Introduction

In late ‘70s three Canadian physicians introduced a novel genodermatosis syndrome characterized by the presence of multiple benign cutaneous neoplasms [1]. Birt-Hogg-Dubé syndrome (BHDS) (MIM# 135150), is an autosomal dominant disorder with unknown incidence. BHDS has as three main clinical features—the so called clinical triad; cutaneous hair follicle tumours, known as fibrofolliculomas or trichodiscomas, spontaneous recurrent pneumothorax in a pathological substrate of lung cysts and increased lifetime risk for kidney cancer [2–8]. Apart from the classic signs additional findings have been reported occasionally such as parotid oncocytoma, oral papules, skin melanoma, thyroid cancer and/or adenomas [2, 6, 7, 9–11]. The initial report in the literature regarding the association between BHDS and colon cancer with/or colon polyps is dated in mid ‘70s [12]. Since then, several cases of colon cancer and colon polyps in BHDS patients have been reported [3, 6, 13, 14], without strong evidence proving the association of this type of cancer with the syndrome due to conflicting outcomes.

The responsible genetic locus of BHDS was mapped with linkage analysis to chromosomal region 17p11.2 [13, 15], which subsequently led to identification of the causative FLCN gene [16]. The FLCN gene encodes the folliculin protein, a 64-kDa protein consisting of 579 amino acids that is expressed in a variety of tissues such as skin appendages, kidneys and lungs [16, 17]. Even though FLCN’s function is not fully understood, it has been suggested that FLCN acts as a tumour suppressor gene [18] interacting with FNIP1 and FNIP2 proteins, thus playing a key role in mammalian target of rapamycin (mTOR) signalling pathway [19–21].

The true incidence of the BHDS is still unknown due to clinical heterogeneity and variable penetrance, thus omitting a potential serious hereditary condition with significant pulmonary and kidney manifestations. Since its initial description approximately a few hundred families have been described. Reports of BHDS cohorts are valuable since they facilitate understanding of this condition, thus contributing to a better genotype-phenotype correlation. Herein, we present a large Swedish cohort of patients that were referred for genetic screening and consultation regarding BHDS.

Material and methods

Results

Genetic analysis

In total 278 individuals were tested for pathogenic variants in the FLCN gene between 2007 and 2019. Of these 186 (67%) (Table 1, S1 Table) were found to have a pathogenic variant in the FLCN gene. The variant spectrum revealed eleven different pathogenic variants encompassing six of the 14 exons of the gene (Fig 1) where ten of the variants were identifiable using sequencing analysis. Four separate families carried a deletion of the non-coding exon 1 that was detected using MLPA. The most common variant among families was the splicing variant c.779+1G>T that was found in 57% (44 of 78) of the families (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic figure illustrating the different variants in FLCN.

https://doi.org/10.1371/journal.pone.0264056.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Proportions of the different variant detected in our cohort.

https://doi.org/10.1371/journal.pone.0264056.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Description of included cases in the study.

https://doi.org/10.1371/journal.pone.0264056.t001

The, to our knowledge unrelated families, that carried the most common splicing variant were further analysed using haplotype analysis. Results from the haplotype analysis revealed that they shared one common haplotype surrounding the FLCN gene. The region spans between markers D17S740 to D17S620. This corresponds to a minimal common region of 0,58Mb to a maximal common region of 2,29Mb (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Haplotype analysis in families carrying the c.779+1G>T variant in the FLCN gene.

The shared haplotype is marked with italic and spans at least 0,58Mb up to 2,30Mb.

https://doi.org/10.1371/journal.pone.0264056.t002

Discussion

According to our knowledge the present study of 78 families is one of the largest cohorts ever published in current literature for BHDS [3–6, 10, 23–30]. In our series of 186 BHDS carriers we identified eleven different pathogenic and likely pathogenic variants that matched the clinical picture of carriers and their families (Table 1, S1 Table).

The most interesting finding in our study was the detection of a splice site variant c.779+1G>T in 44 Swedish families (96 individuals). This particular pathogenic splice-site variant has been previously reported in four families [6, 11]. In the first report [6], describing three of the families, the variant was associated with a variety of cutaneous manifestations, renal tumours, lung cysts and spontaneous pneumothorax. The proband of the fourth family [11], had only cutaneous symptoms and further investigation by imaging techniques revealed cystic changes of the lungs and a single renal cyst. This patient did not have either a medical record of pneumothorax or family history for pneumothorax and/or renal cancer. In our cohort, 54% of the carriers of this specific splice site pathogenic variant exhibited cutaneous manifestations, 49% had a history of at least one pneumothorax episode by the time of genetic test and 19% were affected by kidney tumours. Among 44 families, 37 (84%) had members with a skin phenotype, 33 (75%) had members with a lung phenotype and in 17 families (39%) there was at least one member that was diagnosed with renal tumour. In seven of 44 families (16%) there was a history for colon cancer and/or polyps (S2 Table).

According to population databases (gnomAD) this variant has a carrier frequency of 5/64548 in Europe and of these 4 are of Swedish origin. This means that the carrier frequency of this variant in the Swedish population (1/3265) is significantly higher compared to the rest of the included populations (gnomAD) [31]. However, the frequency of this pathogenic variant in our cohort was strikingly high in this group of patients, which raised the suspicion that it might be a founder mutation. Haplotype analysis of the 44 families verified this hypothesis proving they are distant relatives, thus implicating that this variant is a founder mutation in the Swedish population. A speculation that the previously reported four families in the United States could be of Swedish descent is quite tempting.

The frameshift deletion of 4 base pairs, c.235_238del p.(Ser79Thrfs*50), has been described once in a large Finnish pedigree, in which numerous members were tested for the family variant [32]. Remarkably, all carriers showed lung manifestations only, while non-carriers were asymptomatic, illustrating the co-segregation of this variant with a pulmonary phenotype in the family. In our study, the variant was found in two different kindreds with a positive family history for the clinical triad of BHDS. However, whether these two apparently unrelated families have a Finnish origin or could be related to the previously reported pedigree is unclear.

Deletions of the first segment of the gene encompassing exon 1 of the gene have been previously reported in four families/index cases with BHDS [33–35]. Despite that, the three first exons are non-coding, a deletion of the first exon eliminates the putative promotor of the gene, thus resulting in a substantial reduction of its expression [33]. Whole genome sequencing of one of our index cases showed that this deletion encompasses 8,1 kb ([GRCh37] del(17)(p11.2) chr17:g.17138220-17146328del). Phenotypic data in previous studies describe all three clinical hallmarks of the syndrome. In our series, two probands exhibited no significant symptoms but they had a positive family history for fibrofolliculomas, pneumothorax and kidney cancer. The other four had clinical symptoms consistent with BHDS.

An additional finding in our cohort was the detection of a 3 bp intronic deletion in position c.1177-5_1777-3del in one family. This variant has been reported before [28, 36] and is predicted to affect the splice acceptor site, thus leading to skipping of exon 11 and subsequently to a premature termination of folliculin synthesis. All three characteristic BHDS hallmarks have been described for that specific variant. In our family, the proband had only fibrofolliculomas, but there was a family history for skin lesions, spontaneous pneumothorax and lung cysts as well.

The most common duplication/deletion is in the “hot spot” c.1285 of the gene. This frequent variant, which has been observed in various cohorts with a variable frequency [6, 10, 14, 16, 27], was found in 15 unrelated pedigrees in our study. In particular, 14 families had the duplication of a cytosine at c.1285dup p.(His429Profs*27) and one family had the deletion of a cytosine at this position, c.1285del p.(His429Thrfs*39). The clinical spectral of analysed eight families included all typical signs for BHDS.

Other pathogenic variant that we detected in single families were a frameshift 2 bp deletion c.1301_1302del p.(Glu434Valfs*21) that has been reported once before [27] a nonsense variant c.1429C>T p.(Arg477*) (that has been detected previously [10]) and the three different novel variants c.241_242del p.(Met81Valfs*18) and c.1538+1G>A, c.43G>A p.(Gly15Ser).

Regarding the clinical description in general, distinctive cutaneous manifestations are the main features of the syndrome. These skin lesions are benign tumours of hair follicles named as fibrofolliculomas, trichodiscomas and acrochordons. They usually appear in the third or fourth decade of life as multiple white papules on the face, with a more common localization on the nose and cheeks, but they can also be visible on the neck, ears and upper trunk. They do not have any clinical implication and they are excised for psychological and cosmetic reasons with ambiguous outcomes [37]. In our series nearly half of all BHD carriers presented cutaneous lesions by the time of genetic investigation. Confirmation of these lesions performed either by clinical examination from a dermatologist and/or by histopathological examination.

Association between FLCN variants and pulmonary symptoms is well established, indicating that carriers have an up to 32-fold risk for spontaneous pneumothorax compared to general population [3], together with a high predisposition to develop lung cysts [3, 5, 6]. Especially, up to 89% of patients with BHDS can develop multiple pleural blebs and subpleural bullae localized mainly in basal segments of the lungs, that are considered precursor lesions to pneumothorax [6]. Radiographic investigation of these lesions regularly misleads physicians towards the diagnosis of emphysema. In our cohort 40% of carriers had a history of spontaneous pneumothorax. Moreover, an ascertainment with CT-thorax of BHDS carriers with or without a history of pneumothorax revealed that up to 26% of them had emphysematous-like bullae in their lungs. This outcome is lower than in previous reports [5, 6, 13, 23], nevertheless it should be mentioned that not all carriers were subjected to imaging analysis.

Several explications to the underlying molecular mechanism of cysts development in the lungs, have been suggested. FLCN protein (folliculin) has a tumour suppressor activity interfering with the mTOR signalling pathway, a fact that provides a plausible explanation for cyst formation [19]. Additionally, any variant which leads to a significant decrease of FLCN expression may induce the inflammatory process in alveolar walls or even lead to matrix degradation and remodelling [32, 38]. Furthermore, a similar conclusion from another study suggests that pulmonary bullae imply a deficiency in the structural proteins of the lung wall, which may be elements of the cytoskeletal network [16]. Finally, an interesting explanation suggests that the folliculin may have a ciliary role, thus resulting in pulmonary cyst formation [39].

The most important life-threatening consequence of the syndrome is a significantly increased lifetime risk up to 34% for kidney cancer or a 7-fold higher risk compared to normal population [3, 4, 6, 13]. The tumours described in literature are multifocal, bilateral, with a slow progress and low metastatic potentional [37]. Median age of diagnosis is 48 years with a range from 31 to 71 years [10]. The earliest reported case for BHDS is a young patient that was diagnosed with renal cancer at the age of 20 years old [13]. Genetic screening where renal tumour was the only clinical symptom in the index patient revealed a FLCN variant in eleven of 21 cases, emphasizing that BHDS is common in this disease. In the current study, carriers were diagnosed with renal tumour in 16% of cases and the youngest affected individual was 20 years old. Overall incidence of renal tumours in our study corresponds to previous reports, however more cases need to be documented in order to have a more accurate estimation of the occurrence of kidney tumours in BHDS.

According to our knowledge, only few reports for renal cysts in BHDS suggest that up to 45% of carriers may have these usually asymptomatic lesions [7]. In our cohort the frequency of renal cysts in carriers were 9%, but the number of examined radiologically individuals is small and not all carriers performed radiographic kidney investigation during the study period.

Even though the underlying mechanism for renal carcinogenesis is still unclear, several researchers suggest that folliculin’s role in mTOR pathway is related to kidney tumorigenesis by inactivating variants in FLCN gene coupled to a secondary loss of heterozygosity [16, 19, 39].

Since the first description of the concurrence between cutaneous fibrofolliculomas and colorectal cancer/ colon polyps [12], numerous studies have tried to prove this association in BHDS families with contradictory conclusions [3, 6, 7, 10, 13, 14, 28, 40, 41]. In our series, 5% of all carriers were affected by colorectal and/or colon dysplastic adenomas, but the size of the sample may not be adequate enough to draw a confident conclusion regarding this association. Moreover, a positive family history for these manifestations was noticed in 18% of BHDS families, yet this information should be interpreted with caution, since not all persons within each and every family underwent carrier testing for the familial disease causing variant. Whether colon cancer along with colonic polyps, which are quite frequent in the general population, may be a coincidental observation in BHDS or not, is still unclear and needs to be elucidated.

In conclusion, the current cohort provides valuable medical data accompanied by reports of eleven different pathogenic variants for a supposedly rare syndrome, among which one is an obviously common Swedish founder mutation. These data may allow a better understanding and evaluation of patients and families with a putative higher risk for life-threatening conditions such as spontaneous pneumothorax and renal cancer. The weakness in our study to assemble all necessary and essential clinicopathological features, for a more accurate BHDS assessment, is due to lack of national guidelines regarding diagnosis and management of the disorder. This problem may be similar in other Western countries as well. All scientific reports for BHDS though, illustrate the heterogeneity of this syndrome that often poses a diagnostic challenge to physicians. Therefore, a consensus that could provide up-to-date recommendations, regarding diagnostic and clinical management of BHDS is a necessary precondition in order to systematically ascertain the disorder.

Supporting information

Acknowledgments

We would like to thank all patients and families who participated in this study.