Dermal neurofibromas (dNFs) are benign tumors of the peripheral nervous system typically associated with Neurofibromatosis type 1 (NF1) patients. Genes controlling the integrity of the DNA are likely to influence the number of neurofibromas developed because dNFs are caused by somatic mutational inactivation of the NF1 gene, frequently evidenced by loss of heterozygosity (LOH). We performed a comprehensive analysis of the prevalence and mechanisms of LOH in dNFs. Our study included 518 dNFs from 113 patients. LOH was detected in 25% of the dNFs (N = 129). The most frequent mechanism causing LOH was mitotic recombination, which was observed in 62% of LOH-tumors (N = 80), and which does not reduce the number of NF1 gene copies. All events were generated by a single crossover located between the centromere and the NF1 gene, resulting in isodisomy of 17q. LOH due to the loss of the NF1 gene accounted for a 38% of dNFs with LOH (N = 49), with deletions ranging in size from ∼80 kb to ∼8 Mb within 17q. In one tumor we identified the first example of a neurofibroma-associated second-hit type-2 NF1 deletion. Analysis of the prevalence of mechanisms causing LOH in dNFs in individual patients (possibly under genetic control) will elucidate whether there exist interindividual variation. Hum Mutat 32:78–90, 2011. © 2010 Wiley-Liss, Inc.

Neurofibromatosis type 1 (NF1) is an autosomal dominant disorder affecting 1:3,500 individuals. Disease expression is highly variable and complications are diverse. However, currently there is no specific treatment for the disease. NF1 is caused by mutations in the NF1 gene, approximately 2.1% of constitutional mutations identified in our population are deep intronic mutations producing the insertion of a cryptic exon into the mature mRNA. We used antisense morpholino oligomers (AMOs) to restore normal splicing in primary fibroblast and lymphocyte cell lines derived from six NF1 patients bearing three deep intronic mutations in the NF1 gene (c.288+2025T>G, c.5749+332A>G, and c.7908-321C>G). AMOs were designed to target the newly created 5' splice sites to prevent the incorporation of cryptic exons. Our results demonstrate that AMO treatment effectively restored normal NF1 splicing at the mRNA level for the three mutations studied in the different cell lines analyzed. We also found that AMOs had a rapid effect that lasted for several days, acting in a sequence-specific manner and interfering with the splicing mechanism. Finally, to test whether the correction of aberrant NF1 splicing also restored neurofibromin function to wild-type levels, we measured the amount of Ras-GTP after AMO treatment in primary fibroblasts. The results clearly show an AMO-dependent decrease in Ras-GTP levels, which is consistent with the restoration of neurofibromin function. To our knowledge this is the first time that an antisense technique has been used successfully to correct NF1 mutations opening the possibility of a therapeutic strategy for this type of mutation not only for NF1 but for other genetic disorders.

Malignant peripheral nerve sheath tumors (MPNSTs) are soft-tissue sarcomas that can arise either sporadically or in association with neurofibromatosis type 1 (NF1). These aggressive malignancies confer poor survival, with no effective therapy available. We present the generation and characterization of five distinct MPNST orthoxenograft models for preclinical testing and personalized medicine. Four of the models are patient-derived tumor xenografts (PDTX), two independent MPNSTs from the same NF1 patient and two from different sporadic patients. The fifth model is an orthoxenograft derived from an NF1-related MPNST cell line. All MPNST orthoxenografts were generated by tumor implantation, or cell line injection, next to the sciatic nerve of nude mice, and were perpetuated by 7–10 mouse-to-mouse passages. The models reliably recapitulate the histopathological properties of their parental primary tumors. They also mimic distal dissemination properties in mice. Human stroma was rapidly lost after MPNST engraftment and replaced by murine stroma, which facilitated genomic tumor characterization. Compatible with an origin in a catastrophic event and subsequent genome stabilization, MPNST contained highly altered genomes that remained remarkably stable in orthoxenograft establishment and along passages. Mutational frequency and type of somatic point mutations were highly variable among the different MPNSTs modeled, but very consistent when comparing primary tumors with matched orthoxenografts generated. Unsupervised cluster analysis and principal component analysis (PCA) using an MPNST expression signature of ~1,000 genes grouped together all primary tumor–orthoxenograft pairs. Our work points to differences in the engraftment process of primary tumors compared with the engraftment of established cell lines. Following standardization and extensive characterization and validation, the orthoxenograft models were used for initial preclinical drug testing. Sorafenib (a BRAF inhibitor), in combination with doxorubicin or rapamycin, was found to be the most effective treatment for reducing MPNST growth. The development of genomically well-characterized preclinical models for MPNST allowed the evaluation of novel therapeutic strategies for personalized medicine.