This research project was conducted after receiving approval from the Institutional Review Board (IRB) at the Dermatology Institute of Jiangxi Province, The Affiliated Dermatology Hospital of Nanchang University (Nanchang, China), and is in accordance with the Helsinki Declaration. All participants, including their parents or legal guardians, provided written informed consent. Use of the clinical picture of the proband was not permitted.

A total of 46 individuals from a DEB-Pr family were recruited for a comprehensive questionnaire and physical examination in this study at the Dermatology Institute of Jiangxi Province, The Affiliated Dermatology Hospital of Nanchang University between July 2020 and March 2021. Among them, 18 individuals donated their blood, 12 individuals agreed to the use of their samples for WES and all donors agreed to the use of their genomic material for COL7A1 PCR amplification. In addition, in order to have non-DEB controls, blood was also collected from 50 healthy controls who visited hospitals for annual physical examination and had a confirmed lack of DEB-associated genetic disease in the family history (median age 45.5 years; age range, 26-76 years; 32% women), and blood from one 32-year-old pregnant woman at week 37 who also presented as healthy and with a lack of DEB-related disease was collected in order to test whether cffDNA could be used for COL7A1 PCR amplification.

The DNA was extracted by DNA isolation kit (cat. no. DC111; Vazyme Biotech Co., Ltd.) and DNA was measured by Nanodrop2000 (Thermo Fisher Scientific, Inc.). Genomic DNA (1 µg) was sheared to 150-250 bp using a Covaris LE220 instrument with the following parameters: Duty cycle, 20%; intensity, 5; cycles per burst, 200; time, 90 sec; and shearing tubes, Crimp-Cap microtubes with AFA fibers (Covaris Inc.). The DNA fragments were then selected by VAHTSTM DNA Clean Beads (cat. no. N411; Vazyme Biotech Co., Ltd.). The library was constructed using the MGIEasy Exome FS Library Prep Set kit v2.0 (cat. no. 1000009658; MGI Tech Co., Ltd.) following the manufacturer's instructions. Briefly, the end-repair for DNA fragments was performed by adding an ‘A’ nucleotide to the 3' end of each strand. Adapters were then ligated to both ends of the end-repaired/dA-tailed DNA fragments for amplification and sequencing. Size-selected DNA fragments were amplified by ligation-mediated PCR, purified and hybridized to the exome array for enrichment. Non-hybridized fragments were then washed out. Finally, the captured products were then circularized by MGIEasy circularization commercial kit (cat. no. 1000005259; MGI Tech Co., Ltd.). Rolling circle amplification was performed to produce DNA Nanoballs. The library was sequenced on the BGISEQ-500 sequencing platform (MGI Tech Co., Ltd.) and the overall quality control data is presented in Table SI. The raw sequence data reported in this paper were deposited in the National Institute of Biotechnology Information (NCBI) Sequence Read Archive database (accession no. SAMN36949954-SAMN36949963) and are accessible through https://www.ncbi.nlm.nih.gov/sra/PRJNA1000907.

The reads were mapped to the hg19/GRCh37 human reference genome by the BWA.v0.7.12 tool and Picard was used to mark the duplicates. A Genome Analysis Toolkit v 3.7 and Samtools (http://www.htslib.org/) were used to detect single nucleotide variants and short insertions/deletions (Indels). Short Indels in the repeat regions and within the 10 bp range from the start and end of the read were also excluded. The remaining variants were filtered from public databases comprising 1000 Genomes (ftp://ftp-trace.ncbi.nih.gov/1000genomes/ftp/release) and gnomAD (The Genome Aggregation Database; https://gnomad.broadinstitute.org/).

The variants were annotated with the ANNOVAR program (https://annovar.openbioinformatics.org/en/latest/). The in-silico analysis was performed by Sorting Intolerant From Tolerant (SIFT) (https://sift.bii.a-star.edu.sg/), Polymorphism Phenotyping v2 (PolyPhen2) (http://genetics.bwh.harvard.edu/pph2/), mutation assessor (http://mutationassessor.org/r3/), mutation taster (https://www.mutationtaster.org/), FATHMM (http://fathmm.biocompute.org.uk/) and CADD (https://cadd.gs.washington.edu/score) to anticipate the functional effect of missense and nonsense variants. Mega version 11 (11.0.13), Jalview Version 2 (2.11.2.7) and BioEdit software (v7.2) were used to align the sequences against the COL7A1 reference sequence (accession no. NC_000003).

ELISAs for desmogelin (Dsg)1 (cat. no. 7880E), Dsg3 (cat. no. 7885E), BP180 (cat. no. 7695E), BP230 (cat. no. 7613E) and Type VII collagen (cat. no. 7845R2) were performed using commercial kits purchased from Medical and Biological laboratories, Co., Ltd., and the procedures were performed according to the manufacturer's instructions.

For immunoblotting, recombinant proteins including LM332 (cat. no. LN332-0502, Biolamina AB), LM111 (cat. no. LN111-0501, Biolamina AB) and integrin α6β4 (cat. no. 5497-A6-050; BD Biosciences) were used for the antigen source (100 ng per reaction), and final blots were incubated with patients' sera at 1:5-1:800 dilution. The detailed method was performed as previously described (17-19).

Indirect IF studies of normal human skin and 1M NaCl-split normal human skin were performed as previously described (19).

Both cfDNA and cffDNA were extracted from plasma using a QIAamp DNA Blood Mini extraction kit (Qiagen GmbH) according to the manufacturer's instructions. The concentration of the DNA was measured by a Nanodrop 2000 (Thermo Fisher Scientific, Inc.) and the quality of the isolated DNA was checked by 1% agarose gel electrophoresis.

The candidate variants were validated by direct Sanger sequencing. Primers for PCR and sequencing were provided by Tsingke Biological Technology. For PCR amplification, 10 ng of total genomic DNA was used as a template in 20 µl of the reaction mixture using Vazyme Taq DNA polymerase as previously described (20). The primers used in this study and listed in Table SII. The thermocycling conditions were as follows: 95˚C for 3 min, followed by 35 cycles of 95˚C for 15 sec, 61˚C for 15 sec and 72˚C for 30 sec, and a final extension at 72˚C for 3 min. The PCR products were checked by DNA electrophoresis and sent to Tsignke Biological Technology for sequencing. The COL7A1 consensus sequence generated from sanger sequencing results from 50 healthy controls, the reference sequence from NC_000003 and the sequences from affected individuals were also aligned, as shown in Fig. S1.

Three-dimensional structure prediction of selected proteins was performed using I-TASSER (https://zhanggroup.org/I-TASSER/) (21). In details, the complete gene sequence of COL7A1 was retrieved from the NCBI GenBank (https://www.ncbi.nlm.nih.gov/gene/1294) under accession no. NC_000003. The wild-type protein of COL7A1 with an exonic region of 70 to 118 was used for 3D structure prediction. The C-score=-1.33, TM-score=0.55±0.15 and root-mean-square deviation (RMSD)=12.2±4.4 Å were selected, where C-score is a confidence score for estimating the quality of predicted models by I-TASSER and TM-score is a metric for assessing the topological similarity of protein structures; it is designed to solve two major problems in traditional metrics such as RMSD.

A review of the literature was performed using PubMed and Google search with the search terms ‘epidermolysis bullosa pruriginosa’, ‘DEB pruriginosa’, ‘EB pruriginosa’ and ‘COL7A1’ covering the period of 1994 to November 2022, and the systematic reviews were limited to studies published in English.

A 50-year-old female proband presented to the Dermatology Hospital of Jiangxi Province (Nanchang, China) with complaints of itching, skin nodules and blackish discoloration of the skin on both lower limbs for >32 years. Upon physical examination, depigmentation at the center with scarring was seen over the larger nodules and multiple lichenified papules to nodules were present over the lower limbs extending from the knee to the ankle joint. Routine investigations, including chest X-ray, computed tomography, electrocardiogram, upper and lower gastrointestinal endoscopy, complete blood count, liver and renal function tests, as well as serum IgE, ferritin and thyroid function tests, all showed normal results. It was further confirmed that direct immunofluorescence using a biopsy of lesioned skin of the left leg showed no IgG, IgA, IgM and complement C3 deposition at the basement membrane zone or on keratinocyte cell surfaces. This patient's serum showed negative IgG, IgA and IgM results on indirect immunofluorescence using normal human skin and indirect immunofluorescence using salt-split skin (22,23). According to ELISAs using commercially available kits (Medical and Biological Laboratories, Co., Ltd.), IgG autoantibodies against Dsg1, Dsg3, BP180, BP230 and COL7 were all negative. Other diagnostic tests, including immunoblotting of normal human epidermal extract, dermal extract and recombinant proteins of laminin 332, laminin 111 and integrin α6β4 (24-26), which are routinely used in our laboratory for the detection of known autoimmune bullous disease (AIBD) autoantibodies (18,27), showed negative results for this patient's serum. Based on all of these clinical and laboratory data, the diagnosis of AIBD was ruled out for this patient.

The proband and her family under investigation comprised 46 individuals and a questionnaire revealed the absence of consanguineous marriage. Among the first three generations, 14 individuals had similar mild clinical disease phenotypes, as indicated in the pedigree chart (Fig. 1A). Based on the clinical examinations and questionnaires completed for the proband's family, the diagnosis of DEB-Pr was suspected.

To thoroughly investigate the potential genetic basis of DEB-Pr, WES of 12 individuals from the proband's family was performed, including the proband, affected individuals and unaffected individuals (Fig. 1B and C). Analysis of the WES data led to the identification of a G-to-A transition at the first base of codon 6,859 in exon 87 of the COL7A1 gene, resulting in the substitution of Gly (GGA) with Arg (AGA) (Fig. 1C). This variant (c.6859 G>A) in the COL7A1 gene has been reported in the DEB-Pr by Japanese studies (28-30). To predict the potential impact of the variant on the protein functions, different bioinformatic analysis tools were used and returned with predicted values, including the SIFT (value: 0.0), PolyPhen2 (value: 0.998), mutation assessor (value: 4.32), mutation taster (value: 0.999), FATHMM (value: -6.29) and CADD (value: 6.70). Overall, the variant was predicted to be ‘deleterious’ with high confidence, suggesting that COL7A1 mutation is the genetic basis of DEB-Pr.

To confirm the results of the WES analysis and explore the potential use of cfDNA for DEB-Pr diagnosis, target DNA amplification and Sanger sequencing were next conducted. The dominant peaks observed in the cfDNA samples were ~166 and 300 bp, which aligns with the size range reported in previous studies (31,32) (Fig. 2A). The targeted region in exon 87 of COL7A1 was amplified and Sanger sequencing was performed in patients and 50 ethnically matched healthy controls (Fig. 2B-D). The results showed that the affected individuals carried AGA, whereas the unaffected individuals and healthy controls were GGA at this position, further supporting the association between genotype and phenotype. Of note, one affected individual (III-9) carried heterozygous variants according to both WES (Fig. 1C) and Sanger sequencing (Fig. 2D). However, clinical examination indicated that this individual did not show any symptoms of DEB-Pr, which is consistent with other reports, suggesting other factors may also have an important role in disease progression, such as metabolic, genetic, epigenetic or environmental factors (33,34), and follow-up of this individual may be recommended. To further determine whether the cfDNA may be used as a template to amplify other regions, pairs of primers for 12 out of 118 exons were randomly selected. The results indicated that all amplicons may be amplified with target sizes, implying that cfDNA may be used for genetic testing (Fig. 2E). To further extend its potential for prenatal diagnosis, cffDNA was subsequently isolated from a pregnant woman (Fig. 2F) and amplification of the COL7A1 gene were performed. Successful amplification of the COL7A1 gene from cffDNA suggested that cffDNA may be used as a non-invasive means of detecting the COL7A1 variants in suspected hereditary epidermolysis bullosa (Fig. 2G).

Multi-alignments analysis indicated high evolutionary conservation of p2287 glycine and the Gly-X-Y motif across several vertebrates, including human, mouse, pig, chicken, dog, frog, zebrafish and lizard, suggesting critical roles of the amino acid in maintaining normal protein function (Fig. 3A). Next, the possible molecular mechanism of the variant on the COL7A1 gene causing DEB-Pr was investigated using the structure prediction tool I-TASSER, which is considered the finest tool for predicting the tertiary structure of a protein (https://seq2fun.dcmb.med.umich.edu/I-TASSER/) (21). Analysis of the predicted 3D protein structures did not reveal any difference in the hydrogen bond interactions between wild-type and mutant residues. However, the residue size and charge differences may impact the protein folding and thus overall protein stability (Fig. 3B). To gain an overview of the COL7A1 mutations causing DEB-Pr, a total of 65 unique mutations in COL7A1gene were summarized, of which 37 were specifically reported in DEB-Pr cases (Fig. 3C). This summary demonstrated that the major cause of DEB-Pr involves glycine substitutions (86.5%), although deletions or splice-site mutations may appear in certain cases, consistent with previous findings (2,6). Notably, all variants are located within the triple helix of COL7A1 in DDEB. By contrast, the mutations include nonsense, splice site, exon skipping and non-glycine missense mutations within the triple helix or non-collagenous NC-2 domain in RDEB.