2.1. Subjects

A patient was admitted to the Chinese PLA General Hospital with unexplained recurrent pneumothorax and multiple diffuse cystic changes in both lungs. Combined with clinical signs and symptoms, there was a strong clinical suspicion that the onset of the disease was related to BHD. And then next-generation sequencing (NGS) identified an FLCN variation of uncertain significance. Subsequently, Sanger sequencing of the remaining four members of the entire family resulted in the detection of an FLCN mutation at the same locus in one additional member of the family. Genetic diagnosis met with the diagnostic criteria for Birt–Hogg–Dubé due to FLCN mutations. All the patients were clinically assessed by at least two experienced clinicians to clarify the diagnosis and understand the genetic background. All participants collected in this study provided informed consent for the study approved by the ethics committee of the Chinese PLA General Hospital. Patients highly suspected of BHD underwent lung CT scanning, renal MRI, and skin examination.

2.2. Diagnostic Criteria for BHD

Primary diagnostic criteria were as follows: (1) the presence of more than five fibrofolliculomas or trichiliocosm in adults, of which at least one has been confirmed pathologically, and (2) the presence of a germline mutation in the FLCN gene that causes the disease.

Secondary diagnostic criteria were as follows: (1) multiple pulmonary cysts, bilateral occurrence of lesions located at the base of the lungs with no other clear etiology, with or without a history of primary spontaneous pneumothorax; (2) renal tumors: relatively early age of onset (< 50 years old), bilateral or multiple renal tumors, or pathological type of mixed acidophilic cell carcinoma and pheochromocytoma; and (3) confirmed diagnosis of BHD in a first-degree relative. The diagnosis of BHD was made by fulfilling one major or two minor criteria.

2.3. Whole-Exome Sequencing Analysis

The genomic DNA was analyzed by target region capture high-throughput sequencing. The assay was performed on the Illumina sequencing platform. The data obtained were sequenced at an average depth of 90× for known exons and upstream and downstream 5 bp sequences in the human genome. The average sequencing depth of the sequences was 90×, and 98% of the target sequences were sequenced at a depth of 20× or more. Base identification was performed for all sequenced segments. The assay was developed and validated by metallic medicine. Sequencing data were analyzed using the GATK software package. The sequenced segments were compared with the UCSC hg19 reference gene group by BWA (Burrows–Wheeler Aligner). The sequenced segments were compared with the UCSC hg19 reference gene group by BWA. The test uses VEP (Variant Effect Predictor) software to annotate variants. The test was also based on genetic disease databases such as ClinVar, OMIM, HGMD, and gnomAD; variant databases; and human population large-scale sequencing databases. Variants were screened based on genetic disease databases such as ClinVar, OMIM, HGMD, and gnomAD; variant databases; and human-scale sequencing databases.

2.4. Construction of FLCN-wt and FLCN-mut Expression Vectors

The whole gene-synthesized FLCN CDS (coding sequence) was used as the template, and Phage-FLCN-BamHI-F/Phage-FLCN-NotI-R and Phage-FLCN-BamHI-mut-F/Phage-FLCN-NotI-R were used as the primers to amplify to obtain the fragments of BamHI-wt-NotI and BamHI-mut-NotI, which were inserted into the phage vector after double enzyme digestion of BamHI and NotI, and then the Phage-wt and Phage-mut vectors were obtained, and the sequence verification was performed. The primer sequence for vector construction is listed in Table 1.

2.5. Cell Culture and Transfection

The 293 T and BEAS-2B cells purchased from Priscilla (Wuhan, Hubei Province, China) were cultured with DMEM (Gbico, Carlsbad, CA, United States) medium containing 10% fetal bovine serum (FBS) (Yeasen, Shanghai, China) and 0.5% penicillin/streptomycin (Yeasen, Shanghai, China). The cells were grown in six-well plates until they reached 70%–80% confluency, and the constructed recombinant expression vectors were transfected into the 293 T cells or BEAS-2B cells instantaneously using Opti-MEM (Lonza, Bassel, Switzerland) and Lipofectamine 3000 (Carlsbad, CA, United States) transfection reagent. After transfection for 48 h, the samples were collected for qPCR and Western blot (WB) detection, respectively.

2.6. NMD Detection by Cell Transfection of FLCN Mutation

2.7. NMD Detection by Cells Knock-in FLCN Mutation With CRISPR/Cas9

2.8. RT-PCR

Cell samples were collected after transfection with wild and mutant eukaryotic recombinant expression vectors for 48 h. Total RNA was extracted by the TRIzol (TaKaRa, Otsu, Shiga, Japan) method, and cDNA was synthesized after DNA digestion. The expression levels of FLCN were detected by qPCR.

Eukaryotic recombinant expression vectors of wt and mut were transfected into 293 T cells, and then cell samples were collected after CHX treatment or RNA si-UPF1 interference. Total RNA was routinely extracted using the TRIzol method, and cDNA synthesis was performed after DNA removal by digestion. The expression levels of UPF1 and FLCN were detected by qPCR.

FLCN-KI cells were collected from the control and editing groups, and the total RNA was extracted by TRIzol and then reversed into cDNA. The expression levels of target genes in the control group and editing group were detected by RT-qPCR. The primer sequence for RT-qPCR is listed in Table 5.

2.9. WB

Cell precipitation was collected after transfection with eukaryotic recombinant expression vectors for 48 h. The total protein was extracted by radioimmunoprecipitation assay (RIPA) lysate (containing protease/phosphatase inhibitors), the protein concentration was determined by the bovine serum albumin (BSA) kit, and then the protein was denatured. The same amount of total protein was taken for sodium dodecyl polyacrylamide gel (SDS-PAGE) electrophoresis, and the expression levels of wild-type and mutant target proteins were detected by WB.

2.10. Immunohistochemistry

Both one Birt–Hogg–Dubé patient and 15 non-Birt–Hogg–Dubé patients were obtained from the department of pathology of Chinese PLA General Hospital. Specimens of lung tissue of Birt–Hogg–Dubé patients were obtained from the proband, and non-Birt–Hogg–Dubé patients with pulmonary herniation were from patients with lung CT suggestive of typical pulmonary herniation, did not meet the clinical diagnostic criteria for Birt–Hogg–Dubé, and were pathologically diagnosed as simple pulmonary bulla. All lung tissues were fixed in 4% paraformaldehyde and made into 5 μm paraffin sections. Paraffin-embedded tissues were incubated at 72°C prior to dewaxing and rehydration. Antigen retrieval was achieved by placing sections in 0.01 M citric acid (pH 6) and microwaving for 15 min. Endogenous peroxidases were quenched in 15 mL of 3% hydrogen peroxide. Samples were washed again with phosphate-buffered saline (PBS) prior to treatment with 10% BSA blocking solution and incubated at 37°C for 40 min. Tissues were incubated in mouse anti-human FLCN, p-mTOR, p-SMAD3, and β-catenin monoclonal antibodies. The primary antibodies were diluted in TBST (tris buffer solution tween) at 4°C overnight. Samples were washed with PBS on the following day, incubated in a secondary antibody (1:1000), washed, and then treated using the 3,3 ^′-diaminobenzidine (DAB) peroxidase staining kit (Shanghai Biotechnology Company, NO132, Xuhui District, Shanghai City, China) as per the manufacturer’s protocol. For detection, a DAB peroxidase detection kit was used, and color development was monitored using an optical microscope. The development process was terminated by removing DAB and rinsing the sections with ddH₂O for 1 min prior to counterstaining with hematoxylin.

The intensity of immunohistochemical staining of sections of the same specimen was determined by two senior pathologists in a double-blind manner, and if the results were not uniform, the final intensity of staining was determined by repeated reading of the sections by agreement.

2.11. Statistical Analysis

All data were expressed as the mean ± SD (standard deviation), calculated using GraphPad Prism 9.0 software (La Jolla, CA, United States) through one-way analysis of variance (ANOVA), followed by the Dunnett test. The sequencing data was analyzed using SnapGene software. p values were represented in figures as follows:  ^∗p < 0.05,  ^∗∗p < 0.01, and  ^∗∗∗p < 0.001 were considered significant. Data are shown in column bars representing the mean ± SEM of at least three independent experiments.

3.1. The Process of Confirming the Diagnosis of Birt–Hogg–Dubé Patients

The patient was a 50-year-old male with multiple pulmonary cysts in both lungs since 2018 and a history of intermittent pneumothorax. The patient’s health check in August 2018 showed diffuse cystic changes in both lungs on CT examination without chest tightness and breathlessness, cough, and chest pain, and no special treatment was given. In May 2020, the patient had a sudden onset of right-sided chest pain with severe dyspnea after strenuous activity. Lung CT showed that the right pneumothorax with a compression area greater than 50% had diffuse cystic changes in both lungs. Closed chest drainage was given on an emergency basis, and the patient’s symptoms resolved. In August 2021, the patient again developed respiratory distress after being active, and the lung CT indicated right-sided pneumothorax with right-sided compressive atelectasis and diffuse cystic changes in both lungs. He was admitted to the hospital for further treatment.

History of past illness shows appendicitis surgery at Age 11 and closed chest drainage for spontaneous pneumothorax in 2018. The patients have lived in Beijing for a long time, with no history of living in infected areas, epidemics, or epidemic water; no history of living in pastoral areas, mines, high-fluoride areas, or low-iodine areas; no history of exposure to chemical substances, radioactive substances, or toxic substances; and no history of drug exposure. But the proband had a history of smoking for 10 years, about 20 cigarettes/day, and occasional alcohol consumption. He has an age-appropriate marriage, with his spouse and son healthy. His mother had a history of recurrent pneumothorax and died of cor pulmonale.

Physical examination is as follows: decreased breath sounds in the right lung, no dry moist rales were heard in both lungs, and regular heart rate. Multiple small papules were seen on the posterior neck (Figure 1(a)). Pathologic biopsy is as follows: skin grayish-white tissue, size 0.4 × 0.3 × 0.2 cm. Chronic inflammation of squamous epithelial mucosa, epidermal hyperplasia with hyperkeratosis, and significant proliferation of collagenous fibrous tissue seen in the dermis (Figure 1(b)), characteristics consistent with fibrofolliculomas. Pathological examination indicates dermatofibroma, known as acrochordon. Chest CT showed right pneumothorax and multiple diffuse cystic changes in both lungs (Figures 1(c) and 1(d)). Renal magnetic resonance plain scan and dynamic enhancement of renal indicated that the left renal cyst and dynamic enhancement scans showed no signs of abnormal enhancement. No obvious abnormality was found in the right kidney (Figures 1(e) and 1(f)).

The patient underwent thoracoscopic right lung bulla resection in the department of thoracic surgery in September 2021 because of multiple pulmonary cysts and recurrent spontaneous pneumothorax. Multiple pulmonary cysts were seen in the upper and lower lobes of the right lung, which were excised in stages with the Endo Gia Ultra cutting suture and ligated with multiple pulmonary cysts, respectively. The wedge resection of lung tissue postoperative pathology is as follows: right lung wedge resection tissue was off-white, with several cysts. Bullosa formation was observed under the pulmonary membrane, fibrous capsular wall tissue was observed, foam cell aggregation and multinuclear giant cell reaction were observed within it, and vascular dilation and congestion were evident in the surrounding lung tissue (Figure 2(a)). Immunohistochemical analysis of wedge resection of lung tissue illustrated positive cytokeratin (CK) and positive smooth muscle actin (SMA) (Figures 2(b) and 2(c)).

3.2. Genealogical Analysis of Birt–Hogg–Dubé Patients

The proband’s CT examination showed diffuse cystic changes in both lungs, The proband′s renal magnetic resonance plain scan and dynamic enhancement of renal indicated that left renal cyst. The proband was highly suspected of Birt–Hogg–Dubé, and whole exome sequencing was performed to verify the causes of the disease. Gene test revealed a unique deletion mutation in Exon 4 of the FLCN gene (FLCN NM_144997.7:c.21_22del p. Cys8Profs ^∗28) among the mutations associated with the disease phenotype, which was not embodied in the Leiden Open Variant Database and was not reported in the literature.

Chest CT of the proband’s Sister 1 of the prognosticator suggests multiple pulmonary cysts in both lungs in January 2022 (Figures 1(d) and 3(a)). The renal MRI of the proband’s Sister 1 showed that renal cysts of about 15–17 mm in diameter were at the lower pole of the right kidney and the upper level of the left kidney, with no abnormal enhancement on dynamic enhancement scanning (Figures 1(f) and 3(d)). Proband’s Sister 1 gene test also detected mutations in the same locus of the FLCN gene. The proband’s mother has already passed away, so genetic testing was not conducted. Based on her history of recurrent pneumothorax and death due to cor pulmonale, it was speculated that the mutation was inherited from the mother.

No significant abnormalities were seen on CT and MRI of the lungs in other members of the family. FLCN gene tested as wild type in the proband’s father, the proband’s Sister 2, and the proband’s son (Figures 4(a) and 4(b)).

3.3. The mRNA and Protein Levels of FLCN Were Significantly Downregulated by Transfection of FLCN Mutation

FLCN mutation c.21_22del p. Cys8Profs ^∗28 results in the generation of the PTC in Exon 4 (Figure 5(a)). Bioinformatics analysis suggested that the deletion of bases at Positions 21 and 22 in FLCN resulted in a frameshift mutation and the generation of PTC, leading to the generation of the truncated FLCN protein (Figure 5(b)). HA-FLCN was adopted, which was used to measure the relative mRNA levels and protein expression of wild-type and mutant FLCN, respectively. qPCR was performed to determine the relative mRNA levels. In comparison to the wild-type group, the relative mRNA levels in the mutant group decreased significantly (Figure 5(c)). Subsequently, the expression of corresponding proteins was detected by means of WB. Because of the presence of PTC, the mut group would generate truncated proteins. In Phage-HA-FLCN-flag, the mutant protein was significantly decreased, suggesting that the truncated protein may express in very low amounts due to premature termination (Figure 5(d)). Taken together, these data suggested that leads to a decrease in mRNA-expressing levels and lower protein expression. These effects were likely partially responsible for the loss of FLCN protein function in this Birt–Hogg–Dubé patient.

3.4. Transfected FLCN Mutation Reduced FLCN mRNA Expression Through the NMD System

It was hypothesized that 5 ^′ proximal PTCs circumvent NMD by reinitiating translation downstream [15]. FLCN mutations c.21_22del p. Cys8Profs ^∗28 reported in this study were located too close to the initiation codon and therefore may escape NMD, not causing changes in mRNA expression levels. Bioinformatics prediction, NMD assay, PCR, and WB were used to verify whether FLCN mutations affected protein expression and led to the clinical phenotype through the NMD pathway. The online bioinformatics prediction (https://nmdpredictions.shinyapps.io/shiny) results (Figure S1) also suggested that the mutation site was in the NMD-competent region where NMD may occur. The degradation of mRNAs containing premature stop codons has been attributed to NMD, which may explain the observed decrease in mutant FLCN mRNA levels compared to wild type. To investigate this hypothesis, we initially suppressed the expression of UPF1, a crucial constituent of the NMD pathway, and subsequently assessed its impact on the FLCN mRNA abundance. In comparison to the si-NC group, the si-UPF1 group exhibited a significant downregulation in both mRNA levels (Figure 6(a)) and protein levels (Figure 6(d)) of UPF1 in cells of wild type and mutant type. However, the knockdown of UPF1 expression resulted in a significant increase in mutant FLCN mRNA and protein levels, while the levels of wild-type FLCN remained unaffected (Figures 6(b) and 6(d)).

To investigate the impact of NMD, wild-type or mutant FLCN was introduced into 293 T cells and subsequently exposed the cells to CHX, a protein synthesis inhibitor, in order to evaluate its potential to impede NMD-mediated mRNA degradation [16]. In comparison to the wild-type (CHX−) group, the expression of FLCN mRNA was notably reduced in the mutant (CHX−) group. The administration of CHX+ did not significantly impact FLCN mRNA expression in the wild-type group but effectively reversed the downregulation of FLCN mRNA expression in the mutant group (Figure 6(c)). This suggested that CHX prevents the degradation of mutant FLCN mRNA through NMD. These findings provided further support for our hypothesis that NMD plays a role in the decreased expression of mutant FLCN.

3.5. Knock-in FLCN Mutation by CRISPR/Cas9 Reduced FLCN mRNA Expression Through the NMD System

Here, an efficient strategy to knock-in FLCN mutation (c.21_22del p. Cys8Profs ^∗28) into the genomic locus was applied to human bronchial epithelial cell BEAS-2B. The online design software was used to design and select sgRNA site sequences with low miss efficiency (Figure S2(A)). The results of sequencing verification showed that all three targets were successfully edited, and the editing efficiency of Tg1 was the highest (Figure S2(B)). The expression level of FLCN mRNA in the mock group and FLCN-KI group was detected using primer FLCN-qPCR-F/R. The results showed that compared with the mock control, the expression of FLCN-KI cell lines decreased to 52% (Figure S2(C)).

The monoclonal cell lines (Figure S3(A)) were sequenced, and the mutant cell lines were identified for culture (Figure S3(B)). According to the sequencing results, Tg1-FLCN was preliminarily identified as a positive clone. PCR products were connected to the T vector and transformed into DH5α, and monoclonal was selected for sequencing. The sequencing results showed that one DNA strand of Tg1-FLCN was successfully introduced with mutation c.21_22del p. Cys8Profs ^∗28, while the other DNA strand was wt type. FLCN-KI cell lines were successfully constructed in line with heterozygous mutation requirements (Figure S3(C)).

The expression level of FLCN in CHX and RNA si-UPF1-treated mock group and FLCN-KI group samples was detected using primers FLCN-qPCR-F/R. The expression level of UPF1 in RNA si-UPF1-treated mock group and FLCN-KI group samples was detected using primers UPF1-qPCR-F/R. Compared with the mock (CHX−) group, FLCN mRNA expression in the FLCN-KI (CHX−) group was significantly downregulated. CHX+ treatment had no significant effect on FLCN mRNA expression in the mock group. However, the downregulation of FLCN mRNA expression in the FLCN-KI group was significantly reversed (Figure 7(a)). Compared with the si-NC group, the expression of UPF1 in mock and FLCN-KI cells of the si-UPF1 group was significantly downregulated (Figure 7(b)). Compared with the mock (si-NC) group, FLCN mRNA expression in the FLCN-KI (si-NC) group was significantly downregulated. si-UPF1 had no significant effect on FLCN mRNA expression in the mock group but could significantly reverse the downregulation of FLCN mRNA expression in the FLCN-KI group (Figure 7(c)).

3.6. FLCN Mutations Lead to the Occurrence of Birt–Hogg–Dubé Through the AMPK, Wnt/β-Catenin, and mTOR Signaling Pathways

According to the American College of Medical Genetics and Genomics (ACMG) guidelines for classification criteria, to seek additional pathogenic evidence for FLCN mutations, the impacts of FLCN on the expression levels of key proteins in signaling pathways in both Birt–Hogg–Dubé and simple pulmonary bulla patients were investigated. Positive expression of FLCN in epithelial and mesenchymal cells was seen in both Birt–Hogg–Dubé and simple pulmonary bulla patients, and the grayscale value of positive cells of FLCN in Birt–Hogg–Dubé patients was lower than in patients with simple pulmonary bulla (95.000 ± 1.546 vs. 143.067 ± 1.115, Figures 8(a) and 8(b)).

Since the FLCN gene was involved in the synthesis of FNIP-1 as well as FNIP-2, where FNIP-1 was involved in the AMPK pathway [17], FNIP-2 acts on the mTOR pathway and regulates cell growth and reproduction [18]. And the loss of FLCN inhibited canonical Wnt signaling via TFE3 [19]. The expression of AMPK (P-Smad3), Wnt/β-catenin (β-catenin), and mTOR (p-mTOR) signaling pathways was verified by immunohistochemistry.

The grayscale value of positive cells of P-Smad3 (AMPK signaling pathway) in Birt–Hogg–Dubé patients was higher than in patients with simple pulmonary bulla (107.133 ± 1.759 vs. 72.933 ± 6.017, Figures 8(c) and 8(d)). The grayscale value of positive cells of β-catenin (mTOR signaling pathway) in Birt–Hogg–Dubé patients was lower than in patients with simple pulmonary bulla (66.467 ± 1.142 vs. 114.933 ± 1.307, Figures 8(e) and 8(f)). The grayscale value of positive cells of p-mTOR (mTOR signaling pathway) in Birt–Hogg–Dubé patients was lower than in patients with simple pulmonary bulla (98.000 ± 1.630 vs. 67.800 ± 2.209, Figures 8(g) and 8(h)). The gray value differences in FLCN, P-Smad3, β-catenin, and p-mTOR between Birt–Hogg–Dubé patients and non-Birt–Hogg–Dubé patients were statistically significant (p < 0.001, Figure 8(i)).