Mouse Models for Four Types of Waardenburg Syndrome

Authors


* Address reprint requests to Dr Masayoshi Tachibana, Research Division, Saitama Cancer Center, 818 Komuro, Ina, Saitama 362-0806, Japan. E-mail: mtachiba@cancer-c.pref.saitama.jp

Abstract

Waardenburg syndrome (WS) is an auditory-pigmentary syndrome caused by a deficiency of melanocytes and other neural crest-derived cells. Depending on a variety of symptoms associated with the auditory-pigmentary symptoms, WS is classified into four types: WS type 1 (WS1), WS2, WS3, and WS4. Six genes contributing to this syndrome –PAX3, SOX10, MITF, SLUG, EDN3 and EDNRB– have been cloned so far, all of them necessary for normal development of melanocytes. Mutant mice with coat color anomalies were helpful in identifying these genes, although the phenotypes of these mice did not necessarily perfectly match those of the four types of WS. Here we describe mice with mutations of murine homologs of WS genes and verify their suitability as models for WS with special interest in the cochlear disorder. The mice include splotch (Sp), microphthalmia (mi), Slugh−/−, WS4, JF1, lethal-spotting (ls), and Dominant megacolon (Dom). The influence of genetic background on the phenotypes of mice mutated in homologs of WS genes is also addressed. Finally, possible interactions among the six WS gene products are discussed.

Abbreviations –
bHLHZip

basic-helix-loop-helix-leucine zipper

cAMP

cyclic 3′5′-adenosine monophosphate

CREB

cAMP response element binding protein

Dom

dominant megacolon

EDN

endothelin

EDNRB

endothelin B-receptor

ERK

extracellular signal-regulated kinase

GSK3β

glycogen synthase kinase 3β

JF1

Japanese fancy mouse 1

LEF-1

lymphoid-enhancing factor 1

ls

lethal-spotting

MC1R

melanocortin type 1 receptor

MITF

microphthalmia-associated transcription factor

MSH

melanocyte stimulating hormone

PKA

protein kinase A

PKB

protein kinase B

Sp

splotch

WS

Waardenburg syndrome

Introduction

Waardenburg syndrome (WS) is a hereditary auditory-pigmentary syndrome, the major symptoms being congenital sensorineural hearing loss and pigmentary disturbance of eyes, hair, and skin (1). Depending on additional symptoms, WS can be classified into four types (Table 1): WS type 1 (WS1) is associated with facial deformity such as dystopia canthorum (lateral displacement of the inner canthi); WS2 has no other symptoms; WS3 is associated with upper limb deformity; and WS4, with megacolon (Hirschsprung disease). At least one gene responsible for each type of WS has been cloned and for these cloning procedures mice with pigmentation anomalies have contributed greatly. Here we review the mice models of WS, with special interest in the hearing organ.

Table 1.  Symptoms and genes of four types of Waardenburg syndrome (WS)
SyndromeSymptoms additional to main symptomsaResponsible gene Inheritanceb
  1. aMain symptoms = sensorineural hearing loss + pigmentary disturbance of eyes, hair, and skin.

  2. bAD = autosomal dominant; AR = autosomal recessive [modified from Read (2002) (65)].

WS1Dystopia canthorumPAX3 (transcription factor)AD
WS2Main symptoms only  (melanocyte-specific)MITF (transcription factor)AD
SLUG (transcription factor)AD
EDNRB? (suggested by this study)?
EDN3? (speculated)?
WS3 (Klein-WS)Hypoplasia of limb  muscle; contructure of  elbows, fingersPAX3 (transcription factor)AD
WS4 (Shah-WS)Hirschsprung's diseaseEDN3 (peptide hormone)AR
EDNRB (G-protein-coupled receptor)AR
SOX10 (transcription factor)AD

splotch (Sp) Mice as a Model for WS1

The first gene responsible for WS is the PAX3 gene, which encodes a transcription factor with paired domain and homeodomain. This human gene was cloned as a homolog of the mouse Pax-3 gene, which was identified as the gene responsible for splotch (Sp) mice (2). splotch mutations, which are associated with spina bifida and exencephaly, were analyzed at Sp2H and Spr alleles, and linkage analysis using interspecific backcrosses mapped to the Sp locus on mouse Chromosome 1 within the Inha to AKP3 interval. Subsequently the Pax-3 gene, located in this interval, was found mutated in Spr/+ mice; 32 nucleotides in Pax-3, which encodes a part of homeodomain, were deleted. Afterward, mutations of the human homolog, PAX3, were found in WS1 families (3, 4); these mutations of PAX3 in WS families usually compromise paired-domain or homeodomain (5).

Sp mice exhibit a number of characteristic developmental anomalies which predominantly affect the neural tube and neural crest (6). Severe alleles in six types of homozygous Sp mice are fatal at the embryonic stage, and even splotch-retarded (Spd) mice, which have the least severe allele, encoding Pax-3 with a substitution mutation at the paired domain, die at birth (7, 8). It was, therefore, not possible to determine the hearing function of homozygous Sp mice, although their pigmentation anomaly was overt. Heterozygous Sp (Sp/+) mice survive after birth and have white belly spots, but curiously, showed no sign of auditory defects (9); WS1 patients are usually heterozygous at the PAX3 gene and yet many show auditory dysfunction. This apparent distinction between mice and humans may be explained by differences in the amount of Pax-3/PAX3 protein necessary for normal development of cochlear melanocytes, as in the case of Mi/MITF protein (see below). The phenotype of Spd mice varies depending on their genetic background, suggesting the existence of modifier genes (10). Spd mice, coisogenic on the C57BL/6J (B6) genetic background, produce a white belly spot in heterozygotes with 100% penetrance and very few other anomalies, while many Spd BC1 progeny [F1 Spd/+ (Spd/+ B6 ×+/+ Mus spretus) × +/+ B6] exhibit highly variable craniofacial and pigmentary anomalies. It has been estimated that at least two genes interact with Spd to influence the craniofacial features.

microphthalmia (mi) Mice as a Model for WS2

Mice with a transgenic insertional mutation, termed VGA-9, were previously described and proposed as a model for WS2 (11). Homozygous VGA-9 mice were microphthalmic due to the loss of retinal pigmentary epithelial (RPE) cells, white in coat color due to the loss of melanocytes, and deaf. These phenotypes are similar to those of microphthalmia (mi) mice, and the insertional locus was allelic with the mi locus. The deafness of VGA-9 mice was due to a cochlear disorder resulting from the lack of intermediate cells in the stria vascularis. These cells are actually neural crest-derived melanocytes that are essential for the stria vascularis to produce endolymph and endolymphatic potential; a loss of strial intermediate cells results in the loss of endolymph and endolymphatic potential, causing endolymphatic collapse and sensory hair cell degeneration (12). Therefore, three major symptoms of VGA-9 mice – microphthalmia, white coat color, and deafness – could be explained solely by the loss of pigment cells, i.e. RPE cells and melanocytes. Soon after the first publication on VGA-9 mice, the responsible gene, microphthalmia (mi) and its human homolog, Microphthalmia-associated Transcription Factor (MITF), were cloned (13, 14). This gene encodes a protein with a basic-helix–loop–helix-leucine zipper (bHLHZip) structure, which is expressed in melanoblasts and RPE precursors. When these genes were introduced into cultured fibroblasts or neuroretinal cells, they induced a pigmented cell phenotype (15, 16), indicating their essential role in pigment cell differentiation (17). Moreover, MITF directly regulates an anti-apoptotic factor, BCL2, that modulates the survival of melanocytic cells (18).

In the same year (1994) that MITF was cloned, its mutations were identified in WS2 families (19); eight types of mutations have been identified so far and they are always transmitted in an autosomal-dominant manner (5, 20). This form of heredity is due to a haploinsufficiency of normal MITF (21, 22); the dominant negative mutation of MITF causes another auditory–pigmentary syndrome, Tietz syndrome (23). Recent findings have emphasized the importance of the dosage of MITF for its function. MITF, in a dosage-dependent manner, transactivates its own promoter by interacting with lymphoid-enhancing factor 1 (LEF-1), which binds LEF-1-binding sites in the promoter of the MITF gene (24). In contrast, 11 types of mi gene mutations, so far identified in mi mice, are transmitted in either a recessive or semi-dominant manner (20, 25–27). This suggests that a half dosage (compared with wild-type mice) of Mi protein from one normal allele functions for normal pigment cell differentiation in mice, while a half dosage of MITF protein is not sufficient in humans; ‘Human haploinsufficiency – one for sorrow, two for joy’ (28). A similar distinction between mice and humans has been shown in PAX/Pax-3 mutated WS1 individuals and Sp/+ mice (see above).

Slugh−/− Mice as a Model for WS2

Deletion mutations of the SLUG gene were found in two unrelated WS2 individuals, both showing heterochromia irides and profound hearing loss but no dysmorphic features (29). Neither case had a family history of hearing impairment or pigmentation anomaly, and the mode of inheritance was compatible with recessive heredity.

Slug, a zinc finger transcription gene, was first cloned as a Snail family member in chickens (30) and has been implicated in the epithelial-mesenchymal transition (31, 32). Analysis of a targeted null mice lacking murine homolog of SLUG, i.e. Slugh−/− mice, revealed that Slug is not essential for neural crest generation (32), but is required for full pigmentation of the coat. Slugh−/− mice have diluted coat color, a white forehead blaze, and areas of depigmentation on the ventral body, tail and feet (29). Hearing function has not yet been assessed in Slugh−/− mice, but hyperactivity and circling behaviors observed in some Slugh−/− mice suggest a vestibulo-cochlear disorder.

WS4 Mice as a Model for WS4

Homozygous and heterozygous mutations of the gene encoding the endothelin-B receptor (EDNRB) cause WS4 (33, 34). We have recently described mice with a phenotype closely resembling human WS4. Homozygous WS4 mice showed pigmentation anomaly (white coat color with black eye), aganglionic megacolon and cochlear disorder, and we named them WS4 mice (35). Breeding analyses revealed that WS4 mice were allelic with piebald-lethal and Japanese Fancy mouse 1 (JF1), in both of which the Ednrb gene is mutated (36, 37). Consistent with this finding, exons 2 and 3, which encode transmembrane domains III and IV of the Ednrb G-protein-coupled receptor protein, were deleted in these mice. Cochlea of WS4 mice showed endolymphatic collapse, resembling that in VGA-9 mice (11), due to the lack of melanocytes (intermediate cells) in the stria vascularis.

JF1 Mice as a Model for WS2 and WS4

The JF1 mice are an inbred strain of mice derived from Japanese wild mice (37, 38), which are often bred by Japanese laymen as fancy ‘panda’ mice because of their cute appearance with black eyes and white spotting on a black coat (Fig. 1A). While their Ednrb gene is mutated, unlike piebald-lethal mice, JF1 mice are not lethal even in the homozygous state. This non-lethality of JF1 mice is probably due to the fact that the mutation in JF1 mice is an insertional mutation in intron 1 that creates a cryptic splicing acceptor site that results in decreased expression of wild-type Ednrb but does not cause aganglionic megacolon (T. Kunieda, personal communication). However, we noticed that JF1 mice do not show a response to sound. Consistent with this finding, the cochlea showed endolymphatic collapse with no intermediate cells in the stria vascularis (Fig. 1B,C). As JF1 mice have pigmentation anomalies and hearing impairment – but not megacolon or dysmorphogenesis – they constitute a mouse model of WS2. These notions are consistent with the finding that WS2 is occasionally caused by mutations in EDNRB (29).

Figure 1.

(A–C). JF1 mouse and its cochlea. (A) JF1 mouse. JF1 mice have black eyes and black coat with large white spots and deafness. (B) Cochlea of JF1 mice showing endolymphatic collapse. Reissner's membrane (RM) shifted toward stria vascularis (SV) and become attached to tectorial membrane (TM), collapsing endolymphatic space. No hair cells are discernible. (C) Normal cochlea of Japanese wild mice. Outer hair cells (OHC) and inner hair cells (IHC) are discernible. D–F. Megacolon of (C57BL/6 × JF1)F2 mouse. (D) Macroscopic appearance of megacolon in a F2 mice (left) and normal colon of its littermate. Arrow indicates junction of normo- and mega-colon. St, stomach; Ce, cecum. (E) Histologic section around the junction. (F) Histological section of the colon distal to the junction. Note that no Auerbach's ganglion cells, which exist in normal colon (inset), are visible. G–J. Disorder of colon and cochlea of a Dom mouse. (G) One offspring (lower mouse) of heterozygous Dom mice was small compared with its littermate (upper mouse) and showed white patch at the belly. (H) Megacolon of this small Dom mice with white patch. (I) Cochlear disorder of this Dom mice with white patch and megacolon. Although there are vacuoles in the stria vascularis (SV) and spiral ligament (SL) in the area where these two face each other, RM is located in a normal position and no endolymphatic collapse is observed. However, no hair cells were discernible in the organ of Corti (arrow). (J) Cochlea of the normal littermate. OHC and Deiters’ cells (DC) are discernible.

One possible reason for the lack of megacolon in JF1 mice is the fact that mutation in the Ednrb gene affects only the expression level of wild-type Ednrb protein (37). Another possibility is the difference in genetic backgrounds; while the genetic background of JF1 mice is similar to that of Japanese wild mice (38), the genetic background of piebald-lethal mice, which shows megacolon, is that of laboratory mice. To test the latter possibility, we tried to introduce the JF1 allele (EdnrbJF1) into laboratory mice by mating JF1 with C57BL/6 laboratory mice; 92 F2 mice did not show megacolon, but two F2 mice did show it (Fig. 1D–E). This finding supports the notion that the phenotype of the EdnrbJF1 mutation is influenced by genetic background and suggests that laboratory mice carrying the EdnrbJF1 allele are potentially a mouse model for WS4.

lethal-spotted (ls) mice as a model for WS4

Homozygous mutations of the endothelin 3 gene (EDN3) cause WS4 in humans (39, 40), and homozygous mutations of the Edn3 gene cause coat spotting and aganglionic megacolon in ls mice and gene targeted edn3 null mice (41). Some of these mice can survive and mate; they are potentially a model for WS4, although cochlear disorders of these mice remain to be examined.

Dominant megacolon (Dom) mice as a model for WS4

Dom mice are mutant mice produced at the Jackson Laboratory (42). Homozygous Dom mice are lethal and their embryos lack neural crest-derived cells expressing the melanocyte lineage markers such as Kit, Mi and dopachrome tautomerase (43). Heterozygous Dom mice show white spotting and some of them show megacolon. The gene responsible has been mapped to mouse Chromosome 15 (42, 44), and a mutation of the Sry-related transcription factor gene, Sox10, was found (45, 46). Subsequently, mutations of the human ortholog SOX10 were found in human WS4 individuals (47–49). These findings led us to examine cochleas of Dom mice showing megacolon; of 120 offspring born from pairs of heterozygous Dom mice (B6C3Fe-a/aSox10Dom Jackson Lab., Stock no. 00029), two mice were small, had megacolon (Fig. 1G, H), and did not respond to sound. As judged by their black coat color with moderate white spotting at the belly, they seemed to be heterozygous Dom mice. Curiously, unlike other WS model mice, cochleas of these two Dom mice with megacolon did not show endolymphatic collapse, suggesting that their stria vascularis had intermediate cells (melanocytes) sufficient for normal production of endolymph. However, their organ of Corti, which is composed of sensory cells such as outer and inner hair cells and supporting cells such as Deiters’ cells, was missing (Fig. 1I). This may be related to a previous report that Sox10 was strongly expressed in supporting cells of newborn mice (50).

Cascade of Waardenburg Gene Products

As discussed above, so far four transcription factor genes have been cloned as genes responsible for the four types of WS: PAX3, MITF, SLUG and SOX10. We have previously shown that PAX3, a product of the WS1 gene, transactivates a WS2 gene, MITF (51, 52). Consequently, SOX10, a product of a WS4 gene, was shown to transactivate the MITF gene (50, 53–56), and recently MITF was shown to transactivate the SLUG gene, another WS2 gene (29). Thus, there is an epistatic cascade among the four WS genes, i.e. PAX3, SOX10, MITF, and SLUG.

While interaction of EDN3 and EDNRB, products of two WS4 genes, is obvious, coupling of this interaction with products of other WS genes remains to be elucidated. EDN1, which is also a potent ligand for EDNB, causes elevation of cyclic 3′,5′-adenosine monophosphate (cAMP) and melanogenesis in human melanocytes (57). cAMP plays a pivotal role in modulating melanogenesis through activation of protein kinase A (PKA) and cAMP response element binding protein (CREB), leading to up-regulation of MITF (58, 59). Up-regulation of MITF is also conducted by MITF itself through interaction with LEF-1 and LEF-1 binding sites in the promoter of the MITF gene (24). However, elevation of cAMP activates Ras and the extracellular signal-regulated kinase (ERK) cascade in melanocytic cells (59); ERK phosphorylates MITF at serine 73 and serine 409 and promotes its degradation, thereby leading to an inhibition of melanogenesis (61, 62). In addition to these classic pathways, elevation of cAMP, because of interaction between α-melanocyte-stimulating hormone (αMSH) and melanocortin type 1 receptor (MC1R), causes inhibition of phosphatidylinositol 3-kinase, inhibition of a serine/threonine kinase AKT (also termed protein kinase B or PKB) and activation of glycogen synthase kinase 3β (GSK3β), in that order (63). The EDN3–EDNRB interaction may cause elevation of cAMP, leading to activation of GSK3β. Activated GSK3β then phosphorylates MITF at serine 298 and increases the ability of MITF to bind the CATGTG core sequence in the tyrosinase gene promoter and to transactivate the gene (63, 64). As MITF binds the CATGTG sequence in SLUG promoters (29), EDN3–EDNRB interaction may also result in the enforced transactivation of the SLUG gene. Collectively, the gene products of the six WS genes work together, resulting in a cascade-like pathway (Fig. 2) regulating the differentiation, melanogenesis, and survival of melanocytes. Besides these six known WS genes, there may be other genes responsible for WS, occurring somewhere in this cascade-like pathway.

Figure 2.

Hierarchic relationship of gene products of six Waardenburg syndrome genes: EDN3, EDNRB, SOX10, PAX3, MITF and SLUG. Interaction of endothelin 3 (EDN3) and endothelin-B receptor (EDNRB) probably causes elevation of cyclic AMP (cAMP), which results in activation of extracellular signal-regulated kinase (ERK) and glycogen synthase kinase 3β (GSK3β), leading to phosphorylation of Microphthalmia-associated Transcription Factor (MITF). While ERK-phosphorylated MITF tends to degrade, GSK3β-phoshorylated MITF enhances binding activity to the target genes, including the SLUG gene. Elevated cAMP also activates protein kinase A (PKA) and phosphorylates cAMP response element binding protein (CREB), which binds with the MITF promoter. SOX10 and PAX3 also bind with the MITF promoter, while MITF binds with the SLUG promoter.

Ancillary