Localized mosaic neurofibromatosis type 1

Summary

Background: Neurofibromatosis type 1 is one of the more common rare disorders, and its atypical/segmental or mosaic forms are underdiagnosed. Thus far, only a few dozen cases of localized mosaic neurofibromatosis have undergone combined germline and somatic genetic testing for the NF1 gene. Methods: A 65-year-old female patient was referred to our center for multiple neurofibromas on her right shoulder with a clinical diagnosis of localized mosaic neurofibromatosis. One of the neurofibromas was surgically removed. Massively parallel sequencing and multiplex ligation-dependent probe amplification were utilized to identify the germline and somatic variants in the NF1 gene. Results: Heterozygous pathogenic NF1 gene variant c.7549C>T and multiple heterozygous intragenic NF1 gene deletions were detected in the DNA taken from the shoulder neurofibroma but not DNA from blood leukocytes or buccal smear. Conclusion: Germline and somatic genetic testing in localized forms of neurofibromatosis are advisable since it facilitates proper genetic counseling regarding risks to offspring who could inherit a germline pathogenic variant. Another important point to consider is cancer surveillance, which is often underutilized in mosaic forms of neurofibromatosis.

Keywords:

Next-generation sequencing – case report – segmental neurofibromatosis – somatic mosaicism – NF1

Introduction

Localized mosaic neurofibromatosis (LMN), also known as segmental neurofibromatosis or type V neurofibromatosis according to Riccardi´s classification [1], is one of the least common genodermatoses of the neurofibromatosis family. LMN arises due to post-zygotic somatic mosaicism in the NF1 gene [2] and is a member of the mosaic neurofibromatosis 1 (NF1; MIM: 162200) group. The preferred term for the condition is “localized mosaic NF1” [3] – as opposed to (a) mosaic NF1, which is not confined to a specific segment of the body, or (b) germinal NF1, which affects only one segment by pure chance. The terms “segmental neurofibromatosis” and “mosaic neurofibromatosis” are used loosely, worsening the nosological issue. The classic definition of LMN describes the condition as café au lait macules (CALM) and/or neurofibromas present in only one unilateral segment of the body, usually superficially [4]. The distribution of CALM generally follows Blaschko lines [5]. The development of tumor lesions and CALM in NF1 and LMN follows Knudson’s two-hit hypothesis, i.e., two different molecular lesions need to be present to initiate tumorigenesis [6].

Because of its discreet clinical presentation, many LMN cases often go undiagnosed. Hundreds of adult LMN cases [7] and dozens of pediatric LMN cases [8] have been clinically described in the literature. However, only a few individuals have undergone genetic testing, e.g., only 15 out of the adult patients mentioned by García-Romero et al [7] underwent molecular genetic testing for the presence of NF1 mosaicism, eight patients were tested by Marwaha et al [9], two by Messiaen et al [2], and another five cases were reported individually [10–14].

Here we present a sporadic case of LMN in a female patient with multiple cutaneous neurofibromas on her shoulder having a heterozygous pathogenic somatic variant in the NF1 gene, which was identified using massively parallel sequencing (MPS). Additionally, multiple heterozygous intragenic NF1 deletions were detected using multiplex ligation-dependent probe amplification (MLPA). Our case demonstrates the importance of molecular analysis for follow-ups regarding cancer risk and proper genetic counseling concerning reproductive options in affected families.

Methods

A 65-year-old female patient was referred to our center regarding multiple neurofibromas on her right shoulder (Fig. 1). Two similar nodular growths were also present on her nose, although neither were histologically nor genetically examined. No CALM or other NF1 related signs were detected at this stage. The family history was unremarkable, and the patient had two healthy children with no apparent signs of NF1. Lisch nodules were not detected on ophthalmological evaluation. Subsequently, skin excision of one of the nodules was performed, and a 15 x 10 x 5 mm tissue sample, including a 5 x 5 mm suspect neurofibroma, was indicated for histological evaluation. The histological examination confirmed the neurofibroma diagnosis (Fig. 2).

DNA for molecular genetic diagnostics was isolated from formalin-fixed paraffin-embedded (FFPE) tissue biopsy samples from two locations, i.e., the neurofibroma itself and some of the healthy skin adjacent to the neurofibroma. We also isolated DNA from the patient’s peripheral blood lymphocytes and buccal smear cells using standard DNA extraction procedures with Magcore^Tm assays.

Initially, we analyzed DNA from peripheral blood lymphocytes to evaluate germline pathogenic variations. Targeted MPS of the neurofibromin gene NF1 (MIM: 613113) was performed on a MiSeq platform, and data were analyzed using SOPHiA DDM software. Peripheral blood lymphocyte DNA revealed no single nucleotide variants (SNV) or copy number variants (CNV) in the NF1 gene (Classes IV-V). Next, we analyzed the DNA extracted from the excised neurofibroma using the same analytical approach. Sequencing data were analyzed using FinalistDX bioinformatics software. The neurofibroma sample was also analyzed using MLPA (kits P081-NF1 and P082-NF1).

Results

A heterozygous NF1 gene pathogenic variant of interest was found in 13% of the NF1 reads, i.e., NM_001042492.2: c.7549C>T, p.(Arg2517*). The variant was annotated as class 5 (pathogenic) according to ACMG criteria (evidence: PVS1 very strong, PP5 strong, PM2 moderate, PP3 supporting). The variant´s reference SNP cluster ID is rs866445127. This variant has been previously described as pathogenic in both sporadic and familial cases of NF1 and as a “somatic” variant in biopsies [15]. Furthermore, using MLPA, we detected a decrease in peak heights corresponding to exons 3, 5, 6, 7, 9, 11, 15, 16, 21, 23, 24, 25, and 56 of the NF1 gene. This could represent mosaic somatic heterozygous deletions in a subset of sample cells. This finding is consistent with Knudson’s “two-hit” hypothesis. Deletion-based loss of heterozygosity is a common finding in NF1 related neoplasias [16]. Therefore, we concluded that these findings were causative for the observed LMN in our index case.

The SNV and CNVs were not found in DNA extracted from blood lymphocytes and buccal smears, which were analyzed using Sanger DNA sequencing and MLPA.

We tried to analyze the DNA from the adjacent non-neoplastic tissue resected from the neurofibroma sample. Unfortunately, DNA extraction from the FFPE block did not yield adequate amounts of DNA of sufficient quality, and the analysis could not be completed successfully.

Discussion and conclusion

Here we present a case of LMN, where both germline and somatic variations were analyzed using MPS and MLPA. The biallelic NF1 pathogenic variants in only the neoplastic tissue strongly support the LMN diagnosis. Since mosaic forms of NF1 can afflict the gonads, it represents a risk of NF1 to the offspring of patients. This makes it crucial to pursue molecular genetic diagnostics so that proper genetic counseling can be provided [17]. When the mosaic form of NF1 is suspected, localized or not, investigations of peripheral blood lymphocytes will often fail to identify the causative variant since the post-zygotic variants are only harbored by a specific subset of the patient’s cells. In this regard, molecular genetic examination of other tissues should follow.

Reports identifying the pathogenic variants in LMN via MPS are still lacking. Ko et al [11] reported a patient diagnosed using a procedure similar to ours. García-Romero et al [7] described four mosaic NF1 patients that underwent testing of the affected tissue and blood lymphocytes; in one case, the variant was only found in the affected tissue, and in three cases, it was found in both tissues. Whether it was the localized form of the disease was not specified. Furthermore, Maertens et al [18] described another patient with mosaic NF1 (though not localized) in which different tissues were examined, including hair, urine, and a buccal smear, and the causative variant was found to varying degrees in the different tissues. In patients described by Marwaha et al. and Freret et al [9,10], both first- and second-hit variants were identified in diseased tissue but not in peripheral blood lymphocytes. In another patient described by Marwaha et al [9], unpigmented skin above a plexiform neurofibroma was examined, but no pathogenic variant was detected. The major limitation of most of these studies was that healthy tissue around the affected area was not tested; this assumes that the identified variant in the diseased tissue was present throughout the entire segment of the patient’s body, i.e., in both healthy and affected cells alike. Moreover, if only one variant is determined, this could lead to erroneous genetic counseling in risk assessment, i.e., when the variant found in the neoplasia is used in preimplantation/prenatal diagnostics to rule out the risk of NF1 due to gonadal mosaicism in the offspring.

Our data makes it impossible to conclude which one of the detected SNV or CNVs was the first or the second hit variant. We did not want to further and unnecessarily stress the patient by asking her to undergo additional multiple skin biopsies after her diagnosis was clinically established. Moreover, a “secondary” diagnostic/research strategy was beyond the scope of our standard clinical diagnostic approach. Additionally, we know that her children are unaffected.

Another critical point to consider is related to oncological prevention in neurofibromatosis. In this regard, the risk of neoplasms in LMN is similar to that in NF1 patients with the typical form of the disease [19]. Female carriers of germline NF1 pathogenic variants have a higher lifelong risk of breast cancer and thus should receive preventive care, e.g., regular mammographic screening. Since LMN skin lesions are often present on the thorax and the abdomen, we suggest oncological screening of female patients with LMN, similar to that for carriers of the germline NF1 pathogenic variant, i.e., using the newest NCCN guidelines [20], taking into account relevant family history.

Finally, our study limitations must acknowledge the substantial diagnostic overlap between localized mosaic NF1 and mosaic NF1 present in multiple areas. This feature complicates investigations, background literature research, and clinical/variant interpretation. Surprisingly, even though NF1 is one of the more common rare diseases (ref. ORPHA: 636), investigations of eventual post-zygotic mosaicism using DNA taken from various tissues remain relatively scarce. Hence, we believe that further specialized investigations of different tissues from LMN patients are warranted.

The Editorial Board declares that the manu script met the ICMJE “uniform requirements” for biomedical papers.
Redakční rada potvrzuje, že rukopis práce splnil ICMJE kritéria pro publikace zasílané do biomedicínských časopisů.

Martin Schwarz, MD
Department of Biology
and Medical Genetics
2nd Faculty of Medicine
Charles University
and Motol University Hospital
V Úvalu 84
150 06 Prague
Czech Republic
e-mail: martin.schwarz@lfmotol.cuni.cz
Accepted for review: 15. 11. 2021
Accepted for print: 18. 1. 2022