Molecular Genetic Dissection of Inflammatory Linear Verrucous Epidermal Naevus Leads to Successful Targeted Therapy

Molecular genetic inflammatory linear verrucous epidermal leads to successful targeted We identify here that heterozygous missense variants in CARD14 are a recurrent cause of this phenotype, leading to successful targeted medical therapy in one patient. Indications for treatment should be made on an individual patient basis. Genetic counselling should be considered in ILVEN, as in these cases it could be passed on as PRP or psoriasis. These findings underline the power of molecular genetic characterisation of rare diseases alongside clinical and histopathological phenotyping. a significant increase in and

TO THE EDITOR Inflammatory linear verrucous epidermal naevus (ILVEN) is a rare skin condition. Classically, it presents at birth or within the first year of life, frequently progressing during early childhood. Diagnostic criteria are erythematous verrucous hyperkeratosis in a fine and whorled Blaschko-linear pattern, intense pruritus, early age of onset, histological features, and resistance to treatment (Morag and Metzker, 1985). The cause of ILVEN has been unknown; however, a single case of mosaicism in gene GJA1 has recently been reported (Umegaki-Arao et al., 2017). We sought to investigate the genetics of ILVEN with a view to new therapeutic angles.
A total of 15 children with ILVEN and six normal controls (from surgery where excess normal skin was available) were recruited with written informed consent by their parents or guardians and Research Ethics Committee approval from the Great Ormond Street Hospital Research and Development office. The patients' parents/guardians consented to the publication of the patients' images. DNA and RNA were extracted from skin biopsies of the affected tissue, DNA was extracted from blood by standard methods and affected skin keratinocytes (KCs) were cultured and immortalized where possible (Lenti-HPV-16 E6/E7 Virus). Deep whole-exome sequencing of blood and affected skin was performed on patient samples, and data were analyzed using an optimized bioinformatic pathway for the detection of low-level somatic variants as previously published (Al-Olabi et al., 2018). Pathogenic GJA1 variants were not found in any patient. The clinical and histological features of patients 1 and 2 are shown in Figures 1  and 2a and b and Supplementary Table  S1.
Heterozygous missense variants in gene CARD14 were detected in 2 of 15 patients (Figure 2c and d). In both patients, the allelic load was compatible with that of a mosaic variant. In patient 1, the variant was present at 20% in both the blood and DNA extracted directly from a whole punch biopsy of the affected skin (c.356T > A, p. (M119K)); and in patient 2, it was present at 1% from DNA extracted directly from the epidermis of the affected skin and it was undetectable in the blood (4/ 313 reads in skin, c.277A>G, p.(K93E)). We had intended that wholeexome sequencing of the epidermis in patient 2 might have increased the mutant allele load; however, this was not the case, and the 1% load may have been due to mainly cornified epidermis being sequenced. However, both variants were convincing on whole-exome sequencing raw data, and both were clearly confirmed by Sanger sequencing (Figure 2e and   Inherited (nonmosaic) heterozygous mutations in CARD14 were recently described as rare causes of psoriasis (Jordan et al., 2012) and pityriasis rubra pilaris (Fuchs-Telem et al., 2012). Variants affecting certain domains of CARD14 were initially described as leading to the activation of NF-kB in the skin (Fuchs-Telem et al., 2012). However, differences between wild-type and variant CARD14 effects on NF-kB are modest (Li et al., 2015), and not all pathogenic variants increase the activation of NF-kB (Bertin et al., 2001). This includes some of those located in the CARD domain (amino acid sequences 15-107) (Israel and Mellett, 2018) such as that in patient 2. Treatment of patients with germline CARD14 variants with Ustekinumab has been highly successful (Eytan et al., 2014;Lwin et al., 2018); however, direct measurement of the effect of CARD14 variants on IL-12 and IL-23 expression has not previously been performed (Teng et al., 2015). Our findings suggest that IL-12 and IL-23 could be increased by CARD14 variants in a non-NF-kB-dependent manner.
Patient 1 had been resistant to multiple therapies (cyclosporine, acitretin, oral prednisolone), and she had faltering growth (height and weight below the 0.4th centile by age 3 years; birth weight 50the75th percentile). With hospital drug and therapeutics committee approval, we started treatment at the age of 6 years with Ustekinumab (0.75 mg/kg/ dose at 0 and 1 months and 3 monthly thereafter, as per psoriasis protocol). She has had a dramatic and sustained improvement in her skin, now 20 months into treatment, but has required an increase to 8-weekly dosing to maintain effect between doses. She also exhibited catch-up growth, with height and weight improving from the  (e, f) Sanger sequencing chromatograms confirm the variants. Cultured patient KCs and SVK 14 cells transfected with a mutant CARD14 construct express increased IL12 and IL23 at mRNA and protein level, proliferate faster than controls, and show variable activity of NF-kB p65. (g, h) QRT-PCR demonstrating a significant increase in IL-12A and IL-23A in cultured KCs from the affected skin from patient 2 and in SVK 14 cells transfected with the mutant CARD14 construct in comparison to control patient KCs (n ¼ 3) and SVK 14 cells transfected with the wild-type CARD14 construct, respectively. Mean relative gene expression of five replicates per patient sample and duplicates per SVK 14 sample was calculated with SD. (i, j) WST-1 proliferation assay showing a proliferation increase in KCs cultured from patient 2 and in SVK 14 cells transfected with the mutant CARD14 construct compared to control patient KCs (n ¼ 3) and SVK 14 cells transfected with the wild-type CARD14 construct, respectively, measured at 450 nm after 2 and 4 hours. The KCs were cultured for 8 days before proliferation measurement. The mean absorbance of five replicates is shown with SD. (k) Nuclear extracts from patient 2 KCs do not show a difference in NF-kB p65 activity when compared to control patient KCs (n ¼ 6). (l) Nuclear extracts from SVK 14 cells transfected with the mutant CARD14 construct show an increase in NF-kB p65 activity when compared with SVK 14 cells transfected with the wild-type CARD14 construct. The mean absorbance of triplicates for patient/control KCs and positive control is shown with SD. (m, n) Patient 2 KCs and SVK 14 cells transfected with the mutant CARD14 construct have significantly increased levels of IL-12 and IL-23 secreted in the supernatant compared to control KC cell lines (n ¼ 4) and SVK 14 cells transfected with the wild-type CARD14 construct, respectively. The mean absorbance of triplicates is shown with SD. All P-values were calculated by Students t-test using Prism, version 7.0 (GraphPad Software, San Diego, CA). Asterisks indicate a P-value < 0.05. <0.4th to 2e9th percentile within 3 months (Figure 1d and f) and no adverse effects. Patient 2 is younger and less symptomatic (Figure 1g and j) and has not required treatment.
Historically, there has been debate about the clinical and histopathological similarities of ILVEN to congenital hemidysplasia with ichthyosiform erythroderma and limb defects syndrome and to psoriasis (Happle, 1991;Ito et al., 1991;Moss and Burn, 1990;Welch et al., 1993). We consider that these debates are likely the result of genetic heterogeneity in ILVEN and that the term ILVEN is a clinical description rather than a single histopathological or genetic entity.
We identify in this study that heterozygous missense variants in CARD14 are a recurrent cause of this phenotype, leading to successful targeted medical therapy in one patient. Indications for treatment should be made on an individual patient basis. Genetic counseling should be considered in ILVEN as in these cases, it could be passed on as pityriasis rubra pilaris or psoriasis. These findings underline the power of molecular genetic characterization of rare diseases alongside clinical and histopathological phenotyping.

Data availability statement
No datasets were generated or analyzed during this study.   Written informed consent was centrally obtained for all UK Biobank participants. We correlated 200,000 PLCD1 exome sequences with inpatient diagnosis of TC. Of the 1,389 PLCD1 variants, six met the preselected threshold (P < 5 Â 10 e8 ) for association with TCs (Table 1). To determine variants independently associated with TCs, we estimated pairwise linkage disequilibrium among these six single nucleotide variants. Four associated single nucleotide variants were in high linkage disequilibrium with decreasing P-values adjacent to PLCD1 p.Ser460Leu. Heat maps (r 2 and D') for pairwise linkage disequilibrium of these single nucleotide variants are presented in Supplementary Figure S1. When PLCD1 p.Ser460Leu subjects were removed from the association analysis, only PLCD1 p.Glu455Lys was independently associated with TCs (P ¼ 9.35 Â 10 e95 ). Therefore, this targeted interrogation of the six significant single nucleotide variants revealed two independent risk alleles, p.Ser460Leu (minor allele frequency ¼ 0.030) and p.Glu455Lys (minor allele frequency ¼ 1.305 Â 10 e4 ).
MRI is useful to identify TCs (Adachi et al., 1996;Gossner and Larsen, 2010). TC size and number varied by risk allele status, with the larger and more frequent cysts found in p.Glu455Lys participants followed by p.Ser460Leu participants, with WT participants showing the smallest and fewest cysts (Figure 1). Of p.Glu455Lys participants, 91.7% (n ¼ 12) showed TCs on MRI compared with 24.0% of rs75495843 participants and 3.1% of WT participants (Supplementary Table S2). Among subjects with TCs on MRI, p.Glu455Lys participants had more cysts (mean AE SE, 3.8 AE 0.91) than p.Ser460Leu (2.2 AE 0.26) participants and controls (1.2 AE 0.08). The lone p.Glu455Lys participant without TCs on MRI had physician-diagnosed TCs that were likely excised.
We examined sex differences in TC penetrance as measured by both inpatient diagnosis and presence of cyst on MRI (Supplementary Table S3). Based on our previous study (Kolodney et al., 2020), we classified participants with either of the two PLCD1 risk variants as familial cyst cases and WT as sporadic cyst cases. Females were more likely to be diagnosed with familial TCs (crude OR, 1.35; 95% confidence interval [CI], 1.19e1.54) and less likely to be diagnosed with sporadic cysts (crude OR, 0.72; 95% CI, 0.68e0.77) than