Hearing loss and connexin 26

Hearing impairment is a sensory disability that affects millions of people all over the world. Though not life-threatening, it can become a major burden in social and professional life. In the industrialized world, deafness of infective and/or environmental origin has become less frequent, with a consequent rise in the proportion of hereditary hearing impairment. Deafness occurs in 1:1000 neonates¹ and the cause is hereditary in about half. This type of hearing impairment is sometimes referred to as prelingual, as it affects the child before the age of speech development. A distinction can be made between syndromic deafness, in which the deafness is accompanied by other specific abnormalities, and non-syndromic deafness (about 75%), in which there are no additional abnormalities. Approximately three-quarters of the non-syndromic forms are caused by a recessive disorder^1,2,3,4. Table 1 gives an overview of some epidemiological features.

Table 1.

Epidemiological features of prelingual hearing loss

	Prelingual hearing loss
Incidence	1/1000
Percentage of genetically caused cases	50%
Syndromic v non-syndromic genetic cases	25 v 75%
Autosomal dominant	20%
Autosomal recessive	74%
X-linked	5%
Mitochondrial	1%

Between 1997 and today, many non-syndromic hereditary forms of deafness have been localized on the human genome by genetic linkage techniques. Depending on the pattern of inheritance of the deafness, these loci are designated DFNA (autosomal dominant), DFNB (autosomal recessive) or DFN (X-linked). They are numbered in chronological order of discovery. For the majority of these loci the underlying disease-causing genes have not been identified so far. On the Hereditary Hearing Loss Homepage⁵ all these currently known forms of hereditary deafness are summarized. Tables 2,3,4,5, derived from this homepage, illustrate the achievements in this field of research. Certain research groups, having found preliminary evidence of a new locus, have claimed (‘reserved’) loci in advance. ‘Withdrawn’ indicates those which turned out not to be correct. Most of these genetic types of hearing impairment are quite rare, with the exception of DFNB1. This paper addresses DFNB1, which is caused by mutations in the connexin 26 gene.

Table 2.

Loci and genes associated with autosomal dominant non-syndromic hearing impairment, with the year of publication

Locus	Localization	Ref.	Year	Associated gene	Ref.	Year
DFNA1	5q31	León et al.	1992	HDIA1	Lynch et al.	1997
DFNA2	1p34	Coucke et al.	1994	GJB3 (CX31)	Xia et al.	1998
				KCNQ4	Kubisch et al.	1999
DFNA3	13q12	Chaib et al.	1994	GJB2 (CX26)	Denoyelle et al.	1998
				GJB6 (CX30)	Grifa et al.	1999
DFNA4	19q13	Chen et al.	1995
DFNA5	7p15	Van Camp et al.	1995	DFNA5	Van Laer et al.	1998
DFNA6	4p16.3	Lesperance et al.	1996	WFS1	Bespalova et al.*	2001
DFNA7	1q21-23	Fagerheim et al.	1996
DFNA8/12	11q22-24	Kirschhofer et al.	1996	TECTA	Verhoeven et al.	1998
DFNA9	14q12-13	Manolis et al.	1996	COCH	Robertson et al.	1998
DFNA10	6q22-23	Ò'Neill et al.	1996	EYA4	Wayne et al.	2001
DFNA11	11q12.3-21	Tamagawa et al.	1996	MYO7A	Liu et al.	1997
DFNA12	11q22-24	Verhoeven et al.	1997	TECTA	Verhoeven et al.	1998
DFNA13	6p21	Brown et al.	1997	COL11A2	McGuirt et al.	1999
DFNA14	4p16	Van Camp et al.	1999	WFS1	Bespalova et al.	2001
DFNA15	5q31	Vahava et al.	1998	POU4F3	Vahava et al.	1998
DFNA16	2q24	Fukushima et al.	1999
DFNA17	22q	Lalwani et al.	1999	MYH9	Lalwani et al.	2000
DFNA18	3q22	Boensch et al.	1998
DFNA19	10	Green et al.	1998
DFNA20	17q25	Morell et al.	2000
DFNA21
DFNA22	6q13	Melchionda et al.	2001	MYO6	Melchionda et al.	2001
DFNA23	14q21-22	Salam et al.	2000
DFNA24	4q	Häfner et al.	1999
DFNA25	12q21-24	Greene et al.	1999
DFNA26	17q25	Yang et al.	2000
DFNA27	4q12	Fridell et al.	1999
DFNA28	8q22	Anderson et al.	1999
DFNA29		Reserved			Reserved
DFNA30	15q26	Mangino et al.	1999
DFNA31		Withdrawn			Withdrawn
DFNA32	11p15	Li et al.	2000
DFNA33		Reserved			Reserved
DFNA34	1q44	Kurima et al.	2000
DFNA35		Reserved			Reserved
DFNA36	9q13-21	Kurima et al.	2000
DFNA37	1p21	Talebizadeh et al.	2000
DFNA38	4p16.3	Young et al.	2001	WFS1	Young et al.*	2001
DFNA39	4q21.3	Xiao et al.	2001	DSPP	Xiao et al.	2001
DFNA40		Reserved			Reserved

^*Unpublished observations

Table 3.

Loci and genes associated with autosomal recessive non-syndromic hearing impairment, with the year of publication

Locus	Localization	Ref.	Year	Associated gene	Ref.	Year
DFNB1	13q12	Guilford et al.	1994	GJB2 (Cx26)	Kelsell et al.	1997
DFNB2	11q13.5	Guilford et al.	1994	MYO7A	Liu et al.	1997
					Weil et al.	1997
DFNB3	17p11.2	Friedman et al.	1995	MYO15	Wang et al.	1998
DFNB4	7q31	Baldwin et al.	1995	SLC26A4	Li et al.	1998
DFNB5	14q12				Fukushima et al.	1995
DFNB6	3p14-p21				Fukushima et al.	1995
DFNB7	9q13-q21				Jain et al.	1995
DFNB8	21q22	Veske et al.	1996	TMPRSS3	Scott et al.	2001
DFNB9	2p22-23	Chaib et al.	1996	OTOF	Yasunaga et al.	1000
DFNB10	21q22.3	Bonné-Tamir et al.	1996	TMPRSS3	Scott et al.	2001
DFNB11	9q13-q21	Scott et al.	1997
DFNB12	10q21-q22	Chaib et al.	1996	CDH23	Bork et al.	2001
DFNB13	7q34-36	Mustapha et al.	1998
DFNB14	7q31	Mustapha et al.	1998
DFNB15	3q21-q25*	Chen et al.	1997
	19p13*
DFNB16	15q21-q22	Campbell et al.	1997
DFNB17	7q31	Greinwald et al.	1998
DFNB18	11p14-15.1	Jain et al.	1998
DFNB19	18p11	Green et al.	1998
DFNB20	11q25-qter	Moynihan et al.	1999
DFNB21	11q	Mustapha et al.	1999	TECTA	Mustapha et al.	1999
DFNB22					Reserved
DFNB23	10q11.2-q21				Reserved
DFNB24	11q23				Reserved
DFNB25	4p15.3-q12				Reserved
DFNB26	4q31	Riazuddin et al.	2000
DFNB27	2q23-q31	Pulleyn et al.	2000
DFNB28	22q13	Walsh et al.	2000
DFNB29	21q22	Wilcox et al.	2001	CLDN14	Wilcox et al.	2001
DFNB30	10p				Reserved

^*The two loci yielded similar LOD scores

Table 4.

Loci and genes associated with X-linked non-syndromic hearing impairment, with the year of publication

Locus	Localization	Ref.	Year	Associated gene	Ref.	Year
DFN1*	Xq22	Tranebjaerg et al.	1995	DDP	Jin et al.	1996
DFN2	Xq22	Tyson et al.	1996
DFN3	Xq21.1	De Kok et al.	1995	POU3F4	De Kok et al.	1995
DFN4	Xp21.2	Lalwani et al.	1994
DFN5					Withdrawn
DFN6	Xp22				del Castillo et al.	1996
DFN7					Withdrawn
DFN8					Reserved

^*Later recognized as syndromic

Table 5.

Mitochondrial mutations associated with non-syndromic hearing impairment, with the year of publication

Gene	Mutation	Ref.	Year
		Prezant et al.	1993
12S rRNA	1555A→G	Usami et al.	1997
		Estivill et al.	1998
		* Reid et al.	1994
	1445A→G	* Fischel-	1995
		Ghodsian et al.	1998
		* Sevior et al.
		* Tiranti et al.	1995
tRNASer	7472insC	* Jaksch et al.	1998
(UCN)		* Schuelke et al.	1998
		Verhoeven et al.	1999
	7510T→C	Hutchin et al.	1999
	7511T→C	Friedman et al.	1999
		Sue et al.	1999

^*Additional symptoms were present in some patients

HEARING IMPAIRMENT

Although the connexin 26 gene GJB2 is also involved in an autosomal dominant form of deafness (DFNA3), most mutations in this gene cause recessive hereditary bilateral deafness/hearing impairment, so-called DFNB1. This form of sensorineural non-syndromic hearing loss is prelingual and its severity varies from mild to profound, depending to some extent on the type of mutation^6,7. Hearing loss in the high-tone range has recently been described as a characteristic feature, but all frequencies are affected⁸. In two-thirds of cases, the hearing loss is non-progressive and there are usually no vestibular and/or labyrinthine abnormalities.

GENETICS

DFNB1 was the first locus incriminated in autosomal recessive deafness; in 1997 GJB2 was found to be responsible⁹. GJB2 is a small gene situated on chromosome 13q11; it has a length of about 5.5 kilobases. There are two exons, of which only one contains the coding sequence. The mRNA is 2.4 kilobases long and translates into a protein with 226 aminoacids. This protein belongs to the connexin family, which currently has more than a dozen members¹⁰.

THE PROTEIN (FIGURE 1)

Figure 1.

Schematic representation of a gap junction. Six connexins form a connexon. Two connexons of neighbouring cells form pores, which allow intercellular transport of small molecules (Adapted from Ref. 22)

Connexins are membrane proteins with four transmembrane domains. Six chains of these proteins form a complex (a hexamer), called connexon. Two hexamers in the membranes of adjacent cells form a cell-to-cell channel, a so-called gap junction, which allows the transport of small molecules and ions between cells. A hexamer can contain various types of connexin, and various types of hexamer can form cell-to-cell channels. The channel constituents determine which molecules or ions can pass through¹¹.

THE CELL (FIGURE 2)

Figure 2.

Schematic section through the human cochlea showing K⁺ recycling pathway and the expression of connexin 26 (GJB2). (Adapted from Ref. 23)

Recently, the hypothesis was put forward that CX26 protein is essential for maintaining the high K⁺ concentration in the endolymph of the inner ear. Sound stimulation of the ossicular chain causes vibrations in the endolymph. K⁺ ions enter the hair cells under the influence of these vibrations and the vibration signal is ultimately converted into a neural signal. The system is regenerated by the release of K⁺ from the hair cells into the supporting cells. The K⁺ ions are then passed from cell to cell via gap junctions and are eventually released into the endolymph. Except for sensorineural cells, the CX26 protein is present in gap junctions connecting all cell types in the cochlea, including the spiral limbus, the supporting cells, the spiral ligament and the basal and intermediate cells of the stria vascularis. It is therefore very likely that connexin 26 is involved in K⁺-recycling in the cochlea¹¹.

EPIDEMIOLOGY

Mutations in CX26 are the most common cause of autosomal recessive deafness throughout the world. This gene is believed relevant to half of all cases of hereditary deafness. CX26 shows diverse mutations, but one mutation occurs very frequently in Europe—the 35delG mutation. Average carrier frequency in Europe is 1:51 (north/middle Europe 1:79, south Europe 1:35)¹² (Table 6). In the Mediterranean countries the carrier frequency exceeds even that of the ΔF508 mutation in the CFTR gene which causes cystic fibrosis. Carrier frequencies in North America and Australia are comparable to those in north/middle Europe. In oriental populations and Ashkenazi Jews, other mutations in the same gene play a more important role (234delC¹³ and 176delT¹⁴, respectively). The high frequency of connexin-26-related hearing impairment in certain populations may be the result of the tradition of marriages between hearing-impaired persons¹⁵. The 35delG mutation gives rise to a severely shortened, non-functional protein¹⁶. More than sixty other, far less frequent, mutations have been described in CX26¹⁷. Uncertainty about the pathogenicity of some of the mutations complicates interpretation of mutation analysis¹⁸.

Table 6.

Carrier frequency of mutation 35delG in the GJB2 gene in 17 European countries (adapted from Ref. 12)

Country	Carrier frequency
Northern and central Europe
Norway	1/47.5
Denmark	1/190
Estonia	1/22.5
UK	0/119
Germany	1/50
Belgium	1/190
Holland	1/44.5
France (Brittany)	1/96
France	1/200
Czech Republic	1/48.7
Slovenia	1/182
Bulgaria	1/157
Total	1/79.3
Southern Europe
Portugal	1/45
Spain	1/40
Italy	1/32
Italy (Sardinia)	1/29.5
Malta	1/36
Greece	1/33
Turkey	1/37.5
Total	1/35.2
Average for Europe	1/51.1

Denoyelle et al.⁷ found mutations in the CX26 gene in 49% of the families from France, Great Britain and New Zealand who had severe to profound prelingual hearing loss. CX26 mutations were present in 51% of the group with, versus 31% in the group without, a clear familial history of hearing impairment; 86% of the CX26 mutations were 35delG mutations. Mueller et al.¹⁹ studied a group of 284 English patients with early childhood hearing impairment or deafness, with and without hereditary causes. They found CX26 mutations in 27.8% of the familial cases and in 7.9% of the sporadic cases; 70% of the CX26 mutations were 35delG mutations. This difference can be explained by the fact that families with different ethnic backgrounds were included in the study. The prevalence of non-familial, sporadic hearing impairment based on CX26 mutations in an English—Belgian population of 68 children was 10%²⁰.

DIAGNOSIS

An increasing number of medical centres can perform mutation analysis to determine involvement of the CX26 gene in congenital hearing impairment. This method has been available for several years at the department of medical genetics in Nijmegen. We retrospectively analysed the outcome of ninety-one CX26 mutation analysis requests covering a fixed period of time. Nineteen unrelated cases were shown to have two mutations in the gene. Twelve of them turned out to be homozygous, whereas four others were heterozygous for the 35delG mutation. Overall, the 35delG mutation was involved in 84% of the cases; thirteen cases originated from multiaffected families, whereas three others were sporadic cases. Information on the remaining three families could not be retrieved. Table 7 gives an overview of the CX26 mutations found in Nijmegen.

Table 7.

Overview of CX26 mutations found in Nijmegen

Patient	Mutation allele 1 →	Implication for protein structure	Mutation allele 2→	Implication for protein structure
1	35delG	Shortened	35delG	Shortened
2	35delG	Shortened	109G>A	V371
3	35delG	Shortened	35delG	Shortened
4	35delG	Shortened	313del14	Shortened
5	35delG	Shortened	35delG	Shortened
6*(f)	71G>A	W24X	407insA	Shortened
7*(m)	71G>A	W24X	427C>T	R143W
8*(d)	71G>A	W24X	407insA	Shortened
9	35delG	Shortened	35delG	Shortened
10	35delG	Shortened	35delG	Shortened
11	35delG	Shortened	35delG	Shortened
12	35delG	Shortened	35delG	Shortened
13	35delG	Shortened	35delG	Shortened
14	35delG	Shortened	35delG	Shortened
15	35delG	Shortened	449delT	Shortened
16	35delG	Shortened	35delG	Shortened
17	101T>C	M34T	427C>T	R143W
18	35delG	Shortened	IVS1+1G>A	Unknown
19	35delG	Shortened	35delG	Shortened
20	109G>A	V371	109G>A;	V371
21	35delG	Shortened	35delG	Shortened

^*Belong to same family; (f), father; (m), mother; (d), daughter

Mutation analysis applies not only to children with a clear family history, but also to children whose parents have normal hearing (sporadic cases). Moreover, if a mutation in CX26 is present, genetic counselling can be offered to provide information on the aetiology answers and on the likelihood of recurrence in future offspring. When a mutation analysis is positive there will usually be no need for further investigations such as imaging and ophthalmological tests, because other causes of congenital deafness no longer have to be excluded. In these cases, attention can immediately be focused on optimizing the child's hearing. Histopathological examination of the cochlea in a patient with confirmed CX26 mutation has revealed an intact acoustic nerve²¹. This means that these patients are suitable candidates for cochlear implantation, provided that their hearing loss is sufficiently profound. Early diagnosis leads to early treatment, which gives the best results with cochlear implantation.

CONCLUSION

Unlike many other genes CX26 is small, so that screening for mutations is fast and relatively simple. Besides, the overall high involvement of CX26 mutations in autosomal recessive non-syndromic forms of deafness, and even in sporadic cases, makes mutation analysis distinctly worth-while. CX26 mutation analysis has therefore secured a place as a useful tool in clinical practice. So far, many different mutations in the CX26 gene causing DFNB1 have been identified¹⁷. The uncertainty about the pathogenicity of the mutation demands close collaboration with geneticists who are familiar with deafness¹⁸. Nevertheless, CX26 mutation analysis provides a good starting-point in the molecular diagnosis of patients with non-syndromic congenital deafness.