Article Text

Download PDFPDF

Original article
Mutation in KANK2, encoding a sequestering protein for steroid receptor coactivators, causes keratoderma and woolly hair
  1. Yuval Ramot1,2,
  2. Vered Molho-Pessach1,2,
  3. Tomer Meir3,
  4. Ruslana Alper-Pinus1,
  5. Ihab Siam1,
  6. Spiro Tams4,
  7. Sofia Babay2,
  8. Abraham Zlotogorski1,2
  1. 1Department of Dermatology, Hadassah—Hebrew University Medical Center, Jerusalem, Israel
  2. 2The Center for Genetic Diseases of The Skin and Hair, Hadassah—Hebrew University Medical Center, Jerusalem, Israel
  3. 3Department of Nephrology, Hadassah—Hebrew University Medical Center, Jerusalem, Israel
  4. 4Faculty of Medicine, The Palestinian Al Quds University, Abu Dis, The Palestinian Authority
  1. Correspondence to Professor Abraham Zlotogorski, Department of Dermatology, Hadassah—Hebrew University Medical Center, Jerusalem 9112001, Israel; zloto{at}


Background The combination of palmoplantar keratoderma and woolly hair is uncommon and reported as part of Naxos and Carvajal syndromes, both caused by mutations in desmosomal proteins and associated with cardiomyopathy. We describe two large consanguineous families with autosomal-recessive palmoplantar keratoderma and woolly hair, without cardiomyopathy and with no mutations in any known culprit gene. The aim of this study was to find the mutated gene in these families.

Methods and results Using whole-exome sequencing, we identified a homozygous missense c.2009C>T mutation in KANK2 in the patients (p.Ala670Val). KANK2 encodes the steroid receptor coactivator (SRC)-interacting protein (SIP), an ankyrin repeat containing protein, which sequesters SRCs in the cytoplasm and controls transcription activation of steroid receptors, among others, also of the vitamin D receptor (VDR). The mutation in KANK2 is predicted to abolish the sequestering abilities of SIP. Indeed, vitamin D-induced transactivation was increased in patient's keratinocytes. Furthermore, SRC-2 and SRC-3, coactivators of VDR and important components of epidermal differentiation, are localised to the nucleus of epidermal basal cells in patients, in contrast to the cytoplasmic distribution in the heterozygous control.

Conclusions These findings provide evidence that keratoderma and woolly hair can be caused by a non-desmosomal mechanism and further underline the importance of VDR for normal hair and skin phenotypes.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The combination of palmoplantar keratoderma in association with woolly hair is characteristic of Naxos disease (OMIM 601214) and Carvajal syndrome (OMIM 605676), two rare autosomal-recessive disorders. In both conditions, cardiomyopathy is an additional feature, which can often lead to ventricular arrhythmia and sudden death. Skin and hair manifestations appear prior to cardiac disease, and considering the potential risk of early sudden death in these patients, cardiac evaluation should be performed in any patient who presents with woolly hair and palmoplantar keratoderma.1 Naxos disease is caused by mutations in JUP, encoding plakoglobin (PKG), a desmosomal plaque constituent.2 Carvajal syndrome is caused by mutations in desmoplakin (DSP), another desmosomal component.3 More recently, mutations in desmocollin 2 (DSC2), an additional desmosomal protein, have also been found to underlie a cardiocutaneous phenotype similar to Naxos disease.4

While palmoplantar keratoderma and woolly hair are the characteristic features of Naxos and Carvajal syndromes, different disorders have been reported to result from mutations in JUP or DSP, depending on the type and site of mutation. For example, specific mutations in JUP can cause only skin fragility, keratoderma and woolly hair but no cardiomyopathy,5 and a homozygous nonsense mutation (p.Q539X) leads to lethal congenital epidermolysis bullosa.6 Such variation in clinical phenotypes was also observed in patients with DSP mutations,7 which can lead to only heart symptoms without skin involvement,8 skin disease without heart involvement9 and acantholytic disease with skin fragility,10 ,11 which can be lethal.10 While some understanding of the genotype–phenotype correlations with DSP and JUP mutations has recently been acquired, they are still in the most part poorly understood.12

Nuclear receptors serve major functions in the cells, translating hormonal signals into changes in gene expression. They are composed of a ligand-binding domain and DNA-binding domain, which interact with regulatory regions on target genes, leading to their activation or repression.13 Nuclear receptors play an important role in maintaining skin homeostasis, and they are involved in a myriad of cellular functions, including epidermal differentiation, proliferation, formation of the permeability barrier, follicular cycling and inflammatory reactions.14 Their function is tightly regulated by a number of corepressors and coactivators, of which proteins from the steroid receptor coactivator (SRC) family are especially important.15 These proteins participate in transcription initiation, elongation, RNA splicing, receptor and coregulator turnover, and mRNA translation.15 Therefore, it is not surprising that SRCs, and specifically SRC-2 and SRC-3, have been shown to be important also in the skin. They are especially crucial for epidermal differentiation, acting as coactivators of vitamin D receptor (VDR).16 Indeed, in human epidermis, SRC-3 is expressed in the granular layer of the epidermis17 and knockdown of SRC-2 and SRC-3 leads to downregulation of epidermal differentiation markers.18 Additionally, SRC-2 and SRC-3 are important for sphingolipid production and permeability barrier formation,19 and SRC-3 null mice show irregularities in permeability barrier formation and the innate immune response.20

The already complex regulatory machinery of nuclear receptors have become even more complicated, when it was revealed that the SRCs themselves are under tight regulation by the SRC-interacting protein (SIP), an ankyrin repeat-containing protein, responsible for sequestering SRCs in the cytoplasm.21 Through its ankyrin repeat domains, this protein allows buffering of the availability of these coactivators in the cells and thus helps to regulate transcription activation. Nevertheless, the exact function of this protein and its relevance to the skin are still obscure.

We have previously reported on a large consanguineous Arab family (family B in the original report22), with affected individuals showing palmoplantar keratoderma and hypotrichosis/woolly hair, suggesting Naxos disease or Carvajal syndrome. Nevertheless, symptoms of cardiomyopathy such as dyspnoea, syncope or sudden death were not recorded in this family.22 Mutations in JUP and DSP have been excluded in this family by direct sequencing, and other desmosome-related proteins were excluded by homozygosity mapping.22 We have recently identified an additional consanguineous Arab family with similar symptoms from the same geographical region. Hence, we have evaluated patients from both families, reassessed their clinical features and performed genetic and molecular studies in order to elucidate the causative gene for the disorder in these families.

Subjects and methods


The current study includes seven patients from two families from an Arab descent living in the same geographical region. The study was approved by the Institutional Review Board, and written informed consent was obtained from the subjects or their parents. All patients were carefully examined by the same study personnel. Blood samples were collected from 7 patients and 17 unaffected family members available for the study.

Genetic analysis

Exome sequencing was performed at Otogenetics corporation using Roche NimbleGen V2 (44.1 Mbp) paired-end sample preparation kit and Illumina HiSeq2000 at a 50× coverage. Sequence reads were aligned to the human genome reference sequence (build hg19) and variants were called and annotated using the DNAnexus software package. Sanger sequencing of suspected mutations was performed on ABI 3130 automated genetic analyser (Applied Biosystems, Foster City, California, USA). The mutation p.A670V was further analysed in 100 normal individuals of the same ethnic background and geographical region by restriction fragment length polymorphism with BseYI (following insertion of a mismatch in the primer used converting the original sequence from CTGCACTACTCCGTGTCTCA to GTGCACTACTCCGTGTCTCA).

Immunofluorescence microscopy

Paraffin-embedded skin biopsies were cut into 5 µm sections and then dewaxed in xylene (three times for 5 min) and rehydrated three times in ethanol (99%, 99% and 96% for 1 min) and twice in distilled water. Heat-mediated antigen retrieval was performed in a pressure cooker, filled with 10 mM sodium citrate buffer for SRC-2 antibody (pH 6.0) or TE buffer for SRC-3 antibody (pH 9.0) in which sections were treated at full pressure for 4 min. Subsequently, sections were blocked with 3% normal goat serum (Jackson ImmunoResearch Laboratories), 0.1% triton X-100 (Sigma) and 1% BSA in PBS. Afterwards, sections were incubated overnight with SRC-2 antibody (Clone 29, BD Biosciences, 1:250) or SRC-3 antibody (Clone 34, BD Bioscience, 1:75) in a humidified chamber at 4°C. Following washing, the sections were incubated for 1.5 h at room temperature with goat antimouse FITC-conjugated secondary antibody (DAKO). Nuclei were stained with DAPI (Santa Cruz Biotechnology).

Luciferase assays

Keratinocyte cultures were initiated from 3-mm-punch biopsies, obtained after local anaesthesia following Helsinki approval and signed informed consent. After overnight incubation at 4°C in trypsin–EDTA, the epidermis was separated and disaggregated in trypsin–EDTA to form a single-cell suspension. The cell suspension was inoculated into 25 cm2 Falcon flasks containing 2×106 lethally irradiated 3T3 mouse fibroblasts.23 Before the primary cultures achieved confluence, cells were released by trypsin 0.25%–EDTA 0.05% (1:1) and plated in 96-well tissue culture plates (5×104 cells in each well) in keratinocyte growth medium.24

Vitamin D response element (VDRE)-Luc and mVDRE-Luc plasmids,25 courtesy of V Rotter (The Weizmann Institute of Science, Rehovot, Israel), were used to assess VDRE activity. A TK-renilla luciferase plasmid was received from M Walker (The Weizmann Institute of Science, Rehovot, Israel) and used as a control for transfection efficiency in the luciferase assays. The plasmids were transfected into 60–70% confluent cell cultures using jetPRIME Transfection Reagent (Polyplus, Illkirch, France). Four hours after transfection, 10-8 or 10-7M 1,25(OH)2 vitamin D3 (Leo pharmaceutical products Denmark) or vehicle (EtOH) were added to the medium, and after 24 h the cells were harvested for luciferase assays. Dual-Luciferase Reporter Assay System (Promega, Madison, WI) was used in the luciferase analysis.

Statistical analysis

The data were expressed as mean±SEM. Statistical analyses were performed using a two-sided Student t test, and statistical significance was accepted at p<0.05.


Pedigrees and clinical characteristics

An updated pedigree of the first family previously reported (family 1) and the new recognised family (family 2) is shown in figure 1. All patients presented with a variable degree of striate palmoplantar keratoderma, which was generally more severe on the soles. Additionally, leukonychia was evident, more pronounced on the fingernails than on the toenails. Scalp hair was sparse, and in patients IV-1 and IV-13 also woolly. In addition to scalp hair, patients had sparse body hair as well as sparse eyelashes and eyebrows, especially on the lateral aspects. The fifth toes showed variable degree of pesudoainhum, ranging from external rotation to a deep sulcus at the digitoplantar fold, accompanied by bulbous appearance of the distal toe. The symptoms appeared to be more severe with advancing age, suggesting a progressive course of the disease. Notably, none of the patients had a history of dyspnoea, syncope or weakness. Additionally, there was no familial history of early sudden death. Electrocardiogram was normal in all patients. Echocardiography performed in the two oldest patients (IV-1 and IV-10) was normal. A skin biopsy stained with haematoxylin and eosin taken from patient IV-12 revealed non-epidermolytic keratoderma with regular acanthosis, and a normal granular layer.

Figure 1

Pedigrees and clinical findings. (A) Pedigrees of families 1 and 2 with woolly hair and keratoderma. (B) Clinical findings of patients in families 1 and 2. (i) True leukonychia of fingernails; (ii) striate keratoderma of palmar skin; (iii) keratoderma of plantar skin, more pronounced at pressure sites; (iv, v) sparse and woolly scalp hair; (vi) pseudoainhum showing deep sulcus at the digitoplantar fold of the fifth toe, and bulbous appearance of the fourth and fifth toes.

Genetic analysis reveals a mutation in KANK2

Blood samples were collected from a total of 7 patients and 17 healthy family members after consent was obtained from the Institutional Review Board and the patients. Exome sequencing was performed on genomic DNA from peripheral blood of individuals IV-1, IV-10 (patients) and IV-6 (healthy sibling) from family 1. We assumed an autosomal-recessive mode of inheritance and searched for homozygous coding, non-synonymous or frameshift mutations in both affected cousins (patients IV-1 and IV-10) not present in the healthy sibling (individual IV-6). We identified two such variations: a three base pair deletion in the RYR3 gene (13305-13307delAGA), which was ruled out as the causative gene by Sanger sequencing of DNA of other family members. The second variation identified was a c.2009C>T substitution, predicting a p.A670V missense mutation in KANK2, which encodes SIP (figure 2A). This mutation was validated by Sanger sequencing of all available DNA from family members, and segregated with the clinical phenotype in both families (figure 1A). This variant was absent in dbSNP (, the 1000 Genomes project data ( or the Exome Variant Server ( The mutation was predicted to be disease-causing by MutationTaster with a score of 0.99 (, probably damaging by PolyPhen-2 with a score of 1 (, and deleterious by the Protein Variation Effect Analyzer (PROVEAN,, with a score of −3.543.26–28 Several databases show a strong expression of the protein in the skin, including the PaxDb database (, neXtProt database ( and the human protein atlas ( Restriction analysis with BseYI was used to screen for the mutation (figure 2B). Two hundred control chromosomes from individuals from the same geographic area and with the same ethnic background were screened. One heterozygous individual was identified.

Figure 2

Genetic analysis and structural model of mutation location. (A) Mutation c.2009C>T in the KANK2 gene. (B) BseYI restriction analysis for the mutation c.2009C>T. MWM, molecular weight markers. (C) Schematic representation of the location of the mutation in the first ankyrin repeat (crystal structure is modified from PDB entry 4HBD and was drawn using PyMol (

In silico analysis suggests loss of ankyrin repeats of the steroid receptor coactivator-interacting protein

As reviewed in the ‘Introduction’ section, SIP is an ankyrin repeat-containing protein responsible for sequestering SRCs in the cytoplasm,21 thus regulating transcription activation. The ankyrin repeats in SIP are responsible for the interaction with the SRCs, which leads to their localisation to the cytoplasm. Loss of the ankyrin repeats led to nuclear localisation of SRCs and changed transcriptional regulation of nuclear receptors.21 The mutation reported here, in position 670 of the protein, localises to the first ankyrin repeat of the protein and is predicted to lead to loss of the ankyrin repeat and potential loss of the four additional downstream ankyrin repeats. A schematic representation of the location of the mutation in the first ankyrin repeat is shown in figure 2C (crystal structure is modified from PDB entry 4HBD ( and was drawn using PyMol (

Functional analyses suggest an abnormal sequestration of SRCs, leading to increased transactivation of VDRE

As explained above, SRCs are important components of the epidermis and have been found to take part in the more differentiated functions of keratinocytes.29 It is therefore postulated that the mutated protein might lead to disturbed regulation of SRCs, resulting in the observed keratoderma in our patients.

To determine whether the mutation in KANK2 affects SRC-2 and SRC-3 localisation in epidermal cells, we immunolabelled skin from patient IV-1 and his father (III-2) from family 2 by SRC-3 or SRC-2 antibodies. In the heterozygote father, SRC-3 was found to be expressed in gradually increasing intensity from the basal layer, via the spinous layer, to the granular layer (figure 3A,A′), similar to previous reports from healthy skin.17 Staining in the basal layer was mainly cytoplasmic and gradually became nuclear towards the granular layer. A similar pattern of expression was also observed for SRC-2. However, in the homozygous patient, staining was mainly nuclear for SRC-3 (figure 3A,A′) and SRC-2 (figure 3B,B′) already in the basal epidermal cells, in contrast to the cytoplasmic pattern in the heterozygous father. This suggests that the mutation in KANK2 affects the subcellular localisation of SRC-2 and SRC-3, as was already shown previously in vitro in cells lacking the ankyrin repeats of SIP.21

Figure 3

Subcellular localisation of SRC-2 and SRC-3. Subcellular localisation of SRC-3 (A) and SRC-2 (B) demonstrating nuclear staining in the spinous layer in both the homozygous patient and the heterozygous father, but in contrast to the heterozygous father, which demonstrates mainly cytoplasmic staining in the epidermal basal cells, the homozygous patient shows mainly nuclear staining in the basal cells. Scale bar=50 µm. (A′, B′) High-magnification confocal microscopy demonstrates nuclear staining of SRC-3 and SRC-2 in basal cells of a homozygous patient, but mainly perinuclear staining in the heterozygous father. Scale bar=5 µm. SRC, steroid receptor coactivator.

To examine the functional consequences of the mutation in the ankyrin repeat on epidermal keratinocytes, we used a promoter reporter assay by placing the firefly luciferase gene downstream of three tandem copies of the VDRE promoter, as previously described.25 We examined VDRE promoter activation in cultured keratinocytes from a normal healthy control, from a healthy heterozygous (family 1, patient. III-2) and from a homozygous patient (family 1, patient IV-10). Vitamin D-induced transactivation was significantly increased in the heterozygous sample compared with control. However, much stronger activation was seen in the homozygous patient, which was significantly higher than the control and the heterozygous samples (figure 4). Therefore, the mutation in the ankyrin repeat area of SIP is postulated to lead to increased transactivation of VDRE. A scheme demonstrating the effects of the mutation is provided in figure 5.

Figure 4

Vitamin D response element (VDRE) promoter activation. Vitamin D reporter plasmids were transfected into cultured keratinocytes and were treated with 1α25(OH)2D3 (10 nM or 100 nM) or vehicle for 24 h. Bars represent mean±SEM. *p<0.05; **p<0.01; ***p<0.001. Results represent triplicate determinations of samples. Het, heterozygote; Hom, homozygote.

Figure 5

Model of transactivation of the VDRE due to KANK2 mutations. In the normal state, SRC-2 and SRC-3 are sequestered by SIP, and therefore localise to the cytoplasm and do not activate VDRE. However, the A670V mutation disturbs the sequestration properties of SIP, leading to nuclear localisation of SRC-2 and SRC-3 to the nucleus, resulting in activation of VDRE and transcription of genes responsible for enhanced differentiation of epidermal cells and hair structure. SIP, SRC-interacting protein; SRC, steroid receptor coactivator; VDRE, vitamin D response element.


The importance of the VDR for normal hair growth has been clearly demonstrated in humans with VDR mutations, who have alopecia as one of their defining characteristics,30 ,31 and also in mice with VDR mutations.32 ,33 The hairless protein has been shown to be essential for normal hair growth and acts by corepressing the VDR.34 While the role of SRC on hair growth is still not entirely clear, it has been shown that mice with keratinocytes knocked down for another coactivator of VDR, MED1 (also named DRIP205), had disturbance of hair follicle cycling.20 Therefore, it is reasonable to assume that SRCs, being coactivators of the VDR, will also have a clear effect on hair growth and structure.

We found the mutation in 1 out of 200 control chromosomes that were screened for the mutation. It should be emphasised, though, that all controls were taken from the same region and from the same ethnic background. Therefore, due to the high prevalence of consanguinity in this region, a prevalence of 0.5% of the mutated allele in the observed population is realistic. Indeed, a high prevalence (1%) of a disease-causing mutated allele has been observed previously in another autosomal-recessive condition, the H syndrome, which is prevalent in the same population.35

Keratoderma with woolly hair (KWWH) has always been observed as a desmosome-related disease. However, here we show a skin and hair phenotype imitating the cutaneous features of Naxos disease and Carvajal syndrome, but in a non-desmosomal protein. Interestingly, although transcripts of the gene have been abundantly found also in the heart,21 we did not observe any cardiomyopathy in our patients according to both clinical and laboratory examinations. Nevertheless, many of the patients are still young, and long-term follow-up of these patients is warranted before heart involvement can be ruled out. Another intriguing fact is that only skin and its appendages are involved in this disorder, although SRCs are abundantly expressed and serve as coactivators of a large number of nuclear receptors, including the androgen receptor, progesterone receptor, glucocorticoid receptor and thyroid receptor.15 Of special interest is the observation that no changes were observed in the skeleton, taking into consideration the importance of vitamin D to the formation and homeostasis of bone. This might be explained by the fact that at least in normal state SRC-3 is not expressed in bone tissue,36 which might suggest that steroid coactivators are less important in regulating vitamin D homeostasis of the human skeleton.

In this report, we have highlighted the effects of the KANK2 mutation on the VDR. Usually, activation of VDR by vitamin D acts to induce differentiation in the skin and would theoretically ameliorate symptoms of keratoderma.20 However, recent evidence has emerged pointing at the fact that the VDR can act in a vitamin D-independent manner as an unliganded transcriptional receptor, which can silence the expression of inhibitory genes.37 Such inhibition can lead to hyperproliferation of basal keratinocytes. Furthermore, it is to be assumed that several more steroid receptors are affected, taken that many of them are expressed in the skin, and their activation is important for normal skin homeostasis.38 Indeed, SRCs are strongly implicated in hyperproliferative conditions such as cancer, and part of this process is due to direct coactivation of an array of transcription factors.39 Therefore, enhanced SRC-mediated signals may lead to hyperproliferation of basal keratinocytes and manifest clinically as hyperkeratosis. SIP is a recently characterised protein, and knowledge on its exact function and mechanism of action is still sparse.21 ,40 Here we show that it is of major importance for skin homeostasis, and set the stage for further studies to elucidate its exact function and interaction with nuclear receptor coactivators. Considering the increasing spectrum of genes responsible for KWWH, and taking into account our report on this clinical phenotype without cardiomyopathy, we believe a more accurate classification for this phenotype is warranted. Naxos disease should be classified as KWWH type I, caused by mutations in JUP; Carvajal syndrome should be classified as KWWH type II, caused by DSP mutations; the cardiocutaneous phenotype due to mutations in DSC2 would be classified as KWWH type III; and the disorder reported here should be classified as KWWH type IV due to mutations in KANK2. Thus, KWWH types I–III are characterised by cardiomyopathy and should warrant close cardiac surveillance and most likely defibrillator implantation, while KWWH type IV probably does not necessitate any cardiologic intervention.


We thank the patients and their families for participation in this study.



  • YR and VM-P contributed equally.

  • Contributors YR, VM-P and AZ conceived and designed the experiments. TM, RA-P and SB performed the experiments. YR, VM-P, RA-P, TM, SB, IS and AZ analysed the data. YR, VM-P and AZ wrote the paper. YR, VM-P, IS, ST and AZ undertook patient management, collection of samples and delineation of the phenotype.

  • Competing interests This study was supported in part by the Authority for Research and Development, Hebrew University of Jerusalem (AZ) and the Hadassah-Hebrew University Joint Research Fund (VM-P). We wish to thank Perry Tal (Stambolsky) and Varda Rotter for providing the plasmids, and Malka Chaouat for her excellent technical assistance.

  • Patient consent Obtained.

  • Ethics approval Hadassah Medical Center and the Israeli Ministry of Health.

  • Provenance and peer review Not commissioned; externally peer reviewed.