Study participants and sample collection

This study, including the clinical examination and gingival tissue biopsies, was approved and performed in accordance with the ethical committee standards at the Southern Medical University, China. Patients and relatives analyzed in the project signed informed consent prior to inclusion. All subjects from the same family with non-syndromic HGF and six random subjects (control samples) were recruited from the Department of Stomatology at Nanfang Hospital. The patients’ medical records and medication histories were collected by a skilled clinician. In addition, eight unaffected individuals from the family underwent oral and clinical examinations.

Peripheral venous blood was collected from sixteen family members. Gingival biopsies from the proband with non-syndromic HGF was collected during gingivectomy, which was carried out to excise excessive abnormal proliferation of the coronal/marginal part of the gingiva. Control gingival biopsies were collected from six healthy random volunteers, aged 20–30 years, due to wisdom tooth extraction. After collection, the gingival tissues were immediately placed into DMEM/F12 (10% FBS) and rinsed with sterile phosphate-buffered saline (PBS) three times. Part of the tissue was fixed with 4% paraformaldehyde for 24 h and then embedded in paraffin. A second sample was used for the isolation and culture of gingival fibroblasts, and a third portion was applied to RNA/protein extraction and detection.

Mutation screening and bioinformatics

Array-CGH was performed to determine whether chromosomal aberrations were present in the proband. To identify the pathogenic variant responsible for HGF, we performed whole-exome sequencing with genomic DNA for affected individuals II-1 and III-14 and the control II-5 in collaboration with Novogene-Beijing (China). We discovered two significant mutations located in KIF3C and ZNF513 on chromosome 2 in two patients in the family. Short tandem repeat markers (D2S2168, D2S144, D2S2223, D2S174, D2S2247, D2S365, and D2S170), which flank KIF3C and ZNF513, were selected for haplotype analysis. Co-segregation analysis in 16 family members with Sanger sequencing was used to the inheritance of these two mutations.

To provide further insights into KIF3C and ZNF513 functionality, tertiary structures of the wild-type and mutant protein were predicted using ITASSER and Pymol. SIFT was used to predict the harmfulness of the mutations. ClustalX 2.0 and GeneDoc were used for multiple sequence alignment based on the NCBI database to analyze the conservation of affected amino acids across species. The two mutations were assessed with high-resolution melting (HRM) analysis in 300 unrelated control subjects with matched geographical ancestry.

Plasmid construction, virus packaging, cell infection and transfection

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-Cas9, an adaptive immune system present in a variety of microorganisms, is a self-defense mechanism for bacteria to resist foreign pathogens. The CRISPR system, including non-coding RNA elements and enzymes, can recognize characteristic sequences and cleave foreign nucleic acids and can be used in eukaryotic cells for fast and effective gene editing. The CRISPR type II system is composed of the Cas9 protein and a single-guide RNA (sgRNA). By recognizing protospacer adjacent motifs (PAMs), sgRNAs pair complementarily with target DNA sequences. The Cas9 protein cleaves the sequence under the guidance of the sgRNA, resulting in a double-strand break, resulting in gene editing.34,35

According to the predicted score and specificity, we designed and synthesized three sgRNAs for KIF3C and ZNF513 with a web tool (http://crispr.mit.edu). These were annealed to form double strands. The sgRNA annealing products were ligated with the expression vector (lentiCRISPR v52961), digested by BsmBI, and then CRISPR recombinant plasmids were obtained by competent cell transformation, monoclonal screening, and plasmid extraction. HEK-293T cells were inoculated on 10-cm petri dishes 1 day in advance, and plasmids transfection was carried out when the cells grew to 70%. LentiCRISPR v52961 recombinant plasmid, PSPAX2, and PMD2G in each petri dish were allocated in a ratio of 7.5 μg:7.5 μg:5 μg, while the ratio of PEI to total plasmid was 3 μg:1 μg. Plasmids and PEI were mixed with DMEM separately and then mixed together and added to the cells. After 8 h culturing, the supernatant was discarded, and 10 mL DMEM (with 10% FBS) was added. After 24 h and 48 h of culturing, 4 mL of complete medium was added. After 72 h of culturing, the supernatant of the cells was collected, and the virus suspension was obtained after ultracentrifugation for 1 h at 4 °C at 20 000 r·min−1, and re-suspension with DMEM.

Normal human gingival fibroblasts (NHGFs) were inoculated on 24-well plates in advance, and the cells grew to 70% confluence to be infected with the virus. The old medium was discarded, 500 μL complete medium containing 1/1 000 polybrence was added, and then 30 μL virus suspension was added. After 24 h of culturing, puromycin (final concentration 1–3 μg·mL−1) was added for screening. After continuous screening for 7 d, the infected cells were expanded, and PCR was employed to identify whether the gene was knocked out successfully. With flow cytometry, a single cell was selected from the cell suspension and inoculated in a 96-well plate, and the clone growth was observed after about 10 d of culture. After the cells were further expanded and cultured, DNA, RNA, and protein were extracted to verify whether KIF3C or ZNF513 was knocked out completely.

The reference and mutant sequences of KIF3C and ZNF513 were cloned into the lentivirus overexpression vector (pLenti-Bi-cistronic). The detailed steps of virus packaging and cell infection were the same as above.

In order to confirm the effect of two variants on NHGFs cells from another aspect, we cloned the sequences of KIF3C and ZNF513 into expression vectors, pcDNA3.1flag and pEGFPN1, respectively. After being constructed, pcDNA3.1flag-(m)KIF3C and pEGFPN1-(m)ZNF513 overexpression plasmids were transiently transfect into NHGFs using Lipofectamine 2000 (Invitrogen) in Opti-MEM. 6–8 h after transfection, cells were cultured in fresh complete medium until they can be accessed for different purpose.

Generation and genotyping of C57BL/6-Kif3c R412H and C57BL/6-Zfp513 R250W knock-in mice

Human ZNF513 corresponds to Zfp513 in mice. A C57BL/6 mouse model with a point mutation (R412H) at the Kif3c locus and (R250W) at the Zfp513 locus was generated by Saiye (Suzhou) Biological Technology Co., Ltd. with CRISPR/Cas9-mediated genome engineering. Briefly, the R412H (CGT to CAT) mutation in the NM_008445.2 transcript of Kif3c or the R250W (CGG to TGG) mutation in the NM_175311.4 transcript of Zfp513 in the donor oligo was introduced into each exon by homology-directed repair. Cas9 mRNA, gRNA, generated by in vitro transcription, and donor oligos were co-injected into fertilized eggs for knock-in mouse production. F0 positive mice were selected and confirmed by PCR and DNA sequencing (Supplementary Fig. 7a, b).

The mouse Kif3c is located on mouse chromosome 12, while Zfp513 is located on chromosome 5. Accordingly, six genotypes, Kif3cR412H/R412H/Zfp513+/+, Kif3c+/+/Zfp513R250W/R250W, Kif3c+/R412H/Zfp513+/R250W, Kif3cR412H/R412H/Zfp513+/R250W, Kif3c+/R412H/Zfp513R250W/R250W, and Kif3cR412H/R412H/Zfp513R250W/R250W, of mice were generated from the F1 Kif3c+/R412H/Zfp513+/+ and Kif3c+/+/Zfp513+/R250W self-crossed inbred lines. Mouse DNA was isolated from tail biopsy samples and genotyped by PCR followed by sequence analysis (Supplementary Fig. 7c). All experiments involving mice were conducted in accordance with the Institutional Animal Care and Use Committee of Southern Medical University approved guidelines.

Separation and processing of the murine mandible and peeling the maxillary gingiva

After the mice were sacrificed using approved guidance, we cut both sides of the oral cavity with a straight scissors, pulled down the murine mandible, excised the mandible including mandibular molar, and then trimmed away the irrelevant tissue. The separated mandible was fixed with 4% paraformaldehyde for 24 h and placed in decalcifying fluid for 3–4 weeks (replaced once every one week). The mouse mandibular first molar was cut on the coronal plane and then embedded in paraffin for serial sections and hematoxylin and eosin (HE) staining, as mentioned below.

The scissors were directed perpendicular to the plane of the palatal bone, and we incised tissues 1 mm behind the third molars and 2 mm behind the incisors. Then, we incised the vestibule until the posterior of both sides of the maxilla. We separated the maxilla from the rest of the skull, cut the middle suture, and trimmed the palatal tissue.36 The maxillary gingiva was stripped from the anterior border with tissue forceps for subsequent experiments. One part of the excised gingiva was used for the culture of gingival fibroblasts, and the second part was applied for RNA/protein extraction and detection.

Primary human or murine gingival fibroblasts isolation, culture, and identification

In brief, rinsed human gingival tissues were trimmed into 1 mm × 2 mm squares with scissor in PBS (2% penicillin/streptomycin), and then these squares were digested in dispase II (2 mg·mL−1) at 4 °C for 16–18 h. The obtained subepithelial tissues were plated in DMEM (low glucose) with 20% FBS and 1% penicillin/streptomycin and cultivated for 5–7 days with the tissue adherent method. Identification of the gingival fibroblasts were confirmed by immunofluorescent staining against vimentin (positive) and cytokeratin (CK19, negative).

Separated mouse maxillary gingival tissues were cut into pieces in a tissue culture dish (35 × 10 mm2) with 1 mL PBS, 2% FBS, 2% penicillin/streptomycin, Collagenase Type IV (2 mg·mL−1), and DNAse Type I (1 mg·mL−1) and were incubated for 20 min at 37 °C at 200 r·min−1 in a shaker incubator.36 Next, 20 μL EDTA 0.5 M was added, and the tissues were incubated for another 10 min at 37 °C at 200 r·min−1. The centrifuged cell pellets were plated in 1640 (20% FBS, 1% penicillin/streptomycin) and cultivated for 4–6 days.

Chromatin immunoprecipitation (ChIP)

ChIP is a method to study the interaction between proteins and DNA in living tissues or cells at the whole-genome level. According to the protocol of the EZ ChIPTM Chromatin Immunoprecipitation Kit (Millipore), we collected 2 × 107 NHGFs or patient gingival tumor fibroblasts, followed by formaldehyde cross-linking and ultrasonic fragmentation, pre-cleaning, and antibody incubation, immune complex precipitation and cleaning, purification, and recovery of the DNA samples. The parameters for ultrasonic breaking of DNA by the ultrasonic processor M220 (Covaris) were as follows: peak incident power: 65 W, sample volume: 60 μL, treatment time: 20 s on and off, repeats: 9, and total time: 6 min. Immunoprecipitation antibodies included a positive control (anti-RNA polymerase), negative control (rabbit IgG), and target antibody (ZNF513).

The obtained DNA products were used for qPCR to detect gene expression levels. After ChIP-qPCR, the value of %Input was calculated by the following specific formula (%Input = 2((Cq(IN)−Log2(DF))−Cq(IP)) × 10037; IN: Input; IP: Immunoprecipitated; DF: dilution factor), and agarose electrophoresis was employed to detect the qPCR products. Additionally, DNA products were used for PCR, whose primers were the same as the qPCR primers. After TA cloning with PCR products, positive clones were selected and identified as genes to be tested with Sanger sequencing.

Histopathology and immunohistofluorescence staining

Human gingival biopsies were fixed with 4% paraformaldehyde for 24 h, then dehydrated in a series of alcohols, and finally embed in paraffin. Tissues were serially cut into 2-μm thick sections and stained with HE and Masson. In brief, paraffin gingiva sections were heated at 65 °C for 30 min and kept in xylene for deparaffinization. Next, the tissues were rehydrated in graded alcohols and stained for 10–15 min with hematoxylin, followed by differentiating in 1% hydrochloric acid alcohol for 5 s. 1% eosin was used for a 3 min staining, then the tissues were further dehydrated in graded alcohols and kept in xylene. Masson staining was performed with the trichromatic staining kit. After heating, deparaffinization, and rehydrating in xylene and graded alcohols, paraffin sections of the gingival tissues were stained for 5 min with Masson compound solution. In addition, 5 min phosphomolybdic acid, 5 min aniline blue staining, and 1 min differentiation liquid were used sequentially. After HE and Masson staining, a light microscope was used to carry out histopathological evaluation of human gingival tissues after mounting with neutral balsam.

After deparaffinization, rehydrating, and antigen retrieval, the paraffin sections were blocked with PBS containing 3% bovine serum albumin for 30 min. Polyclonal rabbit anti-KIF3C antibody (1:100, GeneTex) and polyclonal rabbit anti-ZNF513 antibody (1:100, Sigma) were used to incubate tissues at 4 °C overnight. Then, goat anti-rabbit IgG antibody (DyLight594, 1:200, GeneTex) was incubated for 1 h at room temperature away from light. Finally, the tissues were viewed using LSM 880 with Airyscan (Carl Zeiss, Jena, Germany) after mounting with DAPI medium.

Cell proliferation

EdU is a thymidine analogue, and its alkyne group is rare in natural compounds. It can replace thymine (T) to infiltrate into synthetic DNA molecules during DNA replication. Based on the specific reaction of Apollo fluorescent dye with EdU, the replication activity of DNA can be detected directly and accurately. Stable gene overexpression cells, human or murine gingival fibroblasts, in the logarithmic growth phase were inoculated in a 96-well plate and cultured at 37 °C for 24 h. According to the instructions of the Cell-Light EdU Apollo567 In Vitro Kit (Ribobio), EdU labeling, cell fixation, Apollo staining, and DNA staining were performed sequentially. Finally, the positive cells were viewed under a fluorescence microscope and counted with ImageJ software.

Human gingival fibroblasts and stable gene overexpression cells were inoculated in 96-well plates at the same cell density. Transfected NHGFs were transferred at the same initial concentration and cultured in 96-well plates 24 h after transfection. After 12, 24, 36, 48 or 60 h of culturing, 100 μl of the culture medium containing 10 μL Cell Counting Kit-8 (CCK-8, UE) solution was added per well, and cells were incubated at 37 °C for 1.5 h. Living cells were evaluated with a microplate reader at 450 nm.

Migration assay

The Transwell migration assay was used to quantify cellular migration. Human or murine gingival fibroblasts cultured to the third generation were prepared in 300 μL cell suspensions (2.5 × 105 cells in serum-free medium). The migration chambers with a diameter of 6.5 mm and an aperture of 8 μm were placed in a 24-well plate, then 750 µL culture medium (with 10% FBS) was added to the lower chamber, followed by 300 µL cell suspension, and the plate was incubated at 37 °C for 24 h. After incubation, the medium was removed and the chamber was washed twice in PBS. The chamber was stained with 500 μL 0.1% crystal violet at room temperature for 20 min. The crystal violet was removed, and the chamber was washed twice by PBS, then we scraped off nonmigrated cells with cotton swabs, and finally observed the migrated cells under a light microscope. In addition, the stained chamber was placed in 400 μL 33% acetic acid solution shaking at room temperature for 10 min. 200 μL of the eluent was transferred into a 96-well plate, and the absorbance was measured with a microplate reader at 560 nm. The migration rate was calculated according to the cell density standard curve (Supplementary Fig. 4a).

Primer design and detection of gene expression

The main primer sequences involved in this paper are shown in Table S2. The expression of genes, including KIF3C, ZNF513, COL1A1, FN1, SOS1, KCNQ1, PIK3CA, and PIK3CB, was detected by qPCR or western blotting.