Human diseases associated with connexin mutations

Abstract

Gap junctions and hemichannels comprised of connexins impact many cellular processes. Significant advances in our understanding of the functional role of these channels have been made by the identification of a host of genetic diseases caused by connexin mutations. Prominent features of connexin disorders are the inability of other connexins expressed in the same cell type to compensate for the mutated one, and the ability of connexin mutants to dominantly influence the activity of other wild-type connexins. Functional studies have begun to identify some of the underlying mechanisms whereby connexin channel mutation contributes to the disease state. Detailed mechanistic understanding of these functional differences will help to facilitate new pathophysiology driven therapies for the diverse array of connexin genetic disorders.

1. Introduction

Most of the human congenital diseases were described clinically long before the underlying genetic causes were known. Patients with similar characteristics were identified and clinicians then developed diagnostic criteria based on disease features to classify different disorders. Sequencing of the human genome dramatically increased the identification of underlying causative mutations, particularly for rare disorders, and uncovered new roles for the members of many gene families in the function of multiple organ systems. This has proven true for the connexin gene family, where half of its members have currently been linked to a broad spectrum of genetic diseases. In some cases, the same phenotypic disease can be caused by mutations in three different connexin genes. In other cases, mutations within the same connexin gene can result in up to eight distinct clinical disorders. A major challenge for the field is to understand how the specific functional consequences of the gene mutations relate to the clinical similarities and differences between diseases. Here, we will summarize the currently known connexin genetic disorders and review some of the recent mechanistic advances in our understanding of how the underlying connexin mutations contribute to disease.

1.1. Connexins and intercellular communication

Gap junction channels are made from a family of proteins called connexins (Cx), and allow the direct passage of small molecules between adjacent cells, coupling them both metabolically and electrically [1, 2]. Connexins have four transmembrane domains, TM1–TM4, connected by two extracellular loops, E1 and E2, which mediate docking. The N- and C- termini, and a loop connecting TM2 and TM3 are on the cytoplasmic side of the plasma membrane [3, 4].

Connexins oligomerize in the ER-Golgi pathway into hemichannels (half a gap junction channel, also called a connexon) containing six connexin monomers [5, 6]. The hemichannels are transported to the plasma membrane, where they can act as functional channels by themselves, or move to regions of cell contact and find a partner hemichannel from an adjacent cell to form a complete gap junction channel [3]. Unopposed hemichannels become active, or modulate their activity, under conditions of mechanical, or ischemic stress, allowing the flux of small molecules like Ca²⁺, ATP, glutamate, or NAD⁺ across the plasma membrane, thereby eliciting signaling cascades and diverse physiological responses [7, 8].

Channels formed by different connexins are functionally distinct in terms of their gating, conductance and permeability characteristics [3, 9–13]. Genetic studies in mice have shown that these functional differences between connexins are important, since the loss of one isoform cannot be compensated for by replacement with another [14–16]. Virtually all cell types make more than one connexin at any given time, as illustrated by keratinocytes which express Cx26, Cx30, Cx30.3, Cx31, and Cx43 [17, 18]. The co-expression of multiple connexins within a single cell diversifies the composition of the channels that can assemble. Hemichannels can be made either from a single connexin isoform (homomeric), or from more than one type (heteromeric). The formation of heteromeric hemichannels depends on the abilities of different connexins to interact; not all connexins can co-oligomerize with each other [19–22]. Docking of hemichannels to form homotypic and heterotypic gap junction channels adds an additional level of complexity. Docking compatibility is regulated by specific sequences in the E2 domain [21, 23, 24]. Thus, genetic mutation of a single connexin gene may influence many types of channel, which may contribute to the diverse variety of disease that can result.

2. Human disorders caused by connexin mutations

Mutations in ten different human connexin genes have been linked to twenty-eight distinct genetic diseases (Table 1). Eight of these disorders result from mutations in Cx26, and an additional six diseases are caused by mutations in Cx43. The highest number of distinct mutations has been identified in Cx32. Consequently, a great deal of the mechanistic research has focused on these connexins (see section 3). As the sequencing technologies used to identify and link mutations to human disease continue to improve, additional connexinopathies may be identified [25].

Table 1.

Genetic disorders caused by human connexin mutations

Gene	Chromosome	Protein	Disorder(s)	OMIM
GJA1	6q22.31	Cx43	Craniometaphyseal dysplasia, autosomal recessive Erythrokeratodermia variabilis et progressiva Oculodentodigital dysplasia Oculodentodigital dysplasia, autosomal recessive Palmoplantar keratoderma with congenital alopecia Syndactyly, type III	218400 133200 164200 257850 104100 186100
GJA3	13q12.11	Cx46	Cataract	601885
GJA4	1p34.3	Cx37
GJA5	1q21.2	Cx40	Atrial fibrillation, familial, 11 Atrial standstill, digenic (GJA5/SCN5A)	614049 108770
GJA8	1q21.2	Cx50	Cataract	116200
GJA9	1p34.3	Cx59
GJA10	6q15	Cx62
GJB1	Xq13.1	Cx32	Charcot-Marie-Tooth neuropathy, X-linked 1	302800
GJB2	13q12.11	Cx26	Bart-Pumphrey syndrome Deafness, autosomal dominant 3A Deafness, autosomal recessive 1A Hystrix-like ichthyosis with deafness Keratitis-ichthyosis-deafness syndrome Keratoderma, palmoplantar, with deafness Vohwinkel syndrome Porokeratotic eccrine ostial and dermal duct nevus	149200 601544 220290 602540 148210 148350 124500
GJB3	1p34.3	Cx31	Deafness, autosomal dominant 2B Deafness, digenic, (GJB2/GJB3) Erythrokeratodermia variabilis et progressiva	612644 220290 133200
GJB4	1p34.3	Cx30.3	Erythrokeratodermia variabilis et progressiva	133200
GJB5	1p34.3	Cx31.1
GJB6	13q12.11	Cx30	Deafness, autosomal dominant 3B Deafness, autosomal recessive 1B Deafness, digenic (GJB2/GJB6) Ectodermal dysplasia 2, Clouston type	612643 612645 220290 129500
GJB7	6q14.3–q15	Cx25
GJC1	17q21.31	Cx45
GJC2	1q42.13	Cx47	Leukodystrophy, hypomyelinating, 2 Spastic paraplegia 44, autosomal recessive Lymphedema, hereditary, IC	608804 613206 613480
GJC3	7q22.1	Cx30.2
GJD2	15q14	Cx36
GJD3	17q21.2	Cx31.9
GJD4	10p11.21	Cx40.1
GJE1	6q24.1	Cx23

2.1. Cx26 mutations cause non-syndromic deafness, syndromic deafness with skin disease, or skin disease

Mutations in the GJB2 gene encoding Cx26 are common in humans, with carrier frequencies in unaffected individuals reaching 2–4% in several populations [26–29]. Most of the Cx26 mutations result in non-syndromic deafness [30, 31] which can be autosomal recessive (DFNB1A) or autosomal dominant (DFNA3A). Non-syndromic deafness is a partial or total loss of hearing that is not associated with other pathologies, and occurs in ~1 in 1000 newborn children. Cx26 mutations contribute to ~50% of genetic deafness and hearing loss can range from mild to profound, be stable or progressive, and impact different sound frequencies [32, 33]. An abundance of data suggests that either partial or total loss of function mutations in Cx26 contribute to non-syndromic deafness [34, 35]. In addition to non-syndromic hearing loss, five syndromic forms of deafness that include skin disease are caused to mutations in Cx26.

The syndromic deafness disorders are very rare, and can be divided into two broad groups. The first group includes Bart-Pumphrey syndrome (BPS) [36], Vohwinkel syndrome (VS) [37], and palmoplantar keratoderma with deafness (PPK) [38]. In addition to hearing loss and palmoplantar keratoderma (thickening of the skin on the palms and soles), BPS patients typically have leukonychia (white discoloration of the nails) and knuckle pads (verrucous growths on the knuckles) Skin abnormalities generally become noticeable during childhood, while hearing loss is present from birth [39]. Individuals affected with VS have hearing loss and thick, honeycomb-like calluses on the palms and soles that appear in infancy or early childhood. They may also show starfish-shaped patches of thickened skin on the tops of their digits or knees. VS patients develop constrictive bands of abnormal fibrous tissue around their fingers and toes that cut off the circulation to the digits leading to pseudoainhum (auto-amputation) [40]. Individuals with PPK have hearing loss and develop unusually thick skin on the palms of the hands and soles of the feet beginning in childhood with no other ectodermal abnormalities [41]. These three disorders have overlapping spectrums of both clinical similarities and specific pathogenic Cx26 mutations, and may represent phenotypic heterogeneity within a single broader genetic disorder [42].

Hystrix-like ichthyosis deafness syndrome (HID) [43] and Keratitis-ichthyosis-deafness (KID) syndrome [44, 45] make up the second group. HID is characterized by ichthyosis (dry, scaly skin) that can appear thick and spiky, resembling porcupine (hystrix) quills and profound hearing loss. Skin abnormalities are present at birth, become progressively worse, and eventually cover much of the body. HID patients may also have scarring alopecia (hair loss) [46, 47]. KID syndrome patients have hearing loss, keratitis (inflammation of the cornea) that can cause photophobia, scarring, and vison loss. They also have palmoplantar keratoderma, in addition to erythrokeratoderma (reddened skin) and ichthyosis (dry scaly skin). Partial alopecia is a common feature, frequently affecting the eyebrows and eyelashes. Squamous cell carcinoma occurs in ~15% of patients [48, 49]. Interestingly, only one causative mutation for HID syndrome has been identified, Cx26-D50N [43], which is also the most frequent Cx26 mutation found in KID patients [50]. Thus, HID may be a phenotypic variant of KID syndrome, with the most prominent difference being the absence of keratitis [42, 51]. Among KID patients, a strong genotype-phenotype correlation has emerged [52, 53]. Cx26-D50N patients live into adulthood, although they have vison loss from keratitis and are at high risk for developing squamous cell carcinomas [50, 54]. In contrast, ~10% of all KID patients die from septic complications in early childhood, and the most frequent mutations in this cohort are Cx26-G45E or Cx26-A88V [55–61].

The final Cx26 disorder is Porokeratotic eccrine ostial and dermal duct nevus (PEODDN), an epidermal nevus (an overgrowth of cells in a patch of skin) that follows the developmental path of ectodermal precursor cells, known as the lines of Blaschko [62]. PEODDN is caused by somatic mutations in Cx26 within these precursors, producing linear bands of diseased keratinocytes. The somatic mutations identified are the same as the germline Cx26 mutations causing KID syndrome [63, 64]. This suggested that PEODDN is a mosaic form of the epidermal features of KID syndrome caused by somatic mutations affecting only a subset of keratinocytes [65].

2.2. Cx30 mutations cause non-syndromic hearing loss, or skin disease

Mutations in the human GJB6 gene encoding Cx30 are less common than those in GJB2, but also contribute to non-syndromic deafness and skin disease. The first Cx30 mutation was found in a family with autosomal dominant (DFNA3B) middle-to-high frequency hearing loss [66]. This finding was consistent with the fact that the Cx26 and Cx30 genes share extensive sequence identity, are located close to each other on Chromosome 13, and both lie within the DFNA3/DFNB1 locus. It had been observed that a sizable fraction of deaf DFNB1 patients had only 1 mutant Cx26 allele at the GJB2 locus, and that additional DFNB1-linked familial cases with no Cx26 mutations had also been reported. It was then demonstrated that many of these patients were compound heterozygous for a mutation in one copy of Cx26 and a deletion in one copy of the GJB6 gene encoding Cx30. Two patients were homozygous for the Cx30 mutation, which was the second most frequent DFNB1 mutation in the Spanish population, and had no Cx26 mutations. Thus, mutations in Cx30 could contribute to autosomal recessive deafness (DFNB1B) by themselves [67], or cause deafness with a digenic pattern of inheritance in association with Cx26 [68, 69].

Mutations in Cx30 also cause the autosomal dominant disease ectodermal dysplasia 2, Clouston type, also known as Clouston syndrome [70, 71]. The main features of Clouston syndrome are thickened dystrophic nails, hair defects that include brittleness, short wiry hair, or total alopecia, and moderate to severe palmoplantar keratoderma [72]. In rare cases, the backs of the hands and feet can develop enlarged eccrine syringofibroadenomas, a benign eccrine tumor [73].

2.3. Cx30.3 mutations cause skin disease

Mutations in the GJB4 gene encoding Cx30.3 cause erythrokeratodermia variabilis et progressiva (EKVP), a skin disorder that presents at birth, or appears during infancy [74, 75]. EKVP is characterized by areas of hyperkeratosis (rough, thickened skin) that can be widespread over the body or limited to a small area, but are generally symmetric and fixed. Transient erythematous patches (red areas with increased blood flow in superficial capillaries) are a second major feature of EKVP. Approximately half of the patients also have palmoplantar keratoderma [74, 76]. In addition to being caused by Cx30.3 mutations, EKVP can also result from mutations in the GJB3 and GJA1 connexin genes encoding Cx31 and Cx43, respectively [77, 78]. Thus, in the case of EKVP, mutations in any one of three independent connexin genes can result in the same clinically defined disease [79].

2.4. Cx31 mutations cause nonsyndromic hearing loss, or skin disease

Like many of the other B-group connexins described thus far, mutations in the human GJB3 gene encoding Cx31 also contribute to non-syndromic deafness and skin disease. Cx31 mutations were initially found in two families with autosomal dominant (DFNA2B) progressive high frequency hearing loss [80]. As described above for Cx30, digenic inheritance of nonsyndromic deafness caused by mutations in the GJB2 and GJB3 genes was also found in 3 out of 108 Chinese patients with autosomal recessive deafness and only 1 mutant copy of Cx26. The unaffected parents were heterozygous for 1 of the mutant Cx26 or Cx31 alleles, consistent with a digenic mode of inheritance [81]. One of the heterozygous Cx31 mutations has also been found in a deaf Korean patient with a heterozygous Cx26 mutation [82]. Thus, mutations in Cx31 can contribute to autosomal dominant deafness (DFNA2B) by themselves, or cause deafness with a digenic pattern of inheritance in association with Cx26.

In addition to autosomal dominant or digeneic hearing loss, mutations in Cx31 can also cause the same skin disorder, erythrokeratodermia variabilis et progressive (EKVP), as described above for Cx30.3 [78].

2.5. Cx32 mutations cause peripheral neuropathy

X-linked Charcot-Marie-Tooth disease (CMTX1) was the first human genetic disorder linked to a connexin, and is caused by mutations in the GJB1 gene encoding Cx32 [83, 84]. Charcot-Marie-Tooth (CMT) disease is a clinically and genetically heterogeneous group of hereditary motor and sensory peripheral neuropathies which affects ~1 in 2500 people. CMTX1 is the second most common form of CMT, comprising ~5% of all CMT cases and more than 400 distinct Cx32 mutations have been identified to date [85, 86]. CMTX1 is a demyelinating neuropathy characterized by severely reduced motor nerve conduction velocity, and is considered to be a X-linked dominant, or X-linked intermediate disorder, as hemizygous males are usually more severely affected than heterozygous females. The symptoms of CMTX1 include progressive weakness and atrophy in distal leg muscles with onset in early childhood in affected males. Over time, patients develop foot abnormalities, abnormal gait, diminished or absent reflexes, sensory loss in distal limbs, and the thenar muscles of the hands become atrophied [86]. Studies of large cohorts of patients have suggested that clear genotype-phenotype correlations for Cx32 and CMTX1 remain enigmatic, despite the large number of mutations identified, the relatively frequent occurrence of the disease and the 24 year time span that has elapsed since the first mutations were identified [87, 88].

2.6. Cx40 mutations cause atrial fibrillation or standstill

Heterozygous somatic, or germline, mutations in the GJA5 gene encoding Cx40 cause familial atrial fibrillation-11 (ATFB11) [89–93]. ATFB11 is characterized by uncoordinated electrical activity in the atria of the heart, causing the heartbeat to become fast and irregular, in addition to increasing the risk of stroke and sudden death [94]. Gap junctions in the atrial myocardium contain predominately Cx40, with smaller amounts of Cx43 and Cx45 [95, 96].

Cx40 mutations can also contribute familial atrial standstill (ATRST1) with a digenic pattern of inheritance with mutations in the sodium voltage-gated channel alpha subunit 5 (SCN5A) gene [97]. ATRST1 is a rare condition characterized by the absence of electrical and mechanical activity in the atria that can be persistent or transient, as well as diffuse or partial. About half of the patients experience syncope (fainting) [98].

2.7. Cx43 mutations cause development disorders, or skin disease

Mutations in the GJA1 gene encoding Cx43 cause a variety of rare genetic diseases. Most mutations cause autosomal dominant oculodentodigital dysplasia (ODDD) [99, 100], which can also be inherited in an autosomal recessive pattern [101]. ODDD is a complex developmental disorder characterized by syndactaly (webbing of skin between fingers), microphthalmia (small eyes), craniofacial and dental abnormalities. In addition, some patients display microcephaly (small head), hypotrichosis (sparse hair growth), brittle nails, syndactyly of the toes, and a cleft palate [99, 102]. In exceptional cases, patients may also have with palmoplantar keratoderma [103]. In some patients with Cx43 mutations, syndactyly (type III) is the most prominent feature, and other typical ODDD manifestations are greatly reduced, or absent. Thus, it has been suggested that isolated syndactyly type III, which is also caused by Cx43 mutations, and ODDD may not be separate genetic conditions, but instead represent a broader spectrum of phenotypes within a single disease [104].

Cx43 mutations can also cause autosomal recessive craniometaphyseal dysplasia (CMDR) [105]. CMDR is a rare condition, evident at birth, which is characterized by progressive thickening of bones in the skull and abnormalities at the ends of long bones of the limbs. Patients display distinctive facial features including a prominent forehead, wide nasal bridge, wide-set eyes and a prominent jaw. X-rays reveal unusually shaped long bones, with wider metaphyses (ends), particularly in the legs. The typical ocular, digital, or dental features of ODDD are absent in CMDR patients [106, 107].

Cx43 mutations can also be linked to epidermal disorders such as palmoplantar keratoderma and congenital alopecia-1 (PPKCA1, also called keratoderma-hypotrichosis-leukonychia totalis) [108]. PPKCA1 is a rare autosomal dominant disorder characterized by severe hyperkeratosis, congenital alopecia and leukonychia [109].

Finally, mutations in Cx43 can also cause the same skin disorder, erythrokeratodermia variabilis et progressive (EKVP), as described above for Cx30.3 and Cx31 [77]. In Cx43-associated EKVP, patients had two additional features, prominent white lunulae and periorificial darkening, which had not been previously reported in the Cx30.3 and Cx31 linked cases.

2.8 Cx46 mutations cause cataract

Mutations in the GJA3 gene encoding Cx46 have been found to cause autosomal dominant cataract (CTRCT14), multiple types of which have been described as zonular pulverulent, posterior polar, nuclear coralliform, embryonal nuclear, and Coppock-like [110]. Bilateral opacities are diagnosed at a median age of 5 years, but can appear as early as 6 months. Opacities affect the lens nucleus and are often surrounded by dust-like precipitates in the lens cortex. Many patients require cataract surgery in childhood [111, 112]. No other ocular abnormalities have been reported. Mutations in connexins currently account for ~16% of all genetic cataract where the causative gene is known [113, 114].

2.9 Cx47 mutations cause leukodystrophy, spastic paraplegia, or lymphedema

Autosomal recessive mutations in the GJC2 gene encoding Cx47 cause hypomyelinating leukodystrophy-2 (HLD2, also called Pelizaeus-Merzbacher like disease) [115, 116]. HLD2 is an autosomal recessive leukodystrophy characterized by early-onset nystagmus (repetitive, uncontrolled eye movements), delayed motor development, progressive spasticity, ataxia (uncoordinated voluntary muscle movement), and diffuse leukodystrophy (degeneration of myelin) [117, 118].

Autosomal recessive mutations in Cx47 can also cause spastic paraplegia type 44 (SPG44) [119]. SPG44 is an exceptionally rare form of hereditary spastic paraplegia characterized by a milder phenotype than that of HLD2. Nystagmus is not present, and patients display a late-onset, slowly progressive spastic paraplegia associated with mild ataxia, dysarthria (slurred speech), dysmetria (difficulty judging distance, or scale) and mild cognitive impairment [118].

In addition, hereditary lymphedema type IC (LMPH1C) can be caused by autosomal dominant mutations in Cx47 [120, 121]. LMPH1C is caused by anatomic or functional defects in the lymphatic system, resulting in chronic swelling. There may be accompanying nail and skin changes, such as nail dysplasia or papillomatosis. Onset is often at birth, or early childhood, but can also occur later. Swelling occurs most frequently in the legs, but may also be manifest in the arms, face, genitalia and torso [122].

2.10 Cx50 mutations cause cataract

Similar to the situation described above for Cx46, mutations in the GJA8 gene encoding Cx50 have been found to cause multiple types of autosomal dominant cataract (CTRCT1) [123, 124]. These can be congenital, zonular pulverulent, nuclear progressive, nuclear pulverulent, stellate nuclear, nuclear total, total, and posterior subcapsular. In some cases, cataract caused by Cx50 mutations is associated with microcornea [114, 125]. Bilateral lens opacities may be first seen at birth, or early childhood, and often progress. The embryonic and fetal lens nuclei are usually affected and diffuse cortical opacities may appear in some patients. Of historical note, the Cx50 cataract locus was the very first human genetic disease to be assigned to an autosome [126, 127].

3. Functional consequences of connexin mutations

As described in section 2, Cx26 and Cx43 have the greatest number of distinct disorders linked to them, while Cx32 was the first Cx disease gene identified and has the largest number of known mutations. As a consequence, much of the mechanistic research has emphasized mutations in these three connexins. In the case of Cx43, it has been determined that GJA1 gene mutations cause ODDD by at least ten distinct mechanisms [100, 128]. In contrast, for Cx32, most of the GJB2 mutations fail to form functional channels, suggesting a consistent loss of function mechanism [86, 129, 130]. For brevity, we will focus here on the functional consequences of mutations in Cx26. Other recent reviews have examined the various proposed mechanisms of disease caused by mutations in Cx32 and Cx43, and a similar spectrum of functional consequences are found across many of the different connexin disorders. Cx26 mutations occur with the highest frequently in humans, with a vast majority resulting in non-syndromic deafness without skin pathology [30, 31].

3.1. Loss of function mutations

The majority of recessive nonsyndromic deafness Cx26 mutations found in patients are deletions, truncations, or frameshifts suggesting that total loss of gap junction and/or hemichannel function produces hearing loss. In support of this idea, many of the single amino acid substitution mutations are also devoid of channel activity [34, 35]. In the cochlea, Cx26 is expressed in the supporting cells of the Organ of Corti surrounding the sensory hair cells, and in the lateral wall containing the stria vascularis [131–135]. The stria vascularis produces the endolymph and generates the endocochlear potential, which is critical for sound transduction by the hair cells. Gap junctions in the stria vascularis may contribute to K⁺ transport and the generation of the endocochlear potential. However, within the sensory epithelium, the contribution of Cx26 gap junctions to K⁺ transport does not appear to be critical [136–140]. Instead, Cx26 gap junctions are more likely to transfer nutrients and signaling molecules like glucose or IP₃, which, in turn, help maintain sensory epithelial integrity and hair cell function [141–144]. There is also expression of Cx26 in basolateral and apical regions of supporting cells in regions devoid of cell-cell contact, suggesting that hemichannels may communicate with the perilymphatic and endolymphatic compartments [145]. The absence of skin disorders in patients with loss-of-function mutations suggests human skin does not need functional Cx26, either during development or for normal homeostasis, and that syndromic deafness is a Cx26 gain-of-function disease [51, 53, 146].

3.2. Gain of function mutations

Pathogenesis in syndromic deafness and skin disease caused by dominant Cx26 mutations may be caused by several mechanisms. Two gain of function changes have emerged for Cx26 mutations that cause PPK and deafness, the formation of active heteromeric hemichannels with Cx43, and trans-dominant inhibition of gap junction channels formed by other keratinocyte connexins. For six different Cx26 PPK mutations, it has been shown that they all lacked intrinsic gap junction channel activity when expressed as homotypic channels. In addition, five of the six displayed dominant-negative inhibition of coupling when co-expressed with wild-type Cx26, suggesting a mechanism for their contribution to hearing loss. When co-expressed with wild-type Cx30 or Cx43, all of these mutations also resulted in trans-dominant inhibition of gap junction conductance [147–151]. Thus, one common gain of function mechanism for Cx26 skin disease mutations linked to PPK is trans-dominant inhibition of other keratinocyte connexins, i.e. where the non-functional mutated connexin is able to inhibit the function of another wild-type connexin. In addition, one recent study demonstrated the ability of two Cx26 PPK mutations to form functional heteromeric hemichannels, even though both mutations failed to form functional homomeric hemichannels when expressed alone. Co-expression of Cx43 with either of the Cx26 mutants showed a significant increase in hemichannel activity when compared to Cx43 alone. Co-immunoprecipitation confirmed that Cx43 could be pulled down with either Cx26 PPK mutation, confirming the formation of heteromeric hemichannels [150].

The cause of pathogenesis caused by dominant Cx26 mutations linked to KID syndrome is thought to be due to the formation of aberrantly functioning hemichannels. Work from several labs has confirmed that dysregulation of mutant Cx26 hemichannels is a common gain of function, seen for all known KID mutations, with one exception [52, 152–159]. The lone exception was the Cx26-S17F mutation, which a recent study has shown results in increased hemichannel activity in the presence of wild-type Cx43, as described for some of the PPK mutations above [160]. Unlike the PPK mutations, several of the Cx26 KID mutations can also form functional gap junction channels [153–155].

The majority of KID syndrome mutants cluster in protein domains that contribute to the aqueous pore of the hemichannel, and that form the structural elements that regulate gating [53]. Thus, many KID mutations show altered voltage gating and reduced inhibition by extracellular Ca²⁺, which results in increased hemichannel activity. KID mutant residues identified as pore-lining using cysteine scanning result in hemichannels with altered unitary conductance and permeability to Ca²⁺, and likely other signaling molecules such as ATP [155, 156, 161]. Purinergic signaling plays an important role in cochlear function [162] and hemichannel-mediated release of Ca²⁺ and/or ATP has been implicated in proper cochlear development and sustaining auditory function [163, 164]. Ca²⁺ is also an important regulator of keratinocyte differentiation and the epidermis maintains a Ca²⁺ gradient, with the highest levels in the granular layer and an abrupt reduction in the cornified layer [165]. The altered Ca²⁺ permeation/gating properties of KID hemichannels would almost certainly alter this gradient and affect keratinocyte differentiation (Figure 2). Studies of mouse models of KID syndrome (Figure 1) have supported the hypothesis that mutant hemichannels disrupt calcium homeostasis in vivo. Transgenic mice expressing Cx26 KID mutations develop skin disease and display increased hemichannel activity in their keratinocytes [166, 167]. Examination of the epidermal Ca²⁺ gradient in the skin of these mice showed that it was elevated in the cornified layer. This change in the Ca²⁺ gradient was correlated with altered lipid secretion and defects in the epidermal water barrier [168].

Figure 2.

Disruption of the epidermal Ca²⁺ gradient by mutant Cx26 hemichannels in KID syndrome. (A). Cx26 KID mutant hemichannels have a higher net flux of Ca²⁺ and broader expression pattern than wild-type Cx26. (B). In normal epidermis, Cx26 is restricted to cells in the basal layer, and there is a gradient of Ca²⁺ with the highest levels in the granular layer, and low levels in the basal layer. In KID epidermis, mutant Cx26 expression expands across the epithelium, increasing Ca²⁺ flux and eliminating the normal Ca²⁺ gradient. The loss of the Ca²⁺ gradient disrupts the normal regulation of keratinocyte proliferation (favored in low Ca²⁺) and differentiation (favored in high Ca²⁺) resulting in a greatly thickened epidermis and a decrease in differentiated cells.

Figure 1.

A mouse model of KID syndrome. (A). A wild-type mouse of the SKH1 hairless strain. (B). An inducible transgenic SKH1 mouse expressing the Cx26-G45E mutation in keratinocytes develops the epidermal features of KID syndrome. See [166] for details.

3.3. Unifying themes in connexin congenital disorders

A growing body of evidence suggests that many, if not a majority, of the dominant connexin mutations associated with human genetic diseases consistently acquire unique gain of function activities [94, 128]. Eleven of the twenty-eight connexinopathies are skin diseases that are caused by mutations in five different connexin genes [42]. In the case of Cx26, increased hemichannel activity has been compellingly linked to mutations causing both KID syndrome, and more recently PPK.

Accumulating evidence suggests that dysregulated hemichannels may also play a pathogenic role in some of the other connexin skin disorders. For example, studies of Cx30 mutations associated with Clouston syndrome have provided evidence that increased hemichannel function may contribute to this skin diseases. Expression of these mutants in cultured cells induced large voltage-activated currents and caused cell death, consistent with formation of active connexin hemichannels. In addition, higher rates of ATP leakage were measured in cells expressing the mutant connexins than controls, implying that hemichannel mediated ATP release may contribute to Clouston syndrome pathology [169]. In a similar fashion, Cx31 mutations linked to EKVP were shown to result in increased hemichannel activity, ATP leakage, and necrotic cell death when expressed in transfected cells [170].

Additional support for abberant hemichannels being a general pathological mechanism contributing to skin disease came from analysis of a Cx43 mutation (Cx43-G8V) linked to PPKCA1. Expression of the Cx43-G8V mutant in transfected cells resulted in greatly increased whole cell membrane current, Ca²⁺ influx, and cell death when compared to wild-type Cx43, all consistent with increased hemichannel activity due to the Cx43-G8V mutation [108]. Consistent with these observations, we have found that expression of Cx43-G8V in single Xenopus oocytes (Figure 3, our unpublished data) results in readily detectable single channel currents that have a unitary conductance (215 pS) in accordance with what has been previously reported for wild-type Cx43 hemichannels [171].

Figure 3.

Hemichannel activity of the Cx43-G8V mutation causing PPKCA1 [108]. (A). A recording from a cell attached patch containing a single Cx43-G8V hemichannel obtained in response to an 8s voltage ramp applied between −70 and +70 mV. The leak current was subtracted offline. The hemichannel unitary conductance (γ_hc) was ~215 pS. (B). Another recording from a cell attached patch containing a single Cx43-G8V hemichannel obtained in response to a constant voltage step of −70 mV. At this voltage a sub conductance state was also evident. All currents were filtered at 1 kHz, data were acquired at 5 kHz and the pipette solution contained 140 mM KCl.

Taken together, these data suggest that there may be a common pathological mechanism across the diverse connexin mediated skin disorders. Thus far, the most consistent gain of function is the formation of dysregulated hemichannels that either increase ATP leakage, or disrupt Ca²⁺ signaling within keratinocytes. Further study of disease causing mutations in expression systems and transgenic mice will continue to refine our understanding of the epidermal pathophysiology.

4. Summary

Important roles for connexin channels in the generation and maintenance of tissue homeostasis has become increasingly evident through the continuing discovery of human disorders caused by mutations in connexin genes. Of note, the disease phenotypes in many cases are highly restricted, despite the fact that the connexins can be widely expressed. This aspect of connexin-mediated pathology is most evident in disorders like non-syndromic deafness and CMTX1 in which pathology is highly restricted despite the expression of the mutated Cx26 and Cx32 genes across many organ systems. These cases may reflect the possibility that compensatory mechanisms can be at play in a tissue-specific manner. On the other hand, confinement of phenotype to lens cataract for Cx46 and Cx50 mutations is consistent with their expression pattern, both are largely restricted to this organ. Finally, one of the remaining major challenges will be to discern the precise mechanism(s) whereby gene mutation leads to tissue pathology. For connexin channels, this likely comes down to identifying the biological signaling molecules that are transmitted through these channels and how this transmission is altered in mutated channels within an affected tissue.

Highlights.

• Twenty-eight human genetic diseases result from connexin mutations

• Mutations in two connexin genes encoding Cx26 and Cx43 cause fourteen disorders

• Both gap junction and/or hemichannel function can be compromised by mutations

• Diseases are usually restricted, despite many connexins being widely expressed