Histology

Individuals with BHD syndrome are predisposed to develop renal cell carcinoma (RCC) (Toro et al., 1999; Zbar et al., 2002). Unlike other genetic disorders with this predisposition, renal tumours associated with BHD syndrome are histologically diverse. A large study analysing the histology of 130 renal tumours resected from 30 BHD patients found that 50% (65/130) were of hybrid chromophobe/oncocytoma histology; 34% (44/130) were chromophobe; 9% (12/130) were clear cell; 5% (7/130) were oncocytoma; and 2% (2/130) were papillary (Pavlovich et al., 2002). This pattern is strikingly different from the spectrum of sporadic renal cell carcinomas, where 75% are of clear cell histology, 10% are papillary, 5% are chromophobe and 3-5% are oncocytic (Pavlovich et al., 2002).

However, in a cohort of 14 BHD patients with RCC, tumours with chromophobe and clear cell characteristics were found to occur most frequently (Houweling et al., 2011). In a second cohort of 33 BHD patients with kidney cancer, although mixed chromophobe/ oncocytoma was the most commonly observed tumour type, present in 23/33 patients, three patients (9%) had clear cell tumours (Benusiglio et al., 2014b). Together, these studies suggest that BHD-associated renal cell carcinomas are histologically diverse.

Several other mixed patterns can also occur (Pavlovich et al., 2005; Fahmy et al., 2007; Janitzky et al., 2008; Kluijt et al., 2009) and 12 deaths have been reported in BHD patients due to metastatic renal cancer, all of which were of the clear-cell, hybrid clear-cell/papillary or hybrid clear-cell /chromophobe histology (Pavlovich et al., 2005; Toro et al., 2008; Claessens et al., 2010; Houweling et al., 2011; Nakamura et al., 2013). The International Society of Urological Pathology Tumour Panel recently recommended that as hybrid oncocytic/chromophobe tumour only occurs in three settings, a diagnosis of BHD should be considered when this tumour is found (Srigley et al., 2013). However, the case of a Japanese patient with multiple bifocal chromophobe renal cell carcinomas and capsular angiomyolipomas, but no FLCN mutation, suggests that there is at least one chromophobe RCC predisposition gene to be discovered (Sugimoto et al., 2013).

Renal angiomyolipomas, benign tumours which are most commonly associated with tuberous sclerosis complex, have also recently been identified in two BHD patients (Tobino and Seyama, 2012; Byrne et al., 2012). Tumour histology is non-concordant within families, indicating that there is unlikely to be a genotype-phenotype correlation with regards to tumour histology (Pavlovich et al., 2005).

However, a study by Gatalica et al. (2009) analysed three tumours of different histology found in the kidney of a suspected BHD patient with a germline FLCN mutation and identified a genotype-phenotype correlation. An oncocytic tumour was found to have a second FLCN mutation; an oncocytic-papillary tumour showed increased methylation of the FLCN promoter and somatic mutation of the MET gene; and a clear-cell tumour had a somatic mutation in the VHL gene and increased VHL promoter methylation. Inactivating mutations in VHL and activating mutations in MET cause clear cell RCC and hereditary papillary RCC respectively (Latif et al., 1993; Olivero et al., 1999), meaning that these results are perhaps unsurprising. However, the authors do not show whether the germline FLCN mutation observed in this patient (c.1062+6C>T) actually affects splicing. Therefore whilst these results are interesting, their significance is uncertain at present.

A more recent study reports the pathological findings of kidney tumours in a series of six Japanese patients with confirmed BHD (Kuroda et al., 2014). Five patients had multifocal tumours, two had bilateral tumours and one patient had a solitary tumour. The tumour series consisted of one tumour of unclassified histology, but with features resembling hybrid chromophobe/ clear cell histology; three hybrid oncocytic/chromophobe tumours; one collision tumour consisting of chromophobe; clear cell and papillary tumours; a chromophobe tumour; and a clear cell tumour. All tumours had intertumoral peripheral small papillary tufts (ITPSTs) at the interface between the tumour and normal kidney tissue. The authors suggest that the presence of either collision tumours or multiple tumours of different histology, and ITPSTs – as they were universal in this cohort of patients – might be diagnostic hallmarks of BHD-associated kidney tumours.

It is unclear from which part of the kidney the tumours arise. Due to the high percentage of chromophobe tumours, Pavlovich et al. (2002) initially believed the tumours to arise from the distal nephron. However, two independent studies observed FLCN expression in the proximal tubules of murine kidneys, suggesting this is the site of origin (Chen et al., 2008; Hudon et al., 2010).

Work on a marker panel to distinguish BHD tumours from sporadic RCC is ongoing. Iribe et al., 2015 reported that FLCN-associated hybrid tumours show decreased expression of CK7 compared to sporadic chromophobe RCCs but increased expression of Ksp-cadherin and CD82 compared to sporadic oncocytomas. In addition Furuya et al., 2015 reported that tumours from BHD patients showed markedly reduced expression of FLCN and increased expression of GPNMB. Kato et al. (2016) then reported that BHD-chRCC and HOCTs retained chromosome 17 disomy unlike sporadic tumours which are typically monosomic. So far these markers are only able to distinguish between some of the sporadic and BHD-associated tumour subtype (summarised in the table). The identification of further markers will allow for greater understanding of the underlying biology but also more accurate RCC subtype diagnoses.

Sporadic RCC subtype	BHD-associated RCC subtype	Distinguishing markers
Chromophobe RCC	Chromophobe RCC	↓FLCN, ↑GPNMB, 17q/2p/6p disomy
Chromophobe RCC	HOCT	↓CK7, ↓FLCN, ↑GPNMB, 17q/2p/6p disomy
Oncocytoma	HOCT	↑Ksp-Cadherin, ↑CD82, ↓FLCN, ↑GPNMB
Papillary RCC	Papillary RCC	↓FLCN, ↑GPNMB

Prevalence

Renal cancer is the most life-threatening complication associated with BHD syndrome. Previous estimates of RCC prevalence among BHD patients have varied. Houweling et al. (2011) reported that 14 (12%) individuals in a cohort of 115 Dutch BHD syndrome patients developed renal cancer and further analysis of 22 BHD families of Dutch origin found the lifetime risk for RCC to be 16%. However, Pavlovich et al., (2005) and Toro et al. (2008) studied large cohorts of American BHD patients and found the prevalence of RCC to be 34/124 (27%) and 30/89 (34%) respectively. A more recent French study found the prevalence to be 33/124 (27%) (Benusiglio et al., 2014b).

The difference in these estimates may be due to population differences between cohorts, or ascertainment bias: the patients in the Dutch study were recruited predominantly via dermatology clinics, whereas the cohorts in the American and French studies were recruited via both dermatology and urology clinics. Due to these ascertainment differences, the Houweling et al. (2011) estimation is likely to be low, whilst the Pavlovich et al., (2005), Toro et al. (2008) and Benusiglio et al. (2014b) estimations are likely to be high. Therefore, the risk of developing kidney cancer is likely to be between 12 – 34%.

A study of 130 tumours from 30 patients showed the prevalence of multifocal disease to be 60% (18/30) and that of bilateral disease to be 77% (23/30) (Pavlovich et al., 2002).

Somatic FLCN mutations have been reported in sporadic cases of RCC (Khoo et al., 2003; Gad et al., 2007): FLCN mutations were found in 1/39 (2.6%) (Khoo et al., 2003) and 6/92 (6.5%) (Gad et al., 2007) renal tumour samples (five clear cell RCC, one papillary RCC and one oncocytoma), indicating that FLCN mutations only cause a small percentage of sporadic RCCs.

Heterozygous loss of FLCN was also reported to cause the transformation of an oncocytoma to a high grade oncocytic carcinoma. However, there were additional genomic rearrangements seen in this tumour, suggesting that additional factors also contributed towards disease progression in this patient (Sirintrapun et al., 2014).