Genetical, clinical, and functional analysis of a large international cohort of patients with autosomal recessive congenital ichthyosis due to mutations in NIPAL4

Autosomal recessive congenital ichthyosis (ARCI) belongs to a heterogeneous group of disorders of keratinization. To date, 10 genes have been identified to be causative for ARCI. NIPAL4 (Nipa‐Like Domain‐Containing 4) is the second most commonly mutated gene in ARCI. In this study, we present a large cohort of 101 families affected with ARCI carrying mutations in NIPAL4. We identified 16 novel mutations and increase the total number of pathogenic mutations in NIPAL4 to 34. Ultrastructural analysis of biopsies from six patients showed morphological abnormalities consistent with an ARCI EM type III. One patient with a homozygous splice site mutation, which leads to a loss of NIPAL4 mRNA, showed additional ultrastructural aberrations together with a more severe clinical phenotype. Our study gives insights into the frequency of mutations, a potential hot spot for mutations, and genotype–phenotype correlations.


| INTRODUCTION
Autosomal recessive congenital ichthyosis (ARCI) belongs to a heterogeneous group of rare genetic skin disorders and is characterized by abnormal scaling of the skin over the whole body (Fischer, 2009). Patients with ARCI are born as collodion babies or with congenital ichthyosiform erythroderma (CIE) and later develop lamellar ichthyosis (LI) with coarse brown scales or CIE with fine white scales (Oji et al., 2010). In cases of harlequin ichthyosis (HI), a rare, severe form of ARCI, perinatal mortality is still about 50% (Traupe, Fischer, & Oji, 2014). To date mutations in 10 genes have ichthyosis prematurity syndrome, which constitutes another congenital ichthyosis, which is classified as a syndromal form of ARCI, but leads de facto to nonsyndromic ichthyosis (Klar et al., 2009).
All ARCI genes encode proteins that are involved in the formation of the cornified lipid envelope (CLE) and many of them, including NIPAL4, in the synthesis of fatty acids and ceramides (Hirabayashi, Murakami, & Kihara, 2019;Mauldin et al., 2018). There is increasing evidence that NIPAL4 works as an Mg 2+ transporter for FATP4, an acyl-CoA synthetase, which is considered important for the generation of ω-hydroxy ultra-long chain fatty acid-CoA which is involved in the epidermal lipid metabolism (Li, Vahlquist, & Törmä, 2013;Mauldin et al., 2018). The resulting accumulation of long-chain free fatty acids (FFAs) in NIPAL4 deficient skin is assumed to trigger membrane stripping and disruption of organelle membranes (Mauldin et al., 2018), which disturbs overall metabolic and intracellular membrane component processing and eventually the transport of components for the formation of lamellar bilayers to the extracellular space by lamellar bodies. This may explain specific epidermal ultrastructural abnormalities identified in patients with ARCI having mutations in NIPAL4, formerly classified as ARCI EM type III, which is characterized by deposition of complex or empty vesicles and perinuclear elongated membranes in granular layer cells (Arnold, Anton-Lamprecht, Melz-Rothfuss, & Hartschuh, 1988;Dahlqvist et al., 2007). In addition to these cytotoxic events by putatively accumulating FFAs, NIPAL4 deficiency should lead to a lack of sphingolipid end products, resulting in a disturbed formation of the CLE. These two mechanisms may explain the impairment of the epidermal permeability barrier in patients with ARCI and thereby cause the observed clinical phenotype (Mauldin et al., 2018).

| METHODS
In our large collection of over 1,000 families with ichthyosis, 101 families with mutations in NIPAL4 were identified by Sanger sequencing or NGS methods. At least one affected member of each family was analyzed and informed consent was obtained from all patients. This study was conducted according to the Declaration of Helsinki principles.
Mutation screening was performed on genomic DNA isolated from peripheral blood lymphocytes. We applied standard methods for polymerase chain reaction (PCR) amplification, Sanger sequencing, and next-generation sequencing (NGS) to initially identify potential pathogenic variants. In the case of PCR amplification, exon-specific primer pairs were used to amplify all coding exons and flanking intronic sequences of NIPAL4 (reference NM_001099287.1, GRCh37.p13). Sequences of primers can be provided on request.  Hebsgaard et al., 1996), and NNSplice version 0.9 (http://www.fruitfly. org/; Reese, Eeckman, Kulp, & Haussler, 1997). The results of the prediction tools and databases for all NIPAL4 mutations found in our cohort are listed in Table S1.

| Reverse transcription PCR (RT-PCR)
Keratinocytes from patient P79 and control cell line (NHEKnormal human epidermal keratinocytes) were collected when cells reached confluency. Total RNA was extracted using the Roti-Quick kit (Carl Roth, Karlsruhe, Germany). The quality of RNA was analyzed by the NanoDrop spectrophotometer (PEQLAB, Erlangen, Germany) and 500 ng was reverse-transcribed using SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad) and oligo(dT)15 primers (Thermo Fisher Scientific, Waltham) according to the manufacturerʼs protocol.
In a first PCR a touchdown gradient from 63°C with a decrease of 0.5°C per cycle was used for the first 15 cycles followed by another 15 cycles at 55°C. For NIPAL4 mRNA, nested RT-PCR was performed using 5 µl of the initial PCR product under the same thermocycling conditions. The resulting RT-PCR products were analyzed by electrophoresis on a 1.5% agarose gel.

| Histological analysis of skin sections
Biopsies from patients P79, P101, and a healthy individual were fixed in 4% formaldehyde, gradually dehydrated, and embedded in paraffin

| Electron microscopy
Skin biopsies of six patients (P72, P73, P76, P80, P91, and P93) from the back were fixed for at least 2 hr at room temperature in 3% glutaraldehyde solution in 0.1 M cacodylate buffer pH 7.4, cut into 1 mm 3 pieces, washed in buffer, postfixed for 1 hr at 4°C in 1% aqueous osmium tetroxide, rinsed in water, dehydrated through graded ethanol solutions, transferred into propylene oxide, and embedded in epoxy resin (glycidether 100). Semithin and ultrathin sections were cut with an ultramicrotome (Reichert Ultracut E).
Semithin sections of 1 µm were stained with methylene blue. About 60-80 nm ultrathin sections were treated with uranyl acetate and lead citrate and examined with an electron microscope JEM 1400 equipped with a 2K TVIPS CCD Camera TemCam F216. Skin biopsies from individuals not suffering from ichthyosis were taken as controls.

Patients with ARCI from 101 families carrying mutations in NIPAL4
were identified by Sanger sequencing or by next-generation sequencing methods ( Table 1). Twenty of these pedigrees have already been published in Lefèvre et al. (2004), and Pigg et al. (2016). Altogether Year of birth Sex Origin

Consanguineous parents
Mutations in NIPAL4 Consequences of mutation (NM_001099287) (Amino acid level)  Consequently, mutation c.223+5G>C prevents the correct splicing of NIPAL4 and most likely results in a functional "knock-out" of NIPAL4 via nonsense mediated decay.

| Genotype-phenotype correlations in ARCI patients with mutations in NIPAL4
Retrospective assessment of clinical features of patients with mutations in NIPAL4, when available, reflected diverse clinical phenotypes as reported for other mutations in other genes associated with ARCI. In a part of our cohort we revealed more detailed information about the phenotype of our patients (Table 2). In these cases, ichthyosis phenotypes were frequently accompanied by an erythema, a palmoplantar keratoderma (PPK, Figure 3i Figure   4b+d). Hyperkeratosis was more pronounced in P79 than in P101, whereas P101 additionally showed a parakeratosis within the cornified layers (arrows in Figure 4d).

| DISCUSSION
This study summarizes mutation analysis and clinical data from 101 families with ARCI due to mutations in NIPAL4, and was complemented by functional studies on the molecular and ultrastructural level.
Mutation analysis in this cohort expands the current mutation spectrum in NIPAL4 by adding 16 novel pathogenic variants. Among these, there was a potential splice site variant c.612-3del found in a heterozygous state together with c.527C>A, which affects the highly conserved 3′ splicing consensus motif YAG/RNNN (Lewandowska, 2013). In this variant the pyrimidine residue (Y; here C) of the consensus motif is deleted and replaced by a purine residue (here G).
In silico analysis with different splice site prediction tools points towards a pathogenic effect of c.612-3del as it leads to a reduction or loss of recognition of the intrinsic splicing acceptor site.
The novel mutation c.463+5G>A interfered with the recognition of the adjacent donor splice site and therefore a part of intron 2 was detected in patient-derived cDNA. The position c.463+5 is a highly conserved position in human splice regions (Zhang, 1998). In silico translation of the altered mRNA sequence predicts a frameshift at amino In case of splice site variant c.223+5G>C, in vitro examination of NIPAL4 mRNA from P76 showed a complete loss of NIPAL4 mRNA, which confirmed its pathogenicity (Figure 2c). In addition to this, it becomes evident that NIPAL4 is expressed in peripheral blood lymphocytes, which makes it possible to analyze NIPAL4 mRNA without the need for a skin biopsy. Consequently, this offers an easily accessible and simple method to understand the effect of potential splice site mutations on a molecular level.
In our study we were able to provide substantial evidence for the existence of a mutational hot spot at position c.527 in NIPAL4, which had been detected before in smaller cohorts of ARCI patients with mutations in NIPAL4 (Alavi et al., 2012;Dahlqvist et al., 2007;Lefèvre et al., 2004 mutations could be estimated more precisely. Dahlqvist et al. (2007) detected this mutation on 37 out of 54 alleles (69%), and it is present in our cohort in 124 out of 202 alleles (61%). In this context, it has to be noted that c.527C>A, p.(Ala176Asp) was present in 18 out of 19 patients who were compound heterozygotes.
The mechanism, which causes this highly frequent mutation, remains elusive. Position 176 is located in the middle of a predicted transmembrane domain with an alpha-helical secondary structure (UniProt) and this alanine residue is conserved in mammals (crossspecies analysis, data not shown). Thus, a change from alanine, which is a strong alpha former, to aspartate, which is a weak alpha former (HGMD professional), might affect correct secondary structure formation and thereby might impair the proper functioning of NIPAL4 F I G U R E 4 Morphological and ultrastructural examination of skin specimens shows a variety of epidermal alterations caused by mutations in NIPAL4.
Consequently, it is likely that truncating mutations in NIPAL4 also result in a more severe form of ARCI disease. Nevertheless, two patients in our study with severe ARCI carried the highly recurrent missense mutation c.527C>A, p.(Ala176Asp). This reflects findings in two consanguineous Pakistani families (Wajid et al., 2010) with the same homozygous mutation in NIPAL4, who showed severe ichthyosis phenotypes with fine white scales over the whole body and face.
In one of these Pakistani families, some members had a brownish reticulated lamellar ichthyosis similar to patients in our cohort, who carried c.527C>A, p.(Ala176Asp) in a homozygous state. Hence, in case of c.527C>A, p.(Ala176Asp) different clinical outcomes are possible, which might also be due to other genetic modifiers.
Ultrastructural analysis showed characteristic morphological alterations like vesicular complexes, which can also be found in 12R lipoxygenase knockout mice (Epp et al., 2007) and, very rarely, in single patients with mutations in ALOXE3 or CYP4F22. These ultrastructural markers were elaborately described by Arnold et al. (1988) and Dahlqvist et al. (2007); in the latter communication, involving 27 cases of 18 families compared with 18 patients with ARCI but without consistent EM findings, the clear association with mutations in ICHTHYIN, now NIPAL4, was proven. One ARCI patient did not fulfill the EM criteria but carried a NIPAL4 mutation indicating that NIPAL4 deficiency does not necessarily result in a distinct clinical and also morphological phenotype. Concerning the origin of these vacuolar and membrane material, both groups assumed that these structures present defective lamellar bodies. However, there are intact lamellar bodies (in contrast to e.g., harlequin ichthyosis with mutations in ABCA12), and in view of deposition of excess membrane material in ichthyosis with mutations in PNPLA1 and FATP4, a more general disturbance of membrane trafficking and respective fatty acid components is more likely to play a role. Interestingly, the novel missense mutation c.695C>G in P72 caused a morphological phenotype resembling to some degree to PNPLA1 morphology (Grall et al., 2012). These observations hint to a strong association of ARCI pathogenesis and the reaction chain leading to intact ω-esterified ultralong chain fatty acid.
In summary, we present a large cohort of 101 families affected with ARCI carrying mutations in NIPAL4. We expand the mutational spectrum by presenting 16 novel mutations in NIPAL4. We provide insights to the distribution of mutations within NIPAL4, the frequency of their occurrence and potential genotype-phenotype correlations.