a degree of . Synophyrys or medial eyebrow flare In the majority of affected subjects dystopia Broad and high nasal root is readily recognised, but our experience shows that clinical impression is not altogether Hypoplasia of alae nasi reliable. Subjects with WS2 risk being misdiag- Premature greying of hair: scalp hair predominantly white nosed as WS 1 if there is mild hypertelorism. It before age 30 is much better to rely on a biometric index.

Criteria for WS2 were suggested by Lui et al.6 These authors The W index (fig 2) is the best of several recommended that two major features should be present to broadly equivalent indices9 (the bizarre num- make the diagnosis of WS2. The major features are as in the bers in the formula come from a discriminant list above, except for the exclusion of dystopia canthorum and of premature greying analysis). Such indices are based upon the inner canthal, interpupillary, and outer canthal 658 Read, Newton

Table 3 Penetrance (%) of clinicalfeatures of type I and type II Waardenburg syndrome. 2.07, the threshold first adopted by the Data from Liu et alr (where references to published cases are given) Waardenburg Consortium, and was published as a type II family with a PAX3 mutation.' The Type Source No SNHL HetI HypE WF EG Skin HNR Eyb J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from data in fig 3 suggest that a threshold of 1.95, WSI Liu et al 60 58 15 15 48 38 36 100 63 Other reports 210 57 31 18 43 23 30 52 70 averaged across all affected family members, WS2 Liu et al 81 78 42 3 23 30 5 0 7 gives the best distinction, and reclassifies this Other reports 43 77 54 23 16 14 12 14 7 family as WS1. As the figure shows, the eye SNHL=sensorineural hearing loss, HetI= heterochromia irides, HypE=hypoplastic blue eyes, measurements in persons with WS1 and WS2 WF=white forelock, EG=early greying, Skin=white skin patches, HNR=high nasal root, can overlap, so the threshold cannot be taken as Eyb=medial eyebrow flare. absolute, but it has proven a very useful guide. Its practical value is as a guide to whether or not it is worth asking the laboratory to look for a PAX3 mutation (see below).

.-' I,J.. P-Alow,-1 OTHER FACIAL FEATURES A broad, high nasal root, medial hypertrichosis and synophyrys, and hypoplasia ofthe alae nasi AW' are features associated with dystopia cantho- rum in WS1 (table 3). Other facial features described include a patent metopic suture and _A* square jaw."l Strabismus may be more common with WS 1 than normally. 12

EYE COLOUR Iris heterochromia may be complete or partial. In complete heterochromia each iris is a differ- ent colour, while in partial heterochromia the differently coloured area of the iris is sharply demarcated from the remainder and is usually, but not invariably, a radial segment (fig 1C). Partial heterochromia may be unilateral or bilateral and, ifbilateral, may be symmetrical or asymmetrical. Partial heterochromia was found in 4.2% of subjects with WS1 and 27.5% of those with WS2 in our study.6 Hypoplastic blue irises are found where there is deficient iris stroma, and mainly in

association with a severe or profound hearing http://jmg.bmj.com/ loss.6 We found that hypoplastic blue irises were significantly more common in children with WS 1 than WS2. The fundus is reported to show pigmentary changes that correspond to those found on the retina.'2 13

HAIR COLOUR A distinctive white forelock is usually de- on September 30, 2021 by guest. Protected copyright. scribed, but the forelock may be or Figure 1 Facial appearance of Waardenburg syndrome. (A) Type I WS. Note dystopia .3 1' The site of the forelock is usually in canthorum. (B) Type I WS. Typicalfeatures include profound hearing loss, dystopia canthorum, pale blue eyes, white forelock, and white eyelashes on the left. This boy has a the midline but it may be elsewhere on the deletion of 7-8 Mb ofDNA including the entire PAX3 gene. Other clinicalfeatures (mental head. It may vary in size from a few to a retardation, growth retardation) are attributed to deletion of other genes. (C) Type II WS clump of hair and, if present at birth, may per- caused by a splice site mutation in the MITF gene. Note eye colour, with blue left eye and to brown right eye with a sharply demarcated radial blue segment; note normal build offace with sist or disappear only reappear later, usually no dystopia canthorum. She has mild . Other affected relatives with in the teens, when it is considered to represent the same mutation show white forelock, early greying, and varying degrees ofhearing loss. early greying.'5 Complete of the hair may occur in the teens and the hair distances. Indices based on these three meas- may be sparse and of poor quality.7 The urements are more reliable than those based premature greying signifying WS is defined by upon two measures alone, although if there is the Waardenburg Consortium as predomi- strabismus the interpupillary distance cannot nance of white hairs appearing before the age be used. The indices depend upon the of 30 years with the white hairs appearing in relationship between the measurements rather the midline.'6 Pigmentation defects can affect than absolute measures, and so should be the eyebrows and eyelashes as well as scalp unaffected by age, race, or sex. hair.7 Fig 3 shows the justification for placing such emphasis on dystopia. In this data set, every SKIN SIGNS family with mean W>1.95 (averaged over all Hypopigmentation ofthe skin is congenital and affected family members), but no family with may be found on the face, trunk, or limbs. It mean W< 1.95, shows evidence of PAX3 may be associated with an adjacent white fore- involvement from linkage or mutation analysis. lock. has also been One family had a W value between 1.95 and described" and this may develop after birth in Waardenburg syndrome 659

include low frequency and U shaped losses, bilateral or unilateral, and sometimes a combi-

nation of a low frequency sensorineural hearing J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from Al 'A loss in one ear and a profound loss in the other. Radiological investigation of the auditory system has indicated a normal temporal bone or of the lateral semicircular canal a--a associated with a normal bony cochlea.'9 The histological appearance of one temporal bone b._ studied by Fisch" was consistent with the auditory features of type I being caused by a cochleosaccular degeneration. Cochleosaccu- Figure 2 Diagnosis ofdystopia canthorum. To calculate the W index: (1) Measure a, b, c in mm using a rigid ruler lar degeneration and a complete absence of held against theface. (2) Calculate pigmentation was reported at necropsy in the X=(2a-0.2119c-3.909)1c. (3) Calculate inner ears of a 3 year old child with WS4. She Y=(2a-0.2479b-3.909)1b. (4) Calculate W=X+Y+alb. had WS type I is diagnosed if the average Wof all affected extensive skin depigmentation, bilateral family members is 1.95 or more. profound hearing loss and aganglionic mega- colcon, but not dystopia canthorum.74 a previously hypopigmented area. Arias3 noted that hypopigmented areas frequently had Other associated conditions hyperpigmented borders. If depigmented A long list of other conditions have been patches are extensive, (owing to reported in association with Waardenburg mutation in the KIT gene) should be sus- syndrome.20 It is difficult to disentangle to what pected, especially if the patient has normal extent these additional conditions are true rare hearing. manifestations of WS (type I or II), coinciden- tal findings, or features of other neurocristopa- HEARING thies that are really separate syndromes. WS1 The hearing loss in Waardenburg syndrome is carries a small but definite risk of spina sensorineural, congenital, and usually non- bifida20 21 and may occasionally be associated progressive. It may be unilateral or bilateral and with Sprengel's shoulder7 (unpublished obser- can vary in degree from slight to profound. vations) and with cleft lip/palate.22 Hageman and Delleman'7 reported that 25% of Hirschsprung disease (HSCR) is unques- subjects with WS1 and 50% ofthose with WS2 tionably associated with Waardenburg syn- had a bilateral sensorineural hearing loss. drome in WS4, which as described below is Newton'8 found that the penetrance of sen- recessive and caused by mutations in the sorineural hearing loss varied from 69% in EDN3 and EDNRB genes. Whether there is WS1 to 87% in WS2, after omitting the also an association with other forms of

probands (who had been ascertained through Waardenburg syndrome is less clear. Most http://jmg.bmj.com/ their hearing loss). Variation between families reports ofWS-HSCR patients date from before was high. the distinction was made between WS 1 and A profound bilateral loss is the commonest WS2. One might expect WS1, as a neuro- degree of hearing loss in both types, particu- cristopathy, to be the form associated with larly in type I. Unusual audiogram shapes HSCR, but in fact any association that may exist ofWS with HSCR is with type II WS. We 4

are aware ofonly a single case of HSCR among on September 30, 2021 by guest. Protected copyright. the hundreds of WS1 patients with proven PAX3 mutations (R Ramesar, personal com- munication). 3 Homozygous Waardenburg syndrome In view of the frequency of deaf-deaf mar- riages, WS homozygotes are likely to be conceived from time to time. Mouse models x a) (see below) suggest that WS 1 or WS2 homozy- 2 gotes are likely to be very much more severely 3._ affected than heterozygotes, and this is a point to be borne in mind in genetic counselling. Double heterozygosity (for example, WS1- WS2) is probably much less risky, although this view is not based on adequate data. Two cases of homozygosity for WS1 have been described. Zlotogora et a!23 described and 1 illustrated a child with severe hypopigmenta- tion, profound congenital hearing loss, and contractures ofthe forearms, born to a consan- guineous Israeli Arab family. Both parents have Figure 3 Dystopia canthorum: mean and range of W values among affected people in 48 typical WS1, and the child is homozygous for WSfamilies studied by the Waardenburg Consortium. The horizontal dotted lines show proposed thresholds of W=1. 95 or W=2.07for distinguishing WSI and WS2. Note that all the PAX3 mutation S84F. The physical picture families showing evidence ofPAX3 involvement have average W>1. 95, but not all have is strikingly reminiscent of Klein's original W>2. 07. Redrawn from Farrer et al, Am Hum Genet 1994;55:728-37 with permission. WS3 patient. Surprisingly, there is no neural 660 Read, Newton

tube defect, although in the Splotch mouse drome, some WS2 may be melanocyte specific, model homozygotes have lethal whereas WS 1 and the rare variants WS3 and defects. Ayme and Philip24 described a fetus WS4 are neurocristopathies, involving the that was the product of brother-sister incest in frontal bone, limb muscles, and enteric ganglia, J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from a French Gypsy family with typical WS 1. The respectively. All these extra tissues are neural was terminated because of anen- crest derivatives. cephaly diagnosed on scan; the fetus had major abnormalities very reminiscent of homozygous PAX3 and Waardenburg syndrome type I Sp mouse embryos, including severe contrac- Foy et at0 mapped WS 1 to the distal long arm tures and webbing of the limbs. No material of in 1990, using a clue was available from the fetus, but we have found provided by a Japanese patient with de novo a PAX3 mutation N269L in a heterozygous WS 1 and a chromosomal inversion member of the family (M Tassabehji, A P inv(2)(q35q37.3).31 The marker showing link- Read, unpublished data). This mutation is age was ALPP, the placental alkaline phos- typical of those found in WS 1. phatase. ALPP was said32 to map to 2q37, but Hulten et ar3 described a profoundly deaf this was based on radiolabelled in situ hybridi- and severely depigmented child born to first sation results which were not totally unambigu- cousin parents who both had white forelocks ous. ALPP is a difficult marker to use for in situ and white skin patches but normal hearing. hybridisation because distal 2q contains at least The child might have been a WS2 homozygote, three highly homologous alkaline phosphatase although homozygosity for piebaldism (KIT loci, ALPP, ALPI, and ALPPL2, which all mutation) is perhaps more likely. cross react. Physically, the WS 1 gene must be located at one of the inversion breakpoints in The neural crest in auditory-pigmentary the Japanese patient, that is, 2q35 or 2q37.3. syndromes Fluorescent in situ hybridisation has now The association ofhearing loss and pigmentary located the WS1 gene at 2q35.33 abnormalities has long been known in a variety On the basis of this map position, Foy et aP" of .26 In his Origin of Species, Charles suggested that WS1 might be homologous to Darwin asked "What can be more singular than the Splotch mouse mutant. This speculation the relation between blue eyes and deafness in proved correct when three groups identified cats?" In all these auditory-pigmentary syn- PAX3 (originally called HuP2) and its mouse dromes the underlying cause ofthe hearing loss homologue Pax-3 as the gene mutated in WS 1 is a still unexplained requirement for melano- and Sp.34 36 cytes in the stria vascularis of the cochlea.27 Our knowledge of PAX3 expression mostly There is no requirement for (albinos comes from studies of the mouse Pax-3 have normal hearing), but in the absence of gene,37 39 although preliminary studies in the stria is abnormally thin, no human embryos suggest a similar pattern.4"

endocochlear potential is generated, and later Summarising from the extensive studies of http://jmg.bmj.com/ in development Reissner's membrane collapses Gruss's group,3' no Pax-3 transcripts were leading to destruction of the organ of Corti.26 27 detected in any tissues of adult mouse, but Thus auditory-pigmentary syndromes are transcripts were present in embryos from day 8 caused by a physical absence of melanocytes to day 17, peaking at days 9 to 12 during neu- which may affect skin, hair, eyes, or the stria rulation. Transcripts were concentrated in vascularis. Usually the melanocyte deficiency is neuroepithelium, in the dorsal part of the neu- patchy, but alternatively a general dilution of ral groove and in the recently closed neural pigmentation may be seen. In man, hearing loss tube. Pax-3 is expressed longitudinally down on September 30, 2021 by guest. Protected copyright. with uniform dilution of pigmentation is the length of the neural tube from the usually described as Tietz-Smith syndrome hindbrain, but only in mitotically active cells of (MIM 103500) rather than Waardenburg syn- the alar and roof plates, dorsal to the sulcus drome, but in both man and mouse, different limitans. These cells are the source of the neu- alleles of the mi/MITF gene can cause either ral crest. Among neural crest derivatives, Pax-3 spotty or uniformly diluted depigmentation expression was seen in the spinal ganglia and (see below). some craniofacial cells (nasal process and some All melanocytes except those in the retina first and second branchial arch derivatives), but originate in the embryonic neural crest. not in melanocytes, chromaffin granule cells, Absence of melanocytes could be because of a the developing heart, or sympathetic ganglia. In failure of differentiation in the neural crest, a addition to neural tissue, segmented mesoderm failure of melanoblasts to migrate, or a failure also contained Pax-3 transcripts between 8.5 to terminally differentiate and survive in their and 11 days of gestation. Onset of Pax-3 final location. Countless genes must be in- expression coincided closely with the division volved in these processes, and so the genetics of of presegmented mesoderm into discrete auditory-pigmentary syndromes is likely to be somites, and preceded formation of the dermo- complex.28 29 A distinction might be made myotome and sclerotome; as the somites disso- between those syndromes where only melano- ciate, Pax-3 expression is switched off. Note, cytes are involved and those where there is a however, that postnatal appearance or disap- broader malfunction of the embryonic neural pearance of a white forelock and greying in the crest. A condition affecting both skin and reti- teens or twenties have repeatedly been docu- nal melanocytes is likely to be melanocyte spe- mented in WS1 families with defined PAX3 cific, since retinal melanocytes are not derived mutations, so not all effects of PAX3 are from the neural crest. In Waardenburg syn- confined to neurulating embryos. Another site Waardenburg syndrome 661

A Q254X 100% Effective level of PAX3

E251X 91 6del(1) J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from 434del(1 6) R223X Y305X Dystopia 364del(5) A196T 8774ins(1) x2 ,,WS1' 358del(1) R195X \ W2!74X 50% Melanocytes 288del(1) 556del(2) W2!66X Truncating 191 del(17) S mutations Limb buds "WS3" S r77777m-- rm;mm2o r / .-N ... 39 -I Hiw Lx2u lii_- 0% I Neural tube defects

"W266C rhgure S A hypothesis to explain PAX3 dosage effects. G99D Non-truncating The effective level ofPAX3 protein depends both on the W269L mutations amount offfunctional PAX3 protein and on variations in G81A the cellular systems that respond to PAX3 signalling. V78M rR270C Dystopia canthorum is always seen when PAX3 dosage is reduced, melanocyte defects are common in people with 50% "IR270L dosage, limb defects (WS3) are seen only in heterozygotes II185del(18) who have relatively inefficient PAX3 response systems, or in F45L " R271C x2 people homozygousfor loss offfunction PAX3 mutations. R271H tify the DNA targets to which the PAX3 protein binds. Optimal binding sites for PAX3 Paired box 1 Homeobox protein in vitro have been determined by repeated cycles of immunoprecipitation and PCR ofpanels ofrandom oligonucleotides. For the paired domain optimal binding was achieved with the oligonucleotide B CGTCACG(G/G)TT,4 and for the homeodo- main (of the Drosophila paired protein) CCTGAGTCTAATTGATTACTGTACAG.43 Splice R214X Truncating Within these longer sequences, consensus core uonor x2 Spi ice mutations binding sequences of GTTCC or GTT-AC acceptor R259X (paired domain) and ATTA (homeodomain) nt944dell have been recognised." 1 2 3 4 5 6 7 8 9 Exons It is worth reflecting, however, that master ~~~~~~~. switches in development probably depend on http://jmg.bmj.com/ low affinity interactions. Downstream of the master genes, the effector mechanisms need --- S29;38P high affinity binding to make them specific and R203K reliable. However, the master switch, the deci- N278D dei R217] sion point that sends only some of a population Norn-tru ncati ng Y253C m utati o ns of cells down a certain developmental pathway,

S250P needs a finely balanced affinity to work. As with on September 30, 2021 by guest. Protected copyright. haemoglobin and oxygen, binding that is too ::: .Basic donmain HLH domair strong is just as bad as binding that is too weak. zip Thus, the natural targets may be deliberately non-optimal. Most in vitro studies of PAX3 Figure 4 (A) PAX3 mutations in Waardenburg syndrome type I and III. DNA binding have used the e5 sequence from 434del(16) and 916del(1) mutations were found in patients with type III IWST all others were in type I. Note that non-truncating mutations are concentrated in the. 5'part of the the promoter of the Drosophila even-skipped paired box and in the third helix of the homeobox, the two regions criticalfoi?rprotein gene; however, it should be noted that DNA recognition. The A 196T mutation to is likely affect splicing because it Drosophila mutants deficient in or conserved G at the 3' end of exon 447 (unpublished data from our laborator3y)r other prd, ectopi- groups have reported similar data.'6"13 (B) MITF mutations. Thefamily swith the cally expressing prd, show no changes in del(R217) mutation (which is identical to the original mos use mutation) even-skipped regulation.45 To date, only one had Tietz-Smith rather than Waardenburg syndrome47; all otherfamilies hazd typical type natural target of PAX3 is known, the MET II WS. The R203K mutation may be a neutral change: it was seen in the piroband of a 42 four generation family with typical WS2, but did not track with WS throug)rh the oncogene. pedigree.47 Data from our laboratory47 (unpublished data), Nobokuni et al (R214X, Over 50 different PAX3 mutations have been R259X),4 and Morell et al (944de1l1)).65 described in patients with WS 1 or WS3. Fig 4A summarises mutations detected by our of expression was the undifferenti:lated mesen- group'0 35 46 47 (unpublished data). Others have chyme of the limb buds,4' exj plaining the reported similar findings.36 4853 The mutations phenotype of WS3. are almost always different in different affected PAX3 encodes a DNA binding transcription families; the only mutations seen in more than factor, one of a family ofnine human PAX pro- one family are 874ins(G) which inserts a teins defined by the presence of Ea 128 amino seventh G into a run of six Gs, and substitu- acid paired domain. The prototype is the Dro- tions ofR270 or R271 attributable to deamina- sophila paired (prd) gene. PAX3,,PAX6, and tion of CpG. Mutations fall into three classes. PAX7 additionally contain a homeo- (1) Deletions, fErameshifts, splice site muta- domain. An important research goal is to iden- tions, or nonsense mutations that are expected 662 Read, Newton

to act as null alleles. These include complete MITF and its mouse homologue mi encode deletion of the PAX3 gene. Mutations in this proteins belonging to the well known family of

class are scattered across exons 2-6 of the b-HLH-Zip (basic helix-loop-helix zip- J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from PAX3 gene, though in contrast to PAX6 (and per) transcription factors. These proteins for unknown reasons), they are rare in the 3' dimerise as - or heterodimers through part of the gene. their HLH-Zip regions and bind DNA through (2) substitutions in the 5' part of their basic regions. Mi/MITF is one of the few the paired box. These all affect amino acids loci at which more alleles and a richer molecu- known to make important DNA contacts in the lar have been found in mice than in prd-DNA complex.54 man. Mutations in the basic region generally (3) Amino acid substitutions in the third produce molecules that can cause dominant alpha helix of the homeodomain. This helix negative effects by sequestering wild type mol- (the recognition helix) is known to be critical ecules in dimers that cannot bind DNA for recognition by homeodomain proteins of correctly. Mice heterozygous for these muta- their DNA target.55 tions have white spotting or, in some cases, These findings suggest that the mutational dilution of the coat colour. Homozygotes for mechanism is loss of function and that the these, or for the recessive alleles with dimerisa- of WS1 depends on haploinsuffi- tion defects, are mainly or entirely white, and ciency. We favour the model in fig 5. Develop- some alleles produce microphthalmia, ment of the frontal bone must be uniquely sen- defects, , or dental defects. Com- sitive to PAX3 dosage, so that it is virtually pound heterozygotes sometimes have pheno- always disturbed by loss of function mutations types unexpectedly more severe or less severe in PAX3. Differentiation and survival of than expected on this model. The mi protein melanocytes is less sensitive, so that pigmen- dimer has been shown to bind to a DNA tary changes and hearing loss are much more sequence AGTCATGTGCT (the M box) variable features of the syndrome. Develop- found upstream of several melanocyte specific ment of limb buds is relatively insensitive to genes. 2 The CATGTG core ofthis sequence is PAX3 dosage and is normally disturbed only in a target for binding by several b-HLH-ZIP homozygotes. Occasional heterozygotes, how- proteins. Expression of mi in mouse 3T3 ever, do show signs of limb involvement, fibroblasts can cause them to undergo usually minor, and these are the mild WS3 melanocyte-like differentiation.63 It seems pos- patients discussed below. The effect of reduced sible that mi/MITF is a master gene switching PAX3 protein level could be milder or more on melanocyte development, although it is also severe, depending on variations in the un- expressed in heart, and the phenotypes of some known protein or DNA targets of PAX3 action mi mutants suggest additional functions. in development. These are the modifier genes In (fig 4B), MITF mutations have for PAX3 effects. been found in a modest number of families

In general there is no clear correlation with WS2,47 61 64 65 and in one family with the http://jmg.bmj.com/ between genotype and phenotype in WS 1. The phenotype of Tietz-Smith syndrome.47 Tietz- symptoms are very variable even within fami- Smith syndrome (MIM 103500) shows hear- lies, which is perhaps only to be expected if the ing loss combined with uniform non-patchy mechanism is haploinsufficiency; genetic back- dilution of pigmentation; as mentioned above, ground will have important modifying effects. some mi mouse mutants also show dilution However, mutation of 47 in the rather than spotting. In the dominant families

paired domain of PAX3 might have a special that attract the label of WS2, haploinsuffi- on September 30, 2021 by guest. Protected copyright. effect. Two families have been described that ciency rather than a dominant negative effect have atypically severe phenotypes segregating seems the most likely pathogenic with mutations at this site. In the only known mechanism.47 64 The wide range of recessive family with more than one case of WS3,56 four phenotypes seen in mice prompted a search for affected members have the mutation N47H,46 MITF mutations in patients with severe reces- while a small family having a phenotype sive pigmentary syndromes, but so far none has described as craniofacial-deafness- syn- been found. All the human families are drome (MIM 122880) have the mutation dominant, and with the exception of the single N47K.57 The three affected people in this fam- Tietz-Smith family, all have typical WS2. ily all have sensorineural hearing loss, dystopia Perhaps the incidence of hearing loss is rather canthorum, flexion contractures of the fingers, higher than in non-MITF WS2 families, but it and an almost complete absence of nasal is not strikingly different, and we cannot bones, but no pigmentary disturbances. predict which WS2 families will carry MITF mutations. MITF mutations have been found in only about 15% of families fitting the MITF and Waardenburg syndrome type diagnostic criteria for WS2, and the major II WS2 locus or loci remains to be found. Hughes et at8 mapped the mutation in one large WS2 family to 3p12-pl4. At the same PAX3 and Waardenburg syndrome type time, Tachibana et ar5 mapped MITF, the III human homologue of the mouse microphthal- WS3 remains something of an anomaly. Three mia gene to the same location. Microphthalmia rather separate combinations of auditory- had long been seen as a good candidate homo- pigmentary symptoms with hypoplasia or con- logue for WS60 and mutations in MITF were tractures of the upper limbs can be seen. (1) soon found in several WS2 families.6' Klein's original patient and the PAX3 homozy- Waardenburg syndrome 663

gote of Zlotogora et alf have profound hearing PAX3 and MITF illustrate how mutation loss, depigmentation much more severe than in analysis can help elucidate mechanisms of

WS 1, and a severe amyoplasia-like condition . For PAX3 the evidence for hap- J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from affecting the arms. (2) In the family reported loinsufficiency is convincing, with the caveat by Sheffer and Zlotogora,56 people hetero- that mutation of asparagine 47 may have some zygous for the PAX3 mutation N47H have additional effect. For MITF the mechanisms typical WS 1 plus significant amyoplasia, inher- are less certain, but it appears that here too ited as a dominant condition. The amyoplasia haploinsufficiency is a major factor in produc- is identical in pattern to that of Klein's patient, ing the relatively mild abnormalities of WS2 but less severe. (3) Finally, most cases labelled among heterozygous mutation carriers. Domi- WS3 are sporadic or part of WS 1 families, and nant negative effects and pure recessive effects have WS 1 plus quite minor contractures of the may also be found. elbows or fingers. Two cases we have tested are Clinically, the payoff from identifying the heterozygous for a 16 bp deletion and a 1 bp genes and characterising the mutations has so deletion, respectively, in the PAX3 gene, muta- far been relatively modest. Diagnostic labels tions typical of WS 1. have been refined. WS 1 and WS4 are now well defined genetic entities, while the label WS3 is , B, and largely redundant. WS2 remains a heterogene- Waardenburg syndrome type IV ous mix. Only a small proportion of type II Recent progress in identifying HSCR suscepti- families are accounted for by mutations in any bility genes has shed considerable light on WS4 of the genes defined so far. Moreover, the clini- and allowed a new definition of the syndrome. cal definition ofWS2 as a dominantly inherited Patients with mutations in the endothelin 3 patchy phenotype seems likely to exclude some gene, EDN3, or the gene for its receptor, patients with MITF mutations, for example, EDNRB, occasionally show a WS-HSCR phe- patients with dominant partial of the notype, especially if homozygous.66 67 75 76 Het- Tietz-Smith type or with major recessive erozygotes are usually unaffected or have syndromes that include severe depigmentation. isolated HSCR.61-70 A family reported by We can offer families an explanation of why Hofstra et al" is interesting: there is a pattern of they have Waardenburg syndrome, but we are low penetrance isolated deafness or pigmentary still unable to predict what features of WS any disturbances, resembling many families we particular PAX3, MITF, EDNRB, or EDN3 have seen that do not quite meet criteria for mutation will produce in a given person. WS2. In the family of Hofstra et al,7' when Fortunately there is little interest in prenatal cousins married they produced children with diagnosis among WS families. Nor is there typical WS4 who were homozygous for an much prospect for gene therapy, given that the EDN3 mutation. However, we have sought abnormalities of WS arise in the early embryo. EDN3 or EDNRB mutations in our families in Research into WS has allowed definition of vain (M Tassabehji, A P Read, unpublished transcription factors important in human http://jmg.bmj.com/ data). They are certainly not a common cause , not to mention find- ofWS in the absence of HSCR. Some patients ing the first homeobox gene to be implicated in with chromosomal deletions or translocations a human inherited disease. PAX3 and MITF affecting the sites of EDNRB at 13q22 or between them exemplify three major families of EDN3 at 20q13 have pigmentary disturbances transcription factors, the paired domain, without dystopia and with72 or without73 homeodomain, and b-HLH-Zip proteins.

HSCR. Mutations in other unidentified neural Many questions remain, particularly about the on September 30, 2021 by guest. Protected copyright. crest genes may also produce HSCR with pig- precise developmental pathways in which these mentary disturbances and maybe hearing loss, genes act, and the identity of modifier genes with or without dystopia. However, HSCR responsible for the highly variable expression in patients with RET mutations do not have families. It was unexpected that defects in the melanocyte defects (M Seri, personal commu- EDN3-EDNRB system would produce devel- nication), nor probably do those with muta- opmental abnormalities, and the role of tions in the RET , GDNF. in development has yet to be eluci- dated. Auditory-pigmentary syndromes, as Conclusions and summary defects in cell differentiation, always promised Research into Waardenburg syndrome pro- to be biologically interesting, and research so vides some ofthe best examples ofthe interplay far has amply borne out this promise. between mouse and human genetic research. PAX3 was investigated independently as a can- We wish to thank May Tassabehji for her superb laboratory didate gene for Splotch and Waardenburg work, and members of the Waardenburg Consortium for syndrome on the basis of its map location in sharing data and ideas. Our work was supported by the Hearing each species and expression pattern in the Research Trust and Wellcome Trust (grant 035301). mouse. The mouse mi, Ednrb, and Edn3 genes were each cloned when mice with random 1 Waardenburg PJ. Dystopia punctorum lachrimarum, blepharophimosis en partiele irisatrophie bij een doofs- transgene insertions unexpectedly showed the tomme. Ned Tschr Geneeskd 1948;92:3463-5. phenotypes of microphthalmia, lethal, 2 Waardenburg PJ. A new syndrome combining developmen- tal anomalies of the eyelids, eyebrows and nose root with and lethal spotting respectively. MITF and pigmentary defects of the iris and head hair and with con- EDNRB then became positional candidates for genital deafiess. Am J Hum Genet 195 1;3:195-253. 3 Arias S. Genetic heterogeneity in the Waardenburg the 3pl4 linked WS2 and 13q linked WS4 syndrome. Birth Defects 1971;7(4):87-101. respectively, while EDN3 was investigated as a 4 Klein D. Historical background and evidence for dominant inheritance ofthe Klein-Waardenburg syndrome (type II[). pure (non-positional) candidate gene for WS4. Am JrMed Genet 1983;14:231-9. 664 Read, Newton

5 Shah KN, Dalal SJ, Sheth PN, Joshi NC, Ambani LM. embryonic development, a preliminary report. C R Acad White forelock, pigmentary disorder of the irides and long Sci Paris 1995;318:57-66. segment Hirschsprung disease: possible variant of 41 Bober E, Franz T, Arnold HH, Gruss P, Tremblay P. Pax-3 Waardenburg syndrome. Pediatr 198 1;99:432-5. is required for the development of limb muscles: a possible _7 J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from 6 Liu XZ, Newton VE, Read AP. Waardenburg syndrome type role for migration of dermomyotomal muscle progenitor 2: phenotypic findings and diagnostic criteria. Am _7 Med cells. Development 1994;120:603-12. Genet 1995;55:95-100. 42 Epstein JA, Shapiro DN, Cheng J, Lam PYP, Maas RL. 7 Fraser GR. The causes of profound deafness in childhood. Pax3 modulates expression of the c-Met receptor during Baltimore: Johns Hopkins University Press, 1976. limb muscle development. Proc Natl Acad Sci USA 8 Arias S, Mota M. Apparent non-penetrance for dystopia in 1996;93:4213-18. Waardenburg syndrome type 1 with some hints on the 43 Underhill DA, Vogan KJ, Gros P. Analysis of the mouse diagnosis of dystopia canthorum. 7 Genet Hum 1978;26: Splotch-delayed mutation indicates that the Pax-3 paired 101-31. domain can influence homeodomain DNA-binding activ- 9 Newton VE. Waardenburg's syndrome: a comparison of ity. Proc NatlAcad Sci USA 1995;92:3692-6. biometric indices used to diagnose lateral displacement of 44 Chalepakis G, Jones FS, Edelman GM, Gruss P. Pax-3 con- the inner canthi. ScandAudiol 1989;18:221-3. tains domains for transcription activation and transcription 10 Tassabehji M, Read AP, Newton VE, et al. Mutations in the inhibition. Proc NatlAcad Sci USA 1994;91:12745-9. PAX3 gene causing Waardenburg syndrome type 1 and 45 Epstein JA, Cai J, Glaser T, Jepeal L, Maas R. Identification type 2. Nat Genet 1993;3:26-30. of a Pax paired domain recognition sequence and evidence 11 Fisch L. Deafness as part of an hereditary syndrome. .7 for DNA-dependent conformational changes. _7 Biol Chen Laryngol Otol 1959;73:355-82. 1994;269:8355-61. 12 Delleman JW, Hageman MJ. Ophthalmological findings in 46 Tassabehji M, Newton VE, Leverton K, et al. PAX3 gene 34 patients with Waardenburg syndrome. _7 Paediatr structure and mutations: close analogies between Waarden- Ophthalmol Strab 1978;15:341-5. burg syndrome type 1 and the Splotch mouse. Hum Mol 13 Goldberg MF. Waardenburg's syndrome with fundus and Genet 1994;3:1069-74. other abnormalities. Arch Ophthalmol 1966;76:797-809. 47 Tassabehji M, Newton VE, Liu XZ, et al. The mutational 14 Reed WB, Stone VM, Boder E, Ziprkowski L. Pigmentary spectrum in Waardenburg syndrome. Hum Mol Genet disorders in association with congenital deafness. Arch Der- 1995;4:2131-7. matol 1967;95:176-86. 48 Hoth CF, Milunsky A, Lipsky N, Sheffer R, Clarren SK, 15 Di George AM, Olmsted RW, Harley RD. Waardenburg's Baldwin CT. Mutations in the paired domain of the human syndrome. I Pediatr 1960;57:649-69. PAX3 gene cause Klein-Waardenburg syndrome (WS-III)

16 Farrer LA, Grundfast KM, Amos J, et al. Waardenburg syn- as well as Waardenburg syndrome type 1 (WS-1). Am _ drome (WS) type 1 is caused by defects at multiple loci, Hum Genet 1993;52:455-62. one of which is near ALPP on chromosome 2: first report 49 Baldwin CT, Hoth CF, Macina RA, Milunsky A. Mutations of the WS Consortium. AmJ7Hum Genet 1992;50:902-13. in PAX3 that cau,se Waardenburg syndrome type 1: ten new 17 Hageman M, Delleman J. Heterogeneity in Waardenburg mutations and review of the literature. Am _7 Med Genet syndrome. Am _7 Hum Genet 1977;29:468-85. 1995;58:115-22. 18 Newton VE. Hearing loss and Waardenburg syndrome: 50 Morell R, Friedman TB, Moeljopawiro S, Hartono, implications for genetic counselling. . Laryngol Otol 1990; Soewito, Asher JH. A frameshift mutation in the HuP2 104:97-103. paired domain of the probable human homolog of murine 19 Nemansky J, Hageman MJ. Tomographic findings of the Pax-3 is responsible for Waardenburg syndrome type 1 in inner ears of 24 patients with Waardenburg's syndrome. an Indonesian family. Hum Mol Genet 1992;1:243-7. A.7R 1975;124:250-5. 51 Morell R, Carey ML, Lalwani AK, Friedman TB, Asher JH. 20 da-Silva EO. Waardenburg I syndrome: a clinical and Three mutations in the paired homeodomain of PAX3 that genetic study of two large Brazilian kindreds, and literature cause Waardenburg syndrome type 1. Hum Hered 1997;47: review. Am J Med Genet 199 1;40:65-74. 38-41. 21 Pantke OA, Cohen MM. The Waardenburg syndrome. Birth 52 Lalwani AK, Brister JR, Fex J, et al. Further elucidation of Defects 1971;7(7):147-52. the structure of PAX3, and identification of two different 22 Giacola JP, Klein SW Waardenburg's syndrome with point mutations within the PAX3 homeobox that cause bilateral cleft lip. AmI Dis Child 1969;117:344-8. Waardenburg syndrome type 1 in two families. Am .7 Hum 23 Zlotogora J, Lerer I, Bar-David S, Ergaz Z, Abielovich D. Genet 1995;56:75-83. Homozygosity for Waardenburg syndrome. Am 7 Hum 53 Butt J, Greenberg J, Winship I, Sellars S, Beighton P, Ram- Genet 1995;56:1173-8. esar R. A splice junction mutation in PAX3 causes 24 Ayme S, Philip N. Possible homozygous Waardenburg syn- Waardenburg syndrome in a South African family. Hum drome in a fetus with exencephaly. Am .7 Med Genet 1995; Mol Genet 1994;3:197-8. CO. 59:263-5. 54 Xu W, Rould MA, Jun S, Desplan C, Pabo Crystal http://jmg.bmj.com/ 25 Hulten M, Honeyman MM, Mayne AJ, Tarlow MJ. structure of a paired domain-DNA complex at 2.5A Homozygosity in piebald trait. .7 Med Genet 1987;24:568- resolution reveals structural basis for Pax developmental 71. mutations. Cell 1995;80:639-50. 26 Steel KP, Bock GR. Hereditary inner-ear abnormalities in 55 Kissinger CR, Liu B, Martin-Blanco B, Kornberg TB, Pabo animals. Arch Otolaryngol 1983;109:22-9. CO. Crystal structure of an engrailed homeodomain-DNA 27 Steel KP, Barkway C. Another role for melanocytes: their complex at 2.8A resolution: a framework for understanding importance for normal stria vascularis development in the homeodomain-DNA interactions. Cell 1990;63:579-90. mammalian . Development 1989;107:453-63. 56 Sheffer R, Zlotogora J. Autosomal dominant inheritance of 28 Hearing VI. Unraveling the melanocyte. Am _7 Hum Genet Klein-Waardenburg syndrome. Am 7 Med Genet 1992;42: 1993;52: 1-7. 320-2.

29 Barsh GS. Pigmentation, , and genetic pathways 57 Asher JH, Sommer A, Morell R, Friedman TB. Missense on September 30, 2021 by guest. Protected copyright. in humans and mice. Am J Hum Genet 1995;57:743-7. mutation in the paired domain of PAX3 causes 30 Foy C, Newton VE, Wellesley D, Harris R, Read AP. craniofacial-deafness-hand syndrome. Humn Mutat 1996;7: Assignment of WS1 locus to human 2q37 and possible 30-5. homology between Waardenburg syndrome and the 58 Hughes A, Newton VE, Lu XZ, Read AP. A gene for 2 to human chromo- Splotch mouse. Am _ Hum Genet 1990;46:1017-23. Waardenburg syndrome type maps 31 Ishikiriyama S, Tonoki H, Shibuya Y, et al. Waardenburg some 3p12-p14.1. Nat Genet 1994;7:509-12. syndrome type I in a child with de novo inversion (2) 59 Tachibana M, Perez-Jurado LA, Nakayama A, et al. Cloning (q35q37.3). Am _Med Genet 1989;33:505-7. ofMITF, the human homolog of the mouse mzicrophthalmia 32 Martin D, Tucker DF, Gorman P, Sheer D, Spurr NK, gene and assignment to chromosome 3pi4.1-pl2.3. Hum Trowsdale J. The human alkaline phosphatase gene and Mol Genet 1994;3:553-7. related sequences map to chromosome 2 band 2q37. Ann 60 Asher JH, Friedman TB. Mouse and hamster mutants as Hum Genet 1987;51:145-52. models for Waardenburg syndrome in humans. .7 Med 33 Tsukamoto K, Tohma T, Ohta T, et al. Cloning and charac- Genet 1990;27:618-26. terization of the inversion breakpoint at chromosome 2q35 61 Tassabehji M, Newton VE, Read AP. Waardenburg in a patient with Waardenburg syndrome type 1. Hum Mol syndrome type 2 caused by mutations in the human micro- Genet 1992;1:315-17. phthalmia (MITF) gene. Nat Genet 1994;8:251-5. 34 Epstein DJ, Vekemans M, Gros P. Splotch (Sp2H), a muta- 62 Hemesath TJ, Steingrimsson E, McGill G, et al. Microph- tion affecting development of the mouse neural tube, shows thalmia, a critical factor in melanocyte development, a deletion within the paired homeodomain of Pax-3. Cell defines a discrete transcription factor family. Genes Dev 1991;67:767-74. 1994;8:2770-80. 35 Tassabehji M, Read AP, Newton VE, et al. Waardenburg 63 Tachibana M, Takeda K, Nobokuni Y, et al. Ectopic expres- syndrome patients have mutations in the human homo- sion of MITF, a gene for Waardenburg syndrome type 2, logue of the Pax-3 paired box gene. Nature 1992;355:635- converts fibroblasts to cells with melanocyte characteris- 6. tics. Nat Genet 1996;14:50-4. 36 Baldwin CT, Hoth CF, Amos JA, da-Silva EO, Milunsky A. 64 Nobokuni Y, Watanebe A, Takeda K, Skarka H, Tachibana An exonic mutation in the HuP2 paired domain gene M. Analyses of loss-of-function mutations of the MITF causes Waardenburg's syndrome. Nature 1992;355:637-8. gene suggest that haploinsufficiency is a cause of Waarden- 37 Goulding MD, Chalepakis G, Deutsch U, Erselius JR, burg syndrome type 2a. AmI Hum Genet 1996;59:76-83. Gruss P. Pax-3, a novel murine DNA-binding protein 65 Morell R, Spritz RA, Ho L, et al. Apparent digenic inherit- expressed during early neurogenesis. EMBO 7 1991;10: ance of Waardenburg syndrome type 2 (WS2) and 1135-47. autosomal recessive (AROA). Hum Mol 38 Gruss P, Walther C. Pax in development. Cell 1992;69:719- Genet 1997;6:659-64. 22. 66 Puffenberger EG, Hosoda K, Washington SS, et al. A 39 Stuart ET, Kioussi C, Gruss P. Mammalian . of the endothelin-B receptor gene in Annu Rev Genet 1994;28:219-36. multigenic Hirschsprung's disease. Cell 1994;79:1257-66. 40 Gerard M, Abitbol M, Delezoide AL, Dufier JL, Mallet J, 67 Chakravarti A. Endothelin receptor-mediated signaling in Vekemans M. PAX-genes expression during human Hirschsprung disease. Hum Mol Genet 1996;5:303-8. Waardenburg syndrome 665

68 Kusafuka T, Wang Y, Puri P. Novel mutations of 72 Hood OJ, Doyle M, Hebert AA, Oelberg DG. Association of the gene in isolated patients Waardenburg syndrome type II and a de novo balanced 7;20 with Hirschsprung's disease. Hum Mol Genet 1996;5: translocation. Dysmorphol Clin Genet 1989;3:122-3. 347-9. 73 Van Camp G, Van Thienen MN, Handig I, et al. 69 Auricchio A, Casari G, Staiano A, Ballabio A. Endothelin-B Chromosome 13q deletion with Waardenburg syndrome: J Med Genet: first published as 10.1136/jmg.34.8.656 on 1 August 1997. Downloaded from receptor mutations in patients with isolated Hirschsprung further evidence for a gene involved in neural crest function disease from a non-inbred population. Hum Mol Genet on 13q. Y Med Genet 1995;32:531-6. 1996;5:351-4. 74 Nakashima S, Sando I, Takahashi H, Hashida Y Temporal 70 Amiel J, Attie T, Jan D, et al. Heterozygous endothelin bone histopathologic findings ofWaardenburg's syndrome: receptor B (EDNRB) mutations in isolated Hirschsprung a case report. Laryngoscope 1992;102;563-7. disease. Hum Mol Genet 1996;5:355-7. 75 Attie T, Till M, Pelet A, et al. Mutation of the 71 Hofstra RM, Osinga J, Tan-Sindhunata G, et al. A endothelin-receptor B gene in Waardenburg-Hirschsprung homozygous mutation in the endothelin-3 gene associated disease. Hum Mol Genet 1995;4;2407-9. with a combined Waardenburg type 2 and Hirschsprung 76 Edery P, Attie T, Amiel J, et al. Mutation ofthe endothelin-3 phenotype (Shah-Waardenburg syndrome). Nat Genet gene in the Waardenburg-Hirschsprung disease (Shah- 1996;12:445-7. Waardenburg syndrome). Nat Genet 1996;12:442-4. http://jmg.bmj.com/ on September 30, 2021 by guest. Protected copyright.