Novel and recurrent variants of ATP2C1 identified in patients with Hailey-Hailey disease

Hailey-Hailey disease (HHD) is a rare, late-onset autosomal dominant genodermatosis characterized by blisters, vesicular lesions, crusted erosions, and erythematous scaly plaques predominantly in intertriginous regions. HHD is caused by ATP2C1 mutations. About 180 distinct mutations have been identified so far; however, data of only few cases from Central Europe are available. The aim was to analyze the ATP2C1 gene in a cohort of Polish HHD patients. A group of 18 patients was enrolled in the study based on specific clinical symptoms. Mutations were detected using Sanger or next generation sequencing. In silico analysis was performed by prediction algorisms and dynamic structural modeling. In two cases, mRNA analysis was performed to confirm aberrant splicing. We detected 13 different mutations, including 8 novel, 2 recurrent (p.Gly850Ter and c.325-3 T > G), and 6 sporadic (c.423-1G > T, c.899 + 1G > A, p.Leu539Pro, p.Thr808TyrfsTer16, p.Gln855Arg and a complex allele: c.[1610C > G;1741 + 3A > G]). In silico analysis shows that all novel missense variants are pathogenic or likely pathogenic. We confirmed pathogenic status for two novel variants c.325-3 T > G and c.[1610C > G;1741 + 3A > G] by mRNA analysis. Our results broaden the knowledge about genetic heterogeneity in Central European patients with ATP2C1 mutations and also give further evidence that careful and multifactorial evaluation of variant pathogenicity status is essential.


Introduction
Hailey-Hailey disease (HHD, OMIM 16960, or Benign Chronic Pemphigus.) is a rare (incidence 1:50000) autosomal dominant genodermatosis. The symptoms, aggravating periodically, onset in third-fourth decade include blisters, vesicular lesions, crusted erosions, and erythematous scaly plaques, which occur mainly on groins, axillae, neck, and other intertriginous areas, and mucosa may also be involved. Lesions may be odorous and painful and lead to mobility affecting fissures (Li et al. 2016;Zamiri and Munro 2016). In histopathological findings, suprabasalar and intraepidermal keratinocyte acantholysis with a "dilapidated brick wall" appearance is due to abnormal epidermal Ca 2+ distribution by secretory pathway Ca(2+) ATPase 1 (hSPCA1) caused by mutation in its gene: calcium-transporting ATPase type 2C member 1 (ATP2C1) (Cheng et al. 2010;Micaroni et al. 2016;Cialfi et al. 2016). Importantly, ATP2C1 is expressed in all tissues, although HHD clinical symptoms are solely isolated to the skin. Four isoforms differing by alternative processing of the C-terminus are produced, but only few of ATP2C1 mutations localized beyond the core of 26 exons are present in each transcript (Nellen et al. 2017). The majority of ATP2C1 mutations lead to a premature termination codon (PTC); thus the dominant inheritance pattern of HHD seems to result from haploinsufficiency. Nevertheless, as around 1/3 of mutations lead to missenses or in-frame rearrangements, other mechanisms may be involved (Dobson-Stone et al. 2002;Kitajima 2002). Thus, to understand the pathophysiological molecular mechanism of HHD, further investigation is required. Worldwide, only about 300 individuals have been described so far with 179 distinct ATP2C1 variants (Nellen et al. 2017). The majority of them are Asians, and only few cases from Central Europe were published, including a not genotyped case report from Poland (Rácz et al. 2005;Sudbrak et al. 2000;Chlebicka et al. 2012).
Herein, we report the results of the first genetic investigation in 18 Polish HHD patients together with characterization of splicing mutations and in silico structural dynamic modeling of novel missense mutations.

Patients and methods
Eighteen probands of Polish descent (Table 1) with clinical HHD manifestation (according to Matsuda et al. (2014)) have been enrolled in the study, together with their relatives, if available. The average age at diagnosis was 29 years old (range: 15-40). All patients gave informed consent for participation.
All coding exons of ATP2C1 were analyzed using Sanger sequencing (primers and PCR conditions available on request) or panel next generation sequencing (customized KAPA Library Preparation Kit -Roche) using MiSeq (Illumina). The variants were annotated against NCBI RefSeq: NM_014382.3 and checked for presence in the GnomAD, ClinVar, HGMD Professional and ATP2C1 LOVD v.3.0 databases.
Novel missense mutations were analyzed using in silico algorithms: DANN, MutationTaster, FATHMM, FATHMM-MKL, GERP, MutationAssessor, SIFT, Provean, and Poly-Phen2, classified according to ACMG guidelines (Richards et al. 2015) and visualized using dynamic structural modeling (Yasara Structure Package v.15.7.12). Briefly, isoform 1a of hSPCA1 (NP_055197.2) was modeled by homology, using eight closest templates identified in RCSB Protein Data Bank (PDB) records, IDs: 3N5K, 4BEW, 1WPG, 2YN9, 2YFY, 2ZXE, 4RET, and 4HYT. For each template, up to five alternative sequential alignments have been tested. Finally, the best scored models were built on the basis of 1WPG (37.2% of sequence identity and 56.7% of sequence similarity within 802 residues of 919 being aligned) and 3N5K (36.8%, 56.2%, and 810, respectively) PDB records. However, the final hybrid model, which combines the optimal parts of the top models, was scored substantially higher than the latter and thus was further used.
Novel intronic mutations were evaluated with the use of three splice site prediction algorithms: MaxEnt, NNSPLICE, and HSF. In order to confirm the putative cryptic splicing of mutations c.325-3 T > G and c.1741 + 3A > G, RNA was isolated from peripheral blood leukocytes, reverse transcribed, and PCR amplified and analyzed using Sanger sequencing (Fig. 1). As negative and positive controls, we included RNA isolated from a healthy person and from a HHD patient with the already known mutation c.1308 + 1G > T.
Half (4/8) of the novel mutations could easily be classified as pathogenic due to introduction of a premature stop codon (p.Gly850Ter, p.Thr808TyrfsTer16) or change in the conserved consensus sequence of the canonical splice sites (c.423-1G > T, c.899 + 1G > A). The novel missense mutations, p.Gln855Arg, p.Leu539Pro and the p.Thr537Arg detected in cis with 1741+3A>G, were analysed using molecular dynamic modeling and standard in silico tools.
Molecular dynamic modeling showed that conversion of Gln855 into Arg would distort the transmembrane helical structure and form a positively charged region located in the M (1-10) transmembrane helices proximity of the ion channel, which seemingly would affect protein location and Ca 2+ transport. The effect of p.Leu539Pro is less clear; however it is possible that this substitution would destabilize hydrophobic core formed between β-sheet structures and hence influence ATPase alpha subunit interactions, which in turn could affect ATP binding, its hydrolysis, and finally ion transportation. Unfortunately, no family data were available for probands with p.Leu539Pro and p.Gln855Arg; thus the genotype-phenotype segregation could not be performed.
The clinical significance of another missense, the p.Thr537Arg in exon 18 is more difficult to evaluate. In GnomAD, no records for p.Thr537Arg can be found. The prediction algorithms (PolyPhen, SIFT, MutationTaster) indicated possible pathogenic effect of p.Thr537Arg. Contradictory to them, dynamic structural modeling showed that this solvent exposed substitution most probably does not result in significant conformational change. Furthermore, the p.Thr537Arg was found in cis with a novel, mutation c.1741 + 3A > G in intron 18, which leads to exon 18 in- Fig. 1 Results of functional analysis of three splicing mutations in ATP2C1 gene and modeling of altered protein products (ATPase2C1). Results of DNA genotyping: A1, E1, F1, nucleotide substitutions indicated by red arrows; results of cDNA analysis: A2, E2, F2 (patients samples) and control samples (A3, E3, F3); schematic view of altered and normal transcripts (C1, E4, F4); predicted structure models of ATPase2C1 protein: B wild type protein showing organization of transmembrane helices with amino acids 109-120 marked in blue; C2-protein lacking amino acids 109-120; Dwild type protein showing organization of ATP binding domain (amino acids 407-436 marked in magenta, amino acids 524-580 marked in green); E5protein lacking amino acids 407-436, F5 protein lacking amino acids 524-580 frame skipping as we have shown by mRNA analysis. Thus, the protein, if at all synthetized, lacks 57 codons including codon 537. This example of a complex allele containing two variants is not reported before, and c.[1610C > G;1741 + 3A > G], which both were assigned as potentially pathogenic by common prediction algorithms, draws attention on an important issue of careful pathogenicity status evaluation, especially when only selected exons are investigated. Importantly, when p.Thr537Arg status was evaluated alone, it was assigned as "likely pathogenic" using ACMG classification (Richards et al. 2015), which later changed into "uncertain significance" when we detected c.1741 + 3A > G and proved its impact on splicing.
Novel c.325-3 T > G and recurrent c.1308 + 1G > T mutations also lead to in-frame exons skipping (of exons 5 and 15, respectively). Moreover, given that skipping of exons 5 and 15 due to other mutations have been described before (Kitajima 2002;Matsuda et al. 2014;Xiao et al. 2019), our observation indicates that despite distinct molecular lesions, the functional effect of mutations may be similar, which could be significant with regard to the purposes of personalized treatment.
In summary, this is the first report of genetic analysis in Polish HHD patients. Thirteen variants were identified and characterized, including eight unreported before and two recurrent. The results further show heterogeneity in the ATP2C1 mutational spectrum, with possible ethnic-specificity. Last but not the least, by showing a case of complex allele c.[1610C > G;1741 + 3A > G], we also point that careful in silico and extended molecular analysis is essential with respect to proper interpretation of mutation pathogenicity.
Acknowledgments The authors would like to thank Michel van Geel, Ph.D., for his kind help and comments.
Funding information This study was funded by National Science Center (NCN) grant no: 2014/13/D/NZ5/03304. Most of these interactions are lost upon Gln855Arg replacement. Moreover, Gln 855 is proximal to the putative ion transporting channel (Asp, Asp, Asn, Asn), so Gln855Arg replacement most probably will also directly affect Ca transport.

PREDIC TED EFFECT OF SUBSTIT UTION
Unfavorable electrostatic interaction with a proximal Arg403 may slightly affect domain protein stability Disorganization of the hydrophobic core most probably results in a substantial decrease of the protein stability Distortion of transmembrane helices organization together with positively charged region located in the proximity of ion channel will most probably severely disrupt Ca transport .

POLY-PHEN 2
Probably damaging (score: Hum Div 0 Thr537, Leu539 and Gln855 are viewed as ball and stick models and indicated by green arrows.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Statement of informed consent Informed consent was obtained from all individual participants included in the study.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.