Article Text

Download PDFPDF
Short report
Molecular characterisation of oncogenic urothelial mosaic mutations in patients with extensive keratinocytic epidermal naevi
  1. Alejandra Gadea1,
  2. Inmaculada Hernández-Muñoz1,
  3. Asunción Vicente2,
  4. Evelyn Andrades1,
  5. Miriam García-Calvente3,
  6. Laura Camacho4,
  7. Conchi Fernandez-Rodríguez5,
  8. Beatriz Bellosillo4,6,
  9. Ramon Pujol1,7,
  10. Agustín Toll8
  1. 1 Group of Inflammatory and Neoplastic Dermatological Diseases, IMIM, Barcelona, Catalunya, Spain
  2. 2 Dermatology, Sant Joan de Deu Hospital, Barcelona, Catalunya, Spain
  3. 3 Department of Medicine, University of Barcelona, Barcelona, Catalunya, Spain
  4. 4 Department of Pathology, Hospital del Mar, Barcelona, Catalunya, Spain
  5. 5 Department of Pathology, Consorci Parc de Salut MAR de Barcelona, Barcelona, Catalunya, Spain
  6. 6 Cancer Research Program, IMIM (Hospital del Mar Medical Research Institute), Hospital del Mar, Barcelona, Catalunya, Spain
  7. 7 Dermatology, Consorci Parc de Salut MAR de Barcelona, Barcelona, Catalunya, Spain
  8. 8 Dermatology, Hospital Clinic de Barcelona, Barcelona, Spain
  1. Correspondence to Dr Agustín Toll, Dermatology, Hospital Clinic de Barcelona, Barcelona 08036, Spain; atoll{at}


Background: Keratinocytic epidermal naevi (KENs) are congenital benign skin mosaic lesions that share common mutations with some subsets of urothelial carcinomas. Moreover, several patients with extensive KEN who also developed urothelial carcinomas at young ages have been reported. Thus, patients with extensive KEN may harbour mosaic urothelial oncogenic mutations that would favour the early development of urothelial carcinomas. Methods: We selected five patients with extensive KEN involving the lower part of the back and performed a molecular characterisation of urothelial and cutaneous samples using a next-generation sequencing (NGS) custom panel targeting candidate oncogenic genes. Results: Mosaic pathogenic mutations were detected in KEN in all patients. In four out of five patients, mosaic pathogenic mutations in FGFR2 or HRAS were also detected in samples from the urothelial tract. Moreover, we report a patient who developed urothelial carcinomas at age 29 and harboured an HRAS G12S mutation both in skin and urothelial tumour samples. Conclusions: We conclude that patients with extensive KEN involving the lower part of the back frequently harbour oncogenic mutations in the urothelium that may induce the development of carcinomas. NGS panels can be considered as highly sensitive tools to identify this subgroup of patients, which might permit adoption of screening measures to detect malignant transformation at early stages.

  • epidermal nevus
  • urothelial cancer
  • mosaicism
  • genetic screening/counselling

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Keratinocytic epidermal naevi (KENs) are benign congenital and usually single warty lesions of variable size, displaying a linear or whorled distribution following the lines of Blaschko. They originate from pluripotential germinative cells in the basal layer of the epidermis and are considered to represent forms of mosaicism, usually resulting from postzygotic mutations in embryonic cells destined to populate a particular area of the epidermis. Mutations occurring very early in embryonic development give rise to more extensive epidermal naevi and may potentially be associated with a variety of developmental abnormalities, within the group of epidermal nevus syndromes characterised by keratinocytic naevi.1

Approximately 60% of KENs are caused by mutations in mosaicism in the PI3K-Akt (mainly FGFR3 and PIK3CA genes) and Ras-Raf-MEK-ERK (HRAS and NRAS genes) pathways or in both.2 These pathways are closely related and are implicated in the regulation of proliferation and cell survival. In fact, PIK3CA is one of the main effectors of the way of RAS and is necessary for RAS-induced transformation in vitro.3

Somatic mutations frequently found in KEN are also common in urothelial cancer.2 Moreover, several patients with extensive KEN who also developed urothelial cancer at young ages have been reported.4 5 Constitutional mutations in these genes can be found in several developmental disorders, which show a phenotypical overlap and may give predisposition to cancer. Costello syndrome (OMIM #218040), caused by mutations in HRAS, is characterised by facial thickening, epidermal alterations, cardiomyopathy and predisposition to cancer. Thirteen to fifteen per cent of patients with Costello syndrome will develop cancer in their lifetime, among which are rhabdomyosarcoma, bladder cancer and neuroblastoma.6

In this sense, our group had previously reported a patient with extensive KEN who also developed bladder cancer at the—unusually—early age of 19 years. The same G12S mutation in HRAS was found in the KEN tissue and in all tumour samples, including lung metastases.7 Thus, we can hypothesise that the risk of urothelial cancer is increased in some patients with extensive forms of KEN, which may harbour the same oncogenic mutation both in the skin and urothelium. Nonetheless, a non-aggressive genetic diagnostic tool to study the urothelium of patients with KEN has not been explored to date. In the present study, we searched for mutations in Ras and PIK3CA pathways in skin and urine samples from a series of patients with extensive KEN.

Materials and methods


A retrospective review of patients with KEN in the Departments of Dermatology from Hospital del Mar and Hospital Sant Joan de Deu, Barcelona, was performed. Five patients were finally included in the study. Only those patients with involvement of at least 10% of the body surface area and presenting with lesions on the lower part of the trunk were included (figure 1). Two patients presented with isolated KEN lesions with no associated manifestations, and two patients presented with KEN associated with additional extracutaneous manifestations. One case (P02) corresponded to a mosaic form of Beare-Stevenson syndrome (previously reported8). Importantly, one patient (P04) had already developed two urothelial carcinomas (TaG1 and TaG2) starting at age of 29 years. The clinical features of these patients are detailed in table 1.

Table 1

Patient's clinical manifestations and found mutations

Figure 1

Clinical image of patient P05 showing an extensive keratinocytic epidermal naevus involving the lower part of the back.

Skin biopsy specimens were obtained from all patients (formalin-fixed paraffin-embedded or fresh samples), along with urine desquamated squamous epithelial cells and, in one case (P04), also from an urothelial tumour sample. Peripheral blood was also obtained from four patients.

DNA extraction

DNA was extracted with QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the type of sample. The quality and the amount of DNA were measured using Nanodrop 1000 (Fisher, Waltham, Massachusetts, USA) and Qubit 3.0 fluorometer (Life Technologies, Carlsbad, California, USA).

Panel design

A custom panel, QIAseq Targeted DNA Custom Panel (Qiagen), was designed for subsequent sequencing in MiSeq (Illumina, San Diego, California, USA) and included all exons of the following selected genes: PIK3CA, HRAS, NRAS, KRAS, FGFR2, FGFR3, PTEN, AKT1, PTPN11, NF1 and SPRED1 and showed 100% coverage of the region of interest (26 890 bp). The amplicon library was developed following the protocol of the QIAseq Targeted DNA panel (Qiagen), and the amplified libraries were quantified using the Agilent 2100 Bioanalyzer instrument with the Agilent High Sensitivity DNA Kit (Agilent Technologies, Santa Clara, California, USA).


Preparation for sequencing was carried out following the Illumina Experiment Manager V.1.2 guide, with the Illumina Nextera XT V.2 adapter sample index system. For sequencing, we used MiSeq equipment (Illumina). The average sequence coverage was approximately 2.700×.

Variant analysis

To determine the results of the next-generation sequencing (NGS), the variant calling format files (.vcf) were analysed in the Illumina VariantStudio V.3.0 software using the hg19 version of the reference genome. Additional discriminatory parameters were applied to facilitate the interpretation of the data. In summary, all genotypic categories (heterozygous, homozygous and hemizygous) were included, all types of variants (Single Nucleotide Variant (SNV), insertions, deletions and differences with the reference genome) and all the possible consequences of the variants. A quality filter of >99 and a coverage (read depth) of >100 were applied. Only variants with a global population frequency <5% were included to discard common Single Nucleotide Polymorphisms (SNPs). In addition, the variants were reviewed visually by analysing the bam file in the Integrative Genomics Viewer V.2.4 tool.


Pathogenic mutations were detected in all patients (table 1). The variations were located at the following genes: PIK3CA (1), FGFR2 (1), FGFR3 (1) and HRAS (2). The performance of the NGS panel was validated by PCR in three patients (P02, P04 and P05).

Mutations present in the KEN were found in mosaicism in all patients. Pathogenic mutations in samples from the urothelial tract were found in four out of five patients (urine desquamated squamous epithelial cells in three patients and in tumour samples from P04). The mutant allele burden of urine desquamated squamous epithelial cells (P01, P02 and P05) ranged from 2% to 19%. Importantly, in all cases, the percentage of mutations was higher in urine than in blood. The mutant allele burden in two tumour samples from P04 was 31% and 49%, respectively. Two of our patients (P04 and P05) had HRAS hotspot mutations that have previously been reported in urothelial carcinomas.9 P05 had a p.G12V HRAS mutation both in the skin and urine, while P04 harboured a p.G12S HRAS mutation in mosaicism in bladder carcinoma, KEN and blood samples.

A PIK3CA mutation (p.P83S) was found in case P01. To our knowledge, this mutation has only been described in the literature causing CLAPO syndrome10 (OMIM #613089), a disorder within the spectrum of PIK3CA-related overgrowth syndrome (PROS). These patients show lower lip capillary malformations and lymphatic malformation of the face and neck, features that were not present in P01, who, however, had other CLAPO characteristics such as hemibody overgrowth and lip capillary malformation.

The FGFR2 mutation (p.Y375C) found in lesional skin from P02 had been previously reported by our group.8 We found the same mutation (p.Y376C numbering relative to NP_075259 corresponding to the epithelially expressed FGFR2b isoform11) in urine and blood. This variant has been reported in the germline in individuals with Beare-Stevenson syndrome (OMIM #123790).12 Although an association between this mutation and endometrial carcinomas has been reported,13, no association with bladder cancer has been described.

P03 only showed an FGFR3 mutation (p.R248C) in the KEN. Although this mutation has been associated with urothelial carcinoma,14 its absence in the urine sample using high-sensitivity techniques such as NGS probably renders an increased risk of future bladder cancer rather unlikely in this patient.


Bladder cancer usually develops in elder patients, with a peak incidence in the sixth decade of life, with less than 2.4% of the cases presenting under the age of 40 years.15 This early onset subset of neoplasms represents a molecular distinct group, with an uncommon high frequency of HRAS mutations,16 while PIK3CA and FGFR3 are the most common mutations in the majority of bladder cancers.17 Some patients with urothelial cancer at young ages may result from congenital postzygotic mutations, also giving rise to widespread KEN. The association between urothelial cancer and extensive KEN has been previously reported, but little data are available concerning the genetic background of this association.

Here we present the results of a molecular characterisation of patients with extensive KEN by means of an NGS panel targeting genes related to urothelial cancer. We selected five patients who had extensive KEN involving the lower part of the back. Although we are not able to rule out the presence of mosaic mutations in the urothelium of patients with small KEN in other areas, we consider it to be less probable, and it was not the scope of the present study.

Molecular characterisation by means of an NGS panel targeting genes related to urothelial cancer showed shared oncogenic mutations in both the skin and urothelium in one patient who developed bladder carcinoma at early age, and in three out of four patients with no urothelial manifestations.

The prognostic significance of detection of mosaic pathogenic mutations in PIK3CA, FGFR2, FGFR3 and HRAS genes in samples from the urothelial tract has not been determined. However, a periodic and close follow-up of this subgroup of patients seems advisable. Previous mutational data reported in the literature may also aid in establishing a risk grading for patients harbouring extensive KEN and mutated urine desquamative cells. In this sense, the same HRAS mutation (G12S) has been found in two patients with KEN and urothelial cancer (P04; see article by Hafner et al 7) and is frequently reported in young patients with bladder cancer16 and also in patients with Costello syndrome who develop urothelial cancer.9 Therefore, patients with KEN who harbour G12S or G12V HRAS mutations in the bladder tissue probably represent a group at special risk of urothelial carcinoma development.

On the contrary, PIK3CA and FGFR3 are commonly mutated in bladder cancers in the elderly.15 PIK3CA mutations are very common in a large variety of tumours,18 as the PI3K/AKT signalling pathway regulates cell growth, proliferation, migration, metabolism, survival and angiogenesis, among others. Despite this, cancers enriched for such mutations, such as urothelial cancer, are seldom reported in patients suffering from PROS.19 It is believed that, although mosaicism can be present in various organs, only specific tissues are prone to the early development of cancer in the presence of particular mutations.20

In conclusion, the study of molecular alterations by means of NGS panels in urine desquamative cells may be considered a non-invasive screening method to early detect oncogenic mutations in patients with extensive KEN at risk of urothelial cancer. We believe that patients with KEN who harbour G12S or G12V HRAS mutations in the bladder tissue are at special risk of developing urothelial carcinoma. We report the second patient who developed urothelial cancer (P04) at a young age and harboured the HRAS G12S mutation both in skin (KEN) and urothelial tumour samples. The observation of two patients with this association in the same centre would thus argue against its infrequency. Future studies with larger series are needed to determine the optimal screening techniques and recommended follow-up regimes in patients with widespread KEN and urothelial mosaic oncogenic mutations to rule out urothelial cancer at early stages.


The authors thank Dr FX Real and M López-Jurado for their technical support to perform this study.



  • Contributors AG performed all the genetic analysis and designed the genetic panel, and drafted and revised the paper. EA, MG-C, LC, CF-R and BB performed the genetic analysis and revised the paper. AV and RP gathered all clinical data from the patients, and drafted and revised the paper. IH-M analysed the genetic data and revised the paper. AT implemented the study, collected the clinical data and drafted and revised the paper.

  • Funding This study was supported by the grant PI15/00236 from Fondo de Investigación Sanitaria (FIS), Instituto de Salud Carlos III FEDER, Ministerio de Economía y Competitividad, Spain, and from the “Xarxa de Bancs de tumors sponsored by Pla Director d’Oncologia de Catalunya (XBTC)

  • Competing interests None declared.

  • Patient consent for publication Obtained.

  • Ethics approval The study was approved by the clinical ethics committee of the centre and complies with the regulations of the Declaration of Helsinki, Fortaleza, Brasil 2013, and the Organic Law 15/1999 on the Protection of Personal Data. Informed consent has been obtained from patients over the age of 12 years or from parents.

  • Provenance and peer review Not commissioned; externally peer reviewed.