Exploring the genetic basis for clinical variation in neurofibromatosis type 1

KEYWORDS:

• Cancer
• glioma
• germline mutation
• brain tumor
• cancer predisposition syndrome

One of the challenges to implementing individualized approaches to patient care (precision medicine) is the relative paucity of predictive markers of disease development and progression. In this regard, the natural history of a particular disorder or the appearance of specific clinical features of that disease (e.g. cancer) in any given individual is dictated by a myriad of factors, ranging from genomic (e.g. single-nucleotide polymorphisms) and genetic (e.g. specific germline and somatic mutations) determinants to tissue (e.g. tumor microenvironment) and cellular (e.g. cell of origin) influences. Similarly, how a particular cancer responds to targeted therapies also reflects the interplay of these factors.

In order to unravel this complexity, it is necessary to dissect the contributions of each of these disease modifiers. This seemingly overwhelming task is partly simplified by studying human disorders that arise from a single gene. Each person with a monogenic condition shares a common genetic etiology and a spectrum of anticipated clinical features, which provide the foundations for more controlled studies aimed at defining the factors that underlie disease heterogeneity. Understanding the mechanisms by which these contributors create clinical variability also provides unprecedented opportunities to establish improved risk assessment strategies and to identify new therapeutic targets for future clinical trials.

Neurofibromatosis type 1 (NF1) is a common monogenic syndrome affecting one per 2500 individuals worldwide. Individuals with NF1 are prone to develop benign (neurofibromas and optic gliomas) and malignant (glioblastoma, breast cancer, malignant peripheral nerve sheath tumors, and leukemia) tumors as well as cardiovascular defects, autism, cognitive and motor delays, epilepsy, sleep disturbances, and skeletal abnormalities. At present, it is not possible to predict which child or adult will develop which of these medical problems. Moreover, even when clinical abnormalities arise, it is also not currently possible to accurately determine the natural history of that particular disease feature. Aside from previous radiation therapy (plexiform neurofibroma and optic glioma) and young age (<2 years old) or brain location (post-chiasmal optic glioma), few predictors of disease progression exist. These barriers foster a ‘reactive’ (anticipatory management) approach for a lifelong condition and limit our therapeutic options when deciding how to best treat a particular clinical problem when it arises. In this regard, treatment may be delayed in some cases, leading to devastating outcomes, while others, who might never require treatment, may be exposed to repeated sedated radiographic procedures.

However, recent studies using Nf1 genetically engineered mouse (GEM) models, where a single factor can be studied in isolation, have begun to provide important insights into the manners by which each of these factors operates to influence disease pathogenesis and progression. In people with NF1, epidemiologic and genomic investigations have revealed that race [1], sex [2], and single-nucleotide polymorphisms [3] each influence the risk of brain tumor development and progression in children with NF1. Leveraging Nf1 GEM strains that develop high-grade brain tumors (gliomas), several modifier loci have been identified that dictate the penetrance and location of gliomas, while GEM low-grade glioma models have begun to elucidate the impact of sex on optic glioma development, progression, and associated vision loss in mice [4–7].

Nf1 GEM models of plexiform neurofibroma and optic glioma have additionally revealed the critical importance of the cell of origin and the tumor microenvironment in cancer development and progression. In this regard, there are specific cell types that can give rise to these tumors during defined periods of mouse development [8,9]. For example, plexiform neurofibromas are dependent on bone marrow-derived cells (mast cells and macrophages) that produce key stromal factors to facilitate tumor-dependent and continued growth [10]. One of these paracrine circuits involves the chemokine KIT ligand, culminating in a clinical treatment trial using Imatinib [11]. While this trial resulted in few positive effects, further investigation on bone marrow-derived cells and the c-KIT-signaling axis is warranted. Similarly, mouse optic gliomas are maintained by immune system-like cells (microglia) and an analogous paracrine circuit (CCL5 chemokine) [12]. Future dissection of these stromal dependencies may lead to the implementation of complementary drug-targeting approaches.

Another understudied potential contributor to disease heterogeneity is the germline NF1 gene mutation. Individuals with NF1 are born with a germline mutation in the NF1 tumor suppressor gene located on chromosome 17q11.2. While a germline NF1 gene mutation characterizes the disorder, tumor formation requires somatic inactivation of the second NF1 allele. It should be recognized that stromal cells and neurons harbor only a germline NF1 gene mutation and do not exhibit biallelic NF1 loss. As such, the cells most critical for neuronal function and benign tumor maintenance (mast cells, macrophages, neurons, and microglia) are likely to be most affected by potential biological differences conferred by the germline NF1 gene mutation.

To date, there have been over 1300 different germline NF1 gene mutations documented. As the number of NF1 gene mutations identified has increased worldwide, several intriguing genotype–phenotype correlations have been reported. First, individuals with large (~1.4 Mb) genomic microdeletions, spanning the entire NF1 gene locus and neighboring genes, have an increased number of neurofibromas and are at an elevated risk for cardiac malfunction, skeletal anomalies, facial dysmorphism, and malignant tumor development [13]. Second, several groups have reported genotype–phenotype correlations involving two specific germline NF1 gene mutations in people who do not develop cutaneous neurofibromas (c.2970–2972_delAAT; c.5425C>T) [14,15]. This latter codon 1809 mutation is now the most frequently reported missense mutation in NF1, observed in 1% of all individuals with NF1 [15]. Finally, two groups have found clustering of 5ʹ-tertile NF1 gene mutations in individuals with NF1-optic glioma [16,17], which was not replicated in a third study [18].

Based on these findings, our laboratory employed NF1-patient-derived fibroblasts, induced pluripotent stem cells, and derivative neural progenitor cells to demonstrate that different NF1 germline mutations have distinct effects on NF1 protein (neurofibromin) expression and function [19]. Examination of these human cell types revealed that one-third of individuals with NF1 had <25% reductions in neurofibromin expression, while the remaining exhibited >70% reductions. Intriguingly, all of the NF1 patient-derived cell types, regardless of neurofibromin levels, had similar degrees of rat sarcoma (RAS) activation, the protein regulated by neurofibromin and responsible for controlling cell growth. In contrast, the levels of another protein regulated by neurofibromin (dopamine) strongly correlated with neurofibromin expression. These observations, made in human cells, suggest that not all germline NF1 gene mutations are equivalent.

To formally evaluate the impact of the germline NF1 gene mutation on disease pathogenesis, proof-of-principle studies were performed using Nf1 GEM strains harboring two distinct NF1 patient-derived germline Nf1 gene mutations. Both of these mutations are located in the 5ʹ-tertile of the NF1 gene but represent mutations observed in children with optic glioma (c.2041C>T; p.Arg681*) and individuals with spinal neurofibromas (c.2542G>C; p.Gly848Arg). In these experiments, mice were generated, such that the only difference was the germline Nf1 gene mutation, thus controlling for other variables (strain background, cell of origin, and timing of somatic Nf1 loss) [20]. Analogous to the results obtained using NF1 patient-derived cell types, mice with the Gly848Arg germline Nf1 gene mutation exhibited <40% reductions in neurofibromin expression relative to littermate controls, whereas those with the Arg681* germline Nf1 gene mutation expressed ~20% of control neurofibromin levels. When these germline Nf1 gene mutations were coupled with somatic Nf1 loss in neuroglial progenitor cells, only mice with the Arg681* germline Nf1 gene mutation developed optic gliomas. Interestingly, the optic gliomas in the Arg681* mutant mice were larger and exhibited more proliferation than mice in which the germline mutation was created by inserting a neomycin-targeting cassette into the murine Nf1 gene (conventional knockout mice). These conventional knockout mice exhibit ~50% of control neurofibromin expression and, therefore, both under- and overrepresent the effects of actual NF1 patient germline NF1 gene mutations.

Since optic gliomas in children with NF1 can cause visual impairment, the differential impact of these NF1 patient-derived germline NF1 gene mutations was also explored. Similar to the effects observed with optic glioma development and growth, only mice with the Arg681* germline Nf1 gene mutation exhibited reduced retinal ganglion cell survival and nerve fiber layer thickness, which again exceeded those observed with the conventional knockout strain.

Further examination of these mice revealed both cell-intrinsic and stromal effects of the germline Nf1 gene mutation. Whereas astrocytes from Gly848Arg mice exhibited little increase in proliferation following somatic Nf1 loss, those from Arg681* and conventional knockout mice had three- to fourfold greater growth in vitro. Moreover, the optic gliomas from Arg681* mice contained more microglia and microglia-produced CCL5 than those from conventional knockout mice, resulting in increased CCL5-mediated activation of the protein kinase B (AKT)-signaling intermediate in the neoplastic glial cells within the tumor. Together, these findings suggest that the impact of the germline NF1 gene mutation can operate at multiple levels to differentially dictate disease pathogenesis.

In summary, these early-phase preclinical studies provide compelling evidence for an effect of the germline Nf1 gene mutation on optic glioma formation and growth in NF1 and should prompt additional studies employing mice with actual NF1 patient mutations in tandem with patient-derived stem cells where the germline NF1 gene mutation is known. In this manner, it might be possible to define subgroups of patient biospecimens and GEM strains that respond to specific drugs and dosing schedules amenable to future translation to human clinical trials. Moreover, with the availability of accurate NF1 gene mutation testing methods, it now becomes possible to consider incorporating this clinical variable into risk assessment strategies for newly diagnosed children with NF1. Coupled with genomic predictors, patient sex, and other potential risk factors, the future care of individuals with NF1 may evolve into a more ‘proactive’ practice in which we have the ability to better predict disease feature development, clinical progression, and response to treatment.

Declaration of interest

This work was supported by the National Cancer Institute [1-R01-CA195692-01]; Alex’s Lemonade Stand Foundation for Childhood Cancer. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.