Clinical and genetic characteristics of chinese patients with Birt-Hogg-Dubé syndrome

Birt-Hogg-Dubé syndrome (BHD, OMIM #135150) is a rare autosomal dominant disorder, the main symptoms of which are multiple pulmonary cysts followed by recurrent pneumothorax, fibrofolliculomas, and renal cell carcinomas. These three symptoms appear separately [1, 2]. Lung-related symptoms are often the earliest phenotypical manifestations to appear, but most patients are asymptomatic [3, 4]. The pulmonary manifestations of BHD occasionally need to be distinguished from other conditions associated with diffuse cysts lung diseases (DCLD), such as lymphangioleiomyomatosis, Langerhans cell histiocytosis, lymphocytic interstitial and pneumonitis [5]. Lack of a comprehensive understanding of BHD often leads to a high misdiagnosis rate.

The gene responsible for BHD syndrome, the folliculin (FLCN) gene on chromosome 17p11.2, is a tumor suppressor gene that was first reported in 2002 [6] and is known to be involved in the signaling of mammalian target of rapamycin (mTOR) [7]. The FLCN gene consists of 14 exons encoding a 579-amino-acid-long protein, folliculin [8]. Disease-causing mutations in the FLCN gene, including insertions, deletions, missense and nonsense mutations, were found over the entire gene. Multiple in vitro studies focused on FLCN functions suggested that the activation of the AKT-mTOR pathway and increased activity of basic-helix-loop-helix transcription factor TFE3 were related to FLCN-deficient mouse cell lines [9‐11]. To date, 149 unique FLCN germline mutations have been identified in BHD patients and catalogued in the Leiden Open Variation Database. Since Nickerson et al. first described BHD syndrome in 2002 [6], studies of FLCN mutations and related manifestations have been popular in Europe and the United States. Analysis of Caucasian data demonstrated that the frequency of 1-bp deletion or insertion within a hypermutable C8 tract in exon 11 of FLCN was high; further molecular research also confirmed that the poly(C) tract in exon 11 of FLCN is a mutation hot spot [12].

Studies of FLCN mutations for BHD syndrome are less prevalent in China than they are in Europe and America. Japanese researchers reported gene mutations of FLCN in five patients with BHD in 2007; all of the mutations were unique, and four were novel [13]. A recent genetic study of Japanese patients with BHD syndrome published in 2016 included 312 patients from 120 different families and identified 31 FLCN sequence variants; two different mutation hot spots, c.1533_1536delGATG in exon 13 and c.1347_1353dupCCACCCT in exon 12, were found [14]. The difference between races may result in a different mutation spectrum in Asian compared with Caucasian populations. Nevertheless, studies of BHD syndrome in Asia are rare. BHD syndrome mutation analysis in Japanese populations concluded that isolated type with pulmonary involvement and recurrent episodes of pneumothoraces were more informative as diagnostic criteria for BHD in the Asian Japanese population [13, 14]. In China, Ren et al. reported sporadic and familial isolated primary spontaneous pneumothorax (PSP) and found that 10 PSP patients had FLCN gene mutations, none of whom had other features of BHD [15]. However, further studies of BHD syndrome in Chinese populations are warranted. This study aims to discover the mutation spectrum of the FLCN gene in Chinese patients with BHD and attempts to relate the mutation spectrum to the known phenotypes through literature review.

When BHD was described in 1977, its incidence was unknown. A study in Lancet Oncology in 2009 reported that approximately 200 families had been identified worldwide; but the number of affected families now far outnumbers that report [16]. It was not until 2008 that BHD was recognized by physicians in China [15]. No epidemiological data about the incidence rate of BHD in the Chinese population has been available until now. This is the first report of systematic mutation screening analysis of FLCN in a comparably large cohort of Chinese patients with BHD.

In 27 of the 51 patients with symptoms suggestive of BHD, we found 20 mutations in FLCN, including 14 novel and 6 already reported mutations. Two novel small deletions of nucleotides, c.933delT and c.214delA, caused a frameshift mutation, resulting in premature termination codons or triggering a nonsense-mediated mRNA decay (NMD). These mutations would generate possible structural and functional changes in folliculin or mRNA degradation; further functional studies were warranted. Three novel mutations involving the splicing sites of exons have been found in this study, c.780-1G > T, c.249 + 1G > T and c.1179-10_1179-8delTCC. The first two could change the acceptor site of intron 7 or the donor site of intron 4, respectively, which would generally cause exon skipping. The third one might create an intronic ESE site of intron 11 as predicted by Human Splicing Finder; however, ESE finder didn’t get the same prediction. Considering the typical clinical manifestations of BHD and also the positive family history (both her father and brother affected) of this patient, the pathogenicity of this variant might be strong. However, further functional tests are needed to validate this prediction. Additionally, six novel nonsense mutations, p.Gln220*; p.Gln53*, p.Gln339*, p.Glu389*, p.Trp553* and p.Glu49*, have been discovered; the resulting premature termination codons will indeed cause truncated protein production and/or NMD. One missense mutation, c1067T > C; p.Leu356Pro, in exon 10 was noted. PolyPhen2, Mutation Taster, FATHMM and PROVEAN gave “probably damaging/ disease causing/ damaging/ deleterious” results, and its REVEL score is 0.841. All the prediction tools supported its pathogenicity. In the future, however, functional tests of this missense variant will be needed to further provide solid evidence. In addition to the identification of single nucleotide variants and intragenic small indels mentioned above, exon deletion was detected by the MLPA test. A whole exon 8 loss was identified in one patient, and a truncated protein or NMD are highly likely to result from the loss of this exon. All of the variants identified in this study were absent from ExAC database except for the recurrent mutation hot spot (c.1285dupC/delC); they were all classified as pathogenic with some manual adjustments such as checkinging PVS1, PM2, PP3 and/or PP4 (http://​wintervar.​wglab.​org/​) based on ACMG/AMP 2015 guidelines (Table 1).

Similar to the mutation hot spots reported previously in Caucasians [23], 5 duplication and 4 deletion mutations at c.1285, a hypermutable C8 tract in exon 11, were found in our study, accounting for 33.3% (9/27) of all cases and suggesting that this mutation hot spot does not differ between ethnic populations. No other significant mutations were found to be a potential hot spot in the Chinese population. Nevertheless, 14 of 27 (>50% of total cases) FLCN positive patients were reported to have novel mutations, thus demonstrating the diversity of mutation spots along the gene and dramatically broadening the mutation spectrum of FLCN. Thus, the data might suggest race differences in the mutation spots between Chinese and Caucasians. However, considering the inactivation role of FLCN in the etiology of BHD, it is not surprising to see novel ones spreading all over the genes. More studies with larger populations of Chinese patients with BHD are needed to investigate this topic further.

The first early onset symptom in a substantial proportion of our BHD patients was pneumothorax (74%, 20/27), and a similar clinical pattern was reported in Japanese Asian populations in 2016 [14]. Typical skin fibrofolliculomas were only detected in two patients (Fig. 4); a probable explanation for the low detection rate was that the papules were inconspicuous and asymptomatic. Furthermore, no renal cell carcinoma was diagnosed in our BHD patients, except for 1 with hamartomas and 4 with renal cysts (Fig. 5). Patients enrolled in this study were mostly from a respiratory clinic, which could explain the low renal cell carcinoma rate. Though no significant association between FLCN mutation status and lung cyst parameters was stated, Dr. Toro et al. discovered that BHD mutations in exon 9 were associated with more lung cysts than other mutation loci [22]. Nevertheless, that correlation was not identified in our study, and no significant correlations between the extent of clinical manifestations severity and types of gene mutations were noted. Additionally, five of our patients did not undergo renal radiological screening because of the patients’ reluctance. Therefore, we cannot confirm a low prevalence of renal lesions in our group. The small number of patients may also have contributed to a lower rate of renal cancer in this study, although this study has a relatively large sample size compared with other studies of BHD in China.

There are several limitations of our study. Pulmonary cysts exist not only in patients with BHD but also in those with DCLD, such as lymphangioleiomyomatosis and Langerhans cell histiocytosis [5]. FLCN, FBN1, COL3A1, CBS, SERPINA1 and TSC1/TSC2 were involved in different DCLD. Here, we only focused on the FLCN gene and BHD syndrome. Zhang et al. developed a new detecting method using a rapid next generation sequencing-based (NGS) panel to differentiate FLCN gene mutations in patients with PSP [24]. In future efforts, targeted NGS panel or whole exome sequencing analysis on FLCN-negative cases might be necessary. Nevertheless, here, we provide further genetic characterizations of patients with symptoms of BHD.

In conclusion, this study reports 14 novel mutations of FLCN in 27 patients with BHD and is the first study to demonstrate the mutation spectrum of FLCN in a Chinese study population. The mutation spectrum in the Chinese population is even more extensively distributed over the entire FLCN gene than that in Caucasians. These genetic findings provide stronger evidence for the clinical molecular diagnosis of BHD in China. Our results suggest that mutation analysis of the FLCN gene should be systematically conducted in patients with cystic lung diseases.