Autosomal Recessive Keratoderma-Ichthyosis-Deafness (ARKID) Syndrome Is Caused by VPS33B Mutations Affecting Rab Protein Interaction and Collagen Modification

In this paper, we report three patients with severe palmoplantar keratoderma associated with ichthyosis and sensorineural deafness. Biallelic mutations were found in VPS33B, encoding VPS33B, a Sec1/Munc18 family protein that interacts with Rab11a and Rab25 proteins and is involved in trafficking of the collagen-modifying enzyme LH3. Two patients were homozygous for the missense variant p.Gly131Glu, whereas one patient was compound heterozygous for p.Gly131Glu and the splice site mutation c.240-1G>C, previously reported in patients with arthrogryposis renal dysfunction and cholestasis syndrome. We demonstrated the pathogenicity of variant p.Gly131Glu by assessing the interactions of the mutant VPS33B construct and its ability to traffic LH3. Compared with wild-type VPS33B, the p.Gly131Glu mutant VPS33B had reduced coimmunoprecipitation and colocalization with Rab11a and Rab25 and did not rescue LH3 trafficking. Confirming the cell-based experiments, we found deficient LH3-specific collagen lysine modifications in patients’ urine and skin fibroblasts. Additionally, the epidermal ultrastructure of the p.Gly131Glu patients mirrored defects in tamoxifen-inducible VPS33B-deficient Vps33bfl/fl-ERT2 mice. Both patients and murine models revealed an impaired epidermal structure, ascribed to aberrant secretion of lamellar bodies, which are essential for epidermal barrier formation. Our results demonstrate that p.Gly131Glu mutant VPS33B causes an autosomal recessive keratoderma-ichthyosis-deafness syndrome.

. Figure S1. Simplified and hypothetical pedigrees used for linkage analysis. In family 1, the index patient IV-1 (patient 1 in the manuscript), IV-2, III-1 and III-2 were genotyped and included in the linkage analysis. In family 2, the index patient V-3 (patient 2 in the manuscript), her healthy siblings V-2, V-4, V-5, her healthy parents IV-3, IV-4, and patient 3 (V-1) were genotyped and included in the linkage analysis.   shown is the junction between the stratum corneum and the stratum granulosum; that patient stratum corneum contains flattened nuclei and shows a reduction in KLK5 staining. Nuclei are counterstained with DAPI. Scale bars = 10 µm.

Linkage analysis and whole exome sequencing and mutation confirmation
The study was approved by the institutional review board of the Medical University of Innsbruck, and complied with the Declaration of Helsinki Principles. DNA samples from 3 affected individuals and 8 family members as shown in Figure S1 were hybridized to HumanCytoSNP-12v2 BeadChip (Illumina, CA) arrays (featuring approximately 300,000 markers). A whole-genome linkage scan was undertaken, and was based on a subset of 39.000 SNPs, selected by an inter-marker distance of 50 kb and a minor allele frequency of ≥0.15, and on autosomal recessive inheritance of the disorder with full penetrance ( Figure S2).
This resulted in 61,844 and 59,421 variants, respectively, which were annotated using annovar (Wang et al. 2010). Within the linkage region, 56 and 59 high-quality SNPs were found for patients 2 and 3, respectively. After removing synonymous SNPs and SNPs with allele frequencies >10% in the 1000 Genomes dataset, there were 2 SNPs left in each sample.

Antibodies and reagents
The following primary antibodies were used: anti-collagen IV (ab6586, Abcam, UK), anti- For immunostaining paraffin embedded patient skin sections were incubated in Histoclear, dehydrated in serial ethanol dilutions and antigen retrieval performed in citrate retrieval buffer (DAKO, Denmark) for two 5 min periods in a microwave at full power. Sections were blocked with 3% BSA in PBS with 0.5% Tween 20 and primary antibodies were also diluted in this blocking buffer. Sections were co-stained with DAPI, mounted with ProLong Gold (Life Technologies) and imaged on the SP5 system as described previously (Banushi et al. 2016).

In silico homology modeling
Modeling was performed as described previously (Banushi et al. 2016) based on the available structural data on the homologous human VPS33A-VPS16 complex (PDB ID 4BX9) and the C. thermophilum VPS33-VPS16 complex (PDB ID 4JC8). Structural images were created using PyMOL (Schrödinger 2010) and electrostatic potential calculations performed using APBS (Baker et al. 2001).

Mass spectrometry analysis
Urine samples from patients and fibroblasts from patient 2 were cultured and the lysates subsequently analyzed by mass spectroscopy as described previously (Banushi et al. 2016).
Samples from ARC patients (Banushi et al. 2016) were used as positive controls.

Histology
Five-millimeter punch biopsies were taken from the skin of the palms and abdomen.
Specimens were fixed in 4% formaldehyde, embedded in paraffin, sectioned (6 µm), and then stained with H&E. Mouse skin samples were fixed in 10% formalin and embedded in paraffin, sectioned and processed for H&E staining by the Biomedical Research Centre at the Institute of Child Health, London, UK.

Transmission electron microscopy
Patient samples were analyzed following both reduced osmium tetroxide and ruthenium tetroxide postfixation protocols (Elias 1996;Gruber et al. 2011). Murine skin biopsies were fixed in 4% glutaraldehyde in 0.1 M sodium phosphate and incubated with 1% osmium tetroxide 1.5% potassium ferricyanide overnight at 4 °C. A further overnight incubation in 1% tannic acid was performed before serial dehydration and further steps were performed and samples imaged as described previously (Banushi et al. 2016).