Introduction

Neurofibromatosis type 1 (NF1; OMIM 162200) is one of the most common hereditary disorders. It is inherited as an autosomal dominant trait. The estimated incidence at birth is 1/3,000 (1,2). Multiple café-au-lait spots, axillary/inguinal freckling, multiple cutaneous neurofibromas, iris Lisch nodules, and choroidal freckling are the characteristic features of NF1 (3). NF1 also manifests as multisystemic disorders with musculoskeletal, vascular, central nervous, and peripheral nervous system involvement such as scoliosis, tibial dysplasia, vasculopathy, glioma, and malignant peripheral nerve sheath tumors (4).

Clinical diagnosis of NF1 is based on the National Institutes of Health (NIH) diagnostic criteria (5). However, without a family history, these criteria may be insufficient in early childhood as most of the clinical features manifest later in life (3).

The NF1 phenotype is caused by germline heterozygous pathogenic variants of the NF1 gene. NF1 is located at chromosome 17q11.2 and coded neurofibromin 1 protein that acts as a regulator of Ras activity. NF1, one of the largest genes in the human genome, consists of 57 coding exons and 12,362 base pairs transcript length (transcript reference, NM_000267.3).

In recent years, next-generation sequencing (NGS) technology has enabled many genes, regardless of size, to be analyzed systematically, comprehensively, more easily and more cost-effectively. Most studies demonstrated that NGS with in-solution hybridization-based enrichment

provides a high mutation detection rate comparable to that of conventional direct capillary sequencing methods for the molecular diagnosis of neurofibromatosis. In this study, we aimed to present the variant spectrum of NF1 patients from the region of Southeastern Turkey and investigate if there is a clear genotype-phenotype correlation (6).

Materials and Methods

Profile of the Patients

The clinical and genetic data of the NF1 patients who were referred to the Medical Genetics Clinic, Gaziantep Ersin Arslan Training and Research Hospital between May 2016 and December 2019 were evaluated retrospectively. Peripheral blood samples were obtained after taking informed consent from all participants or the legal guardians of those children under the age of 18. The study was approved by the Ethics Committee of the Gaziantep University Medical Faculty (approval number: 65587614-774.99-291, date: 04/10/2017).

Ninety-two Turkish patients from 86 unrelated families who were both clinically and molecularly diagnosed with NF1 were included in this study. Fifty-four (58.7%) patients were male and thirty-eight (41.3%) were female. Age at diagnosis was in the range of 1-46 years.

The clinical diagnosis was performed based on the presence of two or more of the diagnostic criteria proposed by the NIH Consensus Development Conference (5). The diagnosis of NF1 was established in patients who had two or more of the following NIH criteria: Six or more cafe-au-lait macules (one of them must be greater than 5 mm and 15 mm, prepubertal and post-pubertal respectively); two or more neurofibromas or 1 plexiform neurofibroma; freckling in the axillary or inguinal regions; optic glioma; two or more Lisch nodules; a distinctive osseous lesion (sphenoid dysplasia or tibial pseudarthrosis); and a first-degree relative with neurofibromatosis type 1.

Genetic Analysis

Genomic DNA was extracted from whole blood samples using an automated method (RSC whole blood DNA kit) in the Maxwell® 16 (Promega Corporation, Madison, WI). NF1 (57 coding exons, NM_000267.3) amplicons were designed using the AmpliSeq Designer software (Life Technologies, CA, USA), targeting the complete coding sequence of the NF1 gene by 120 amplicons. The design target coverage was 99.49%.

Amplicon library was prepared using the Ion AmpliSeq Library Kit Plus, Xpress Barcode Adapters 1-96 Kit (Thermo Fisher Scientific), then pooled together using Qubit 1X dsDNAhs Assay kit and Qubit 4 Fluorometer (Thermo Fisher Scientific). Emulsion PCR, and Ion Sphere Particles enrichment were carried out in the Ion Chef System, then loaded into an Ion 530 chip. NGS was performed via Ion 510 & Ion 520 & Ion 530 kit-Chef (Thermo Fisher Scientific). Data were processed using Ion Torrent Suite Software (Thermo Fisher Scientific) and Ion Reporter Software (Thermo Fisher Scientific).

The Human Gene Mutation Database (HGMD) (7) and Leiden Open Variation Database (LOVD) were used to determine whether a variant was novel or not. Several prediction algorithms, including SIFT (http://sift.jcvi.org), Polyphen2 (http://genetics.bwh.harvard.edu/pph2/), Human Splicing Finder (http://www.umd.be/HSF/) and Mutation taster (http://www.mutationtaster.org) were used to determine any damaging effects on the protein. The Genome Aggregation Database (https://gnomad.broadinstitute.org) was used to estimate the minor allele frequency score.

Nomenclature of the variants was based on the NM_000267.3 (NCBI transcript number). Variants were reviewed using dbSNP (http://www.ncbi.nlm.nih.gov/projects/SNP/). A minimum 30X coverage for all bases was accepted for a reliable variant calling. Ion Reporter version 5.0 software was used to annotate variants. Integrated Genomics Viewer (http://software.broadinstitute.org/software/igv/) was used for visual assessment of the revealed variants.

Novel variants are reported to the Human Genome Variation Society guidelines and checked by using Mutalyzer tool (https://mutalyzer.nl/about). All variants were classified by using The American College of Medical Genetics and Genomics (ACMG) guidelines criteria (8). Some variants were validated with direct capillary sequencing that was performed by using the BigDyeTerminator kit v3.1 (LifeTechnologies, Darmstadt, Germany) and an automated capillary sequencer (3500xl Genetic Analyzer, Life Technologies). The obtained sequence data were analyzed using the Seq-Scape (Ver. 2.1) program (Applied Biosystems).

Results

Clinical Manifestations of the Patients

We reviewed the clinical data in 73 of the 92 patients. The frequencies of clinical features are sorted by age ranges in Table I. The median age was 8 years. 42.5% of patients were familial and 57.5% were sporadic cases. Seventy-one (97.2%) patients were suffering from café-au-lait spots. Axillary or inguinal freckling was present in 27 patients (36.9%). The other common skin manifestation were cutaneous neurofibromas that accounted for 19.1% (14/73) of cases. Optic glioma was identified in 6.8% (5/73) of the patients. Hamartomas were detected with magnetic resonance imaging in 22 of the patients (30.1%) (Table II).

Characterization of the NF1 Variants

Seventy-six variations of which two were probably somatic were detected in 92 patients from unrelated families. The identified variants were as follows: 36 frameshift variants (41.9%) resulting from small insertions, deletions or indels; 27 non-sense variations (31.4%); 14 missense variants (16.2%); 6 splicing alterations (7%), and 3 in-frame deletions or indels (3.5%) (Figure 1).

Twenty-seven (35%) of the variants were novel and had not been previously reported in HGMD or LOVD. The c.2446C>T, c.3826C>T, c.5839C>T, c.2546dupG showed familial segregation. The c.5107C>T (15%), c.3721C>T (20%) variants were detected at lower fraction percentages; 15% and 20% respectively. The c.2033dupC, c.2446C>T, c.3525_3526delAA, c.3826C>T, c.4084C>T, c.5546G>A, c.6792C>G variations were identified in more than one unrelated family (Table III). Only one of the variations (c.7674G>A) was interpreted as a variant of unknown origin (VUS) based on ACMG criteria. The other three were likely pathogenic and 23 of them were pathogenic. 24 variants produced truncated protein as a result of premature stop codon, which is a significant indication of their pathogenicity. All novel variants were predicted to be deleterious by at least one in-silico analysis.

Discussion

As NF1 is the one of the most common inherited disorders and NF1 is one of the largest genes in the human genome, a great number of variations (over 3,000) have been reported in HGMD to date. Even though previously reported pathogenic variation types show diversity, most of them cause severe truncated gene products (9). Most of the pathogenic variants (93%) are small nucleotide alterations (including non-sense, missense, insertion or deletion) and splicing variants. Intragenic deletions/duplications (2%) and microdeletions (5%) are rarely detected (10). In this study, we identified 76 different mutations in 92 families. Most of the variant types are frameshift, non-sense and splice-site, similar to the literature (10-12). The missense variation rate was relatively high (14 variants, 16%) in accordance with similar studies (11-13). There were some conflicting pathogenicity variants such as c.7674G>A and c.2764G>A which were interpreted as VUS based on ACMG criteria. The patient with c.7674G>A had a typical NF1 phenotype, which was a strong indicator showing the variant’s pathogenicity. However, the case with c.2764G>A variant did not reach the optimum clinical registry.

There is not a well-recognized hot spot region in the NF1 gene (14-16). In this study, we found only one recurrent variant (c.6792C>G) in two unrelated families. It did not show significant variant aggregation in any exon (Figure 2). Although we observed relative variant density in exon 13 according to the exon length (Figure 2), it was insufficient to speculate that exon 13 is a hotspot region. Moreover, the variant frequency was lower both in the first and last few exons which code the constitutional amino acids. This can be attributed to the variants on the distal part of the gene being less effective in causing the NF1 phenotype.

Six functional domains were determined on the NF1 protein: The cysteine and serine rich domain (CSRD; exons 1-22), the tubulin binging domain (TBD; exons 22-27), the domain responsible for interactions with RAS and GTP hydrolysis (GRD; exons 27-34), the bipartite lipid binging domain (first part) (Sec14; exons 35-36), the bipartite lipid binging domain (second part) (PH; exons 35-36), and the carboxyl-terminal domain (CTD; exons 37-52) (17). The distribution of the variants according to functional domains were CSRD (31/76-40.7%), TBD (9/76-11.8%), GRD (11/76-14.4%) Sec 14 and PH (5/76-6.5%), and CTD (20/76-26.3%). We did not find any variants outside of the functional domains.

Our results did not reveal any clear relationships between specific NF1 variants and phenotypes. Furthermore, no complete well-known genotype-phenotype correlation has been reported in the literature to date (12,18-20). Only three clear correlations of clinical significance have been identified in particular pathogenic NF1 variants. NF1 whole-gene deletions are related with early-onset presentation of cutaneous neurofibromas, cognitive abnormalities, somatic overgrowth, and dysmorphic facial features (21,22). The c.2970-2972delAAT variant does not cause cutaneous or surface plexiform neurofibromas (23). Any of the c.5425C>T or c.5425C>A or c.5425C>G variants are related with café-au-lait spots, learning disabilities, short stature, and pulmonic stenosis but not cutaneous neurofibromas (24,25). However, C16 (13 years of age) with the c.2970-2972del variant had neurofibromas, which is a late-onset feature of NF1.

A novel c.1637dupT variation was detected in C85 (20 years of age) who had isolated pilocytic astrocytoma without cutaneous findings. The case of C11 with the c.1721+3A>C variation only had clinical signs of neurofibromas without the accompanying café-au-lait spots. Ocular manifestations such as Lisch nodules and optic nerve glioma were determined in 23% and 6.8% of all cases, which are lower frequencies in comparison with the literature (26,27). Additionally, the frequency of neurofibromas were lower than in the literature.

Mosaic NF1 variants cause mild forms of the NF1 phenotype (28). We detected two different variations with low variation fraction in C87 (c.3721C>T) and C90 (c.5107C>T). Sanger confirmations of these variants were consistent with NGS data. However, a molecular analysis of a second tissue could not be performed in these patients. Both of these patients had classical NF1 symptoms without family history and segmental involvement. We classified both of these patients as mosaic generalized NF1.

Conclusion

In conclusion, NF1 genetic analysis was a supporting tool for the atypical presentation of NF1 cases especially in the prepubertal period. Additionally, genetic analysis before pregnancy provides preimplantation and prenatal genetic diagnosis for families with NF1.

Ethics

Ethics Committee Approval: The study was approved by the Ethical Committee of the Gaziantep University Medical Faculty (approval number: 65587614-774.99-291, date: 04/10/2017).

Informed Consent: Peripheral blood samples were obtained after taking informed consents from all participants or legal guardians of children under the age
of 18.

Peer-review: Externally peer-reviewed.

Authorship Contributions

Concept: E.K., Design: E.K., Data Collection or Processing: H.M.A., E.K., Analysis or Interpretation: H.M.A., E.K., Literature Search: H.M.A., E.K., Writing: H.M.A., E.K.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.