This study collected the clinical data of a proband with PC and AI and his seven unaffected family members. At the same time, laboratory tests such as blood routine, urine routine, C-reactive protein, liver function, renal function, blood lipid, onychomycosis culture, dermoscopy, and skin biopsy were performed on the proband. This study was approved by the Ethics Committee of the Union Hospital of Fujian Medical University. All participants provided written informed consent.

Approximately 2 mL of peripheral blood (EDTA anticoagulant) was collected from the patient and the patient’s parents. A blood genomic DNA extraction kit (Kangwei Century, China) was used to extract genomic DNA from the sample, and the NanoDrop (2000) instrument (Thermo Fisher, United States) was used to analyze the extracted nucleic acids. Genomic libraries were prepared using a standard library construction kit (self-developed by Mackinac), and the libraries were analyzed by NanoDrop (2000) and agarose gel electrophoresis. Using the GenCap liquid phase capture target gene technology (Maginot, China), 723 genes related to skin diseases were captured. The HiSeq 4000 (Illumina) sequencing platform was used for double-ended sequencing. The original sequencing data were removed from contamination and coupling sequences, and the filtered sequences were aligned to the human genome reference sequence (HG19) in the NCBI database using the BWA software (http://bio-bwa.sourceforge.net/). All SNPs (single-nucleotide polymorphisms) and INDELs (insertions and deletions) were analyzed and annotated using the GATK and ANNOVAR software. Missense variants were predicted for pathogenicity and conservation using PolyPhen-2, GERP+, SIFT, and MutationTaster software, and cleaved site changes were analyzed for pathogenicity using Spidex software.

The candidate variant sites obtained by analysis and screening were verified by Sanger sequencing, and the genotype of the parents was determined. The sequencing primers were forward primers: 5′-CTG​CCA​GCT​CAA​TTG​CC-3’ and reverse primers: 5 ‘- TAT​AAG​GAG​AGC​GGC​GGA​AC-3.’ Amplification was performed via polymerase chain reaction (PCR) using the following conditions: pre-denaturation at 98°C for 2 min, denaturation at 98°C for 10 s, annealing at 55°C–65°C for 30 s, and extension at 72°C for 10 s for 35 cycles. After 35 cycles, extension at 72°C for 1 min was performed. PCR products were examined via electrophoresis on a 2% agarose gel. The PCR-amplified target fragments were recovered, purified using a PCR purification kit (QIAGEN, Germany), and sequenced using an ABI 3730XL automated sequencer (Applied Biosystems, United States). The sequencing results were compared with those of the KRT17 (NM_000422.3) gene sequence (genome version GRCh37/hg19), and Mutation Surveyor software was used for analysis. If there were any abnormalities, reverse sequencing was performed.

KRT17 protein sequences were obtained from the National Center of Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/). The three-dimensional structure of KRT17 was obtained from the AlphaFold Protein Structure Database (https://alphafold.ebi.ac.uk/). We predicted the structure of the mutant using I-TASSER (https://zhanglab.ccmb.med.umich.edu/I-TASSER/), and PyMOL was used to analyze and visualize the model by referring to Version 7.6.

The plasmid synthesis scheme is shown in Figure 1. The expression vector used was pLVZG-CMV-MCS-EF1-copGFP-T2A-Puro to synthesize the KRT17 gene with double restriction sites: XbaI/NotI (Figures 1B–D). Next, various mutant plasmids were constructed as follows: wild-type plasmid 1 and pLVZG-CMV-KRT17 (WT)-3xflag-EF1-copGFP-T2A-Puro with double restriction sites XbaI/NotI and mutant plasmid 2 and pLVZG-CMV-KRT17 (c.274A>G)-3xflag-EF1-copGFP-T2A-Puro with double restriction sites XbaI/NotI, encompassing the 274A>G variant site of KRT17. Amplification, validation, and sequencing of the target genes were performed. Human embryonic kidney 293 cell line 293FT cells were used to package plasmids into Lentivirus, the cell culture supernatant produced within 72 h was collected using a sterile filter, and then the virus concentrated and purified. Recombinant lentiviral vectors KRT17 (c.274A>G) and its negative control KRT17 (WT) were constructed by the third-generation lentiviral packaging system (Zolgene Biotechnology, China).

Human immortalized keratinocyte line HaCaT cells (CTCC, United States) were cultured in Dulbecco’s modified Eagle’s medium with high glucose containing 10% fetal bovine serum (Thermo Fisher Scientific, United States). Well-grown HaCaT cells were seeded into six-well plates at 1 × 10⁵ per well and cultivated overnight to reach approximately 50% confluency. Cells were then cultured in a serum-free medium containing 5 μg/mL polybrene (Zolgene, China) at 37°C and 5% CO2 and infected with recombinant lentiviral vectors at a multiplicity of infection (MOI) of 40. After 4–6 h, the medium was replaced with the fresh serum-containing medium. After 72 h of post-infection, cells were cultivated in the medium containing 0.5 μg/mL puromycin (Sigma) for 2 weeks to screen and obtain stable cell lines, which were then cultured in bulk for subsequent assays.

Total RNA was isolated from HaCaT cells using the TRNzol Universal REAGENT (TIANGEN, China). Reverse transcription and qRT-PCR were used to detect differences in the transcript levels of KRT17. cDNA was synthesized from the total RNA using FastKing gDNA Dispelling RT SuperMix (TIANGEN, China). qRT-PCR reactions were performed using BlasTaq™ 2X qPCR MasterMix (ABM, Canada) using cDNA as a template in a reaction system (20 μL): 10 μL BlasTaq™ 2X qPCR MasterMix, 0.5 μL forward primer (10 μmol/L), 0.5 μL reverse primer (10 μmol/L), 2 μL cDNA, and 7 μL RNase-Free ddH2O for 40 cycles. qRT-PCR was performed using the ABI 7500 Real-Time PCR system. The PCR primers were synthesized by Fuzhou Shangya Biotechnology Co., Ltd. The specific sequences were as follows: KRT17, forward 5-GGT​GGG​TGG​TGA​GAT​CAA​TGT-3′ and reverse 5′-CGC​GGT​TCA​GTT​CCT​CTG​TC-3′ and β-actin, forward 5-TCA​CCC​ACA​CTG​TGC​CCA​TCT​ACG​A-3′ and reverse 5'-CAG​CGG​AAC​CGC​TCA​TTG​CCA​ATG​G-3 '. The relative expression levels of KRT17 were calculated by the 2^−ΔΔCT method and normalized to β-actin.

Total protein from HaCaT cells was cultured, lysed (RIPA lysis buffer), and extracted, and the protein expression was detected. Protein samples and marker were loaded into electrophoresis gel wells, wet-transferred to polyvinylidene fluoride (PVDF) membranes after electrophoresis, and soaked in 5% albumin from bovine serum (BSA) at room temperature for 2 h. The membranes were washed once, incubated using primary antibody solutions (KRT17 [ab109725, Abcam, United Kingdom], 1:1,000; Notch1 [ab52627, Abcam, United Kingdom], 1:1,000; Notch2 [D76A6, CST, United States], 1:1,000; Notch3 [D11B8, CST, United States]; AKT [51077-1-AP, Proteintech, United States]; Phospho-AKT [Ser473] [28731-1-AP, Proteintech, United States], 1:1,000; PI3K p85 alpha [AF6241, Affinity, United States], 1:1,000; Phospho-PI3K [AF3242, Affinity, United States],1:1,000; and β-actin [ab8226, Abcam, United Kingdom], 1:1,000), and diluted in 5% bovine serum albumin (BSA) overnight at 4°C. After washing the membranes three times, 5% BSA was added to dilute the horseradish peroxidase-labeled secondary antibody, and the samples were incubated in a shaker at room temperature for 2 h. The sections were visualized after three washes with TBST. The target bands were finally detected using a Thermo ECL Kit (34080, Thermo Fisher). Tanon-5200Multi (Tanon) was used for exposure imaging. Protein expression was analyzed using ImageQuant (LAS-4000, Fujifilm) software. Bands were quantified using ImageJ software.

An 18-year-old male presented with multiple evident cysts of approximately 0.3–0.7 cm diameter over his face, neck, and scrotum, but without ulceration, bleeding, or other symptoms. The onset of skin lesions began after birth. When he was 4 months old, the fingernails and toenails thickened and gradually deformed, showing an opaque yellowish brown color on the deck. Five years ago, he suffered from repeated episodes of papules, nodules, pustules, and cysts on the back. After that, the skin lesions worsened with the development of red papules, nodules, and abscesses on the head and face, as well as bilaterally in the armpits, groin, and buttocks. The most severe involvement was in the axilla with pain, and the Hurley classification was II. His parents did not share a consanguineous relationship, and there were no other individuals in the family with similar symptoms.

He had a body mass index of 20.2. Skin examination revealed papules, pustules, and flake alopecia on the head (Figure 2A). Numerous cysts were present over his face (Figure 2B), neck (Figure 2C), and scrotum (Figure 2D). Multiple black comedones, inflammatory papules, and hyperpigmentation were seen on his back (Figure 2E). Nodules and cysts with discharged pus and scars were present over the axilla (Figure 2F) and inguinal folds (Figure 2D). The patient’s toenails and fingernails were markedly thickened. The toenails were severely affected, and the bunions of both feet were claw-like (Figures 2G, H).

Laboratory examinations showed that the white blood cell count was 11.60 × 10⁹/L (reference value 4.00 × 10⁹–10.00 × 10⁹/L), the percentage of neutrophils was 83.20% (reference value 50.00%–70.00%), and C-reactive protein was 10.43 mg/L (reference value 0–8.00 mg/L). No obvious abnormality was found in other examinations. The microscopic examination and culture of onychomycetes were negative. Nail dermoscopy showed thickened nails, transverse grooves and longitudinal ridges of toenails, and black punctate and linear structures of fingernails (Figures 2I, J). The skin tissue pathology revealed that the intradermal cyst structure is seen, the eosinophilic wavy epidermis can be seen on the cyst wall, and the endothelial cells are arranged in a serrated shape, indicating steatocystoma multiplex (SM) (Figures 2K, L).

The results of whole-exome sequencing indicated that the proband had a heterozygous missense variant in the KRT17 gene, which was a change of nucleotide 274 from adenine A to guanine G (c.274A>G), resulting in the change in the 92nd amino acid from asparagine to aspartic acid (p.N92D). The variant was not found in his parents or 100 unrelated healthy controls (Figure 3A). It was also not found in the gnomAD v4 database, the NHLBI Exome Sequencing Project, or the Exome Aggregation Consortium browser. This supports the idea that this is a de novo causative variant rather than a polymorphism. The results of short tandem repeat (STR) analysis indicated that 18 of the 24 STR loci in the proband were heterozygous, and two loci at the same locus were from his father and mother, supporting the biological relatedness of the proband to his parents.

This variant has been reported in ClinVar. According to the ACMG guidelines, this variant was preliminarily determined to be pathogenic (PS2 + PS4 + PM1 + PM2_Supporting + PM5 + PP3_Strong): 1) PS2: this de novo (both maternity and paternity confirmed) variant has been reported in a patient with the disease and no family history. 2) PS4: this variant has been reported in a large population where it was shown to co-segregate with pachyonychia congenita in multiple affected individuals (McLean et al., 1995). 3) PM1: the variant was located in the mutation hotspot region (helix initiation motif). 4) PM2_Supporting: this variant was absent from controls in Exome Sequencing Project, 1,000 Genomes Project, gnomAD database, or Exome Aggregation Consortium. 5) PM5: many other pathogenic variants in patients with pachyonychia congenita have been reported at the same residues (c.275A>C,p.Asn92Thr; c.275A>G,p.Asn92Ser; and c.274A>C,p.Asn92His) according to the Human Gene Mutation Database (Ofaiche et al., 2014; Zhang et al., 2020; Smith et al., 1997). 6) PP3_Strong: the prediction results of bioinformatics protein function comprehensive prediction software REVEL, SIFT, PolyPhen_2, MutationTaster, and GERP+ indicated harm. We searched the relative protein position of the patient’s variant in the gene bank and found that Asn92 is a conserved amino acid residue in Homo sapiens, Pan troglodytes, Macaca mulatta, Canis lupus familiaris, Mus musculus, Rattus norvegicus, Bos taurus, and Danio rerio (Figure 3E). The 3D structure analysis of KRT17 showed that in the wild-type, Asn92 formed hydrogen bond interaction with amino acids Met88, Gln89, and Ala96. When Asn was replaced by Asp, the amino acid at position 92 changed from neutral to negative, and the hydrogen bond with Gln89 disappeared, forming a new hydrogen bond interaction with Leu95 (Figures 3B–D). This may have affected the structure and function of the protein.

After 72 h of infection with ZVE1003 cells grown in the logarithmic phase with pCDH lentiviral empty load, pLVZG-CMV-KRT17 (WT)-3xflag-EF1-copGFP-T2A-Puro, and pLVZG-CMV-KRT17 (c.274A>G)-3xflag-EF1-copGFP-T2A-Puro lentiviral expression vectors, the cells highly expressed green fluorescent protein (Figure 4A), indicating good growth status. Relative KRT17 mRNA expression levels of the KRT17-WT and c.274A>G mutant of the KRT17 gene in HaCaT cells were examined via qRT-PCR. Green fluorescent protein was used as an external reference to compare the relative differences in the KRT17 mRNA expression group, and the results indicated that KRT17 mRNA expression was downregulated in c.274A>G mutants cells compared to that in control group (P < 0.05) (Figure 4B).

Compared with the negative control group, the protein levels of KRT17-WT and KRT17-c.274A>G transfected groups increased, which was statistically significant (P < 0.01), and there was no difference between the two transfected groups (Figure 5A). The transfection of the KRT17 variant-encoding vector resulted in a significant reduction in NOTCH activity compared with the wild-type vector (P < 0.0001) (Figure 5B), indicating that the KRT17-induced variant resulted in a reduction in NOTCH activity. p-PI3K and p-AKT are the phosphorylation products of PI3K and AKT, respectively. In HaCaT cells transfected with plasmids, the protein levels of AKT were not significantly different from those of the wild-type group, and the protein levels of p-PI3K and p-AKT were not significantly different from those of the wild-type group (P > 0.05) (Figure 5C), indicating that the PI3K–AKT signaling pathway was not activated.