Journal of Neuroscience Research · 2001 · 7 citations
Neurofibromatosis type 1: II. Answers from animal models
AbstractNeurofibromatosis type 1 (NF1), or von Recklinghausen's disease, is a complex disease resulting from mutations in the Nf1 gene. Moreover, the product of the NF1 gene neurofibromin also plays a very important role(s) in various aspects of the normal development and physiology of the organism. Although NF1 has been recognized for centuries, it is only relatively recently that there has been progress in the identification, isolation, and characterization of the NF1 gene and its protein product, neurofibromin (Seizinger et al., 1987; Fountain et al., 1989; Wallace et al., 1990; Marchuck et al., 1991). Nevertheless, progress in understanding the molecular regulation(s) and in identifying the molecular and cellular targets underlying the NF1 disease has lagged. There is much to be learned about the regulation(s) of the Nf1 gene expression and the function(s) and molecular interactions of neurofibromin. The role of neurofibromin in producing the multiple symptoms of the disease in the various cell types remains an enigma. Furthermore, there is still some controversy regarding the primary cellular target that in NF1 leads to neurofibroma and other NF1-associated symptoms and signs. This is the second part of a two-part review on NF1. The first part includes a general overview of the NF1 disease, the NF1 gene, and the genetic and cellular basis of the NF1 disease, with reference to controversial issues that hamper the full understanding of the disease and consequently hinder the efforts to find a cure (Lakkis and Tennekoon, 2000). In this part, the focus is on the animal models of NF1 disease, and we review both in vivo studies on these animals and in vitro studies using cells derived from the Nf1−/− mice. Although these studies have helped us to address some unanswered questions and enhance our knowledge about the molecular and cellular basis of NF1, additional questions have arisen that require further studies. Genetic manipulations in animals, especially in the mouse, have provided very good models to study and improve our understanding of human genetic diseases. The disruption of the Nf1 gene in the mouse is no exception. The targeted disruption of the Nf1 gene in the mouse (Jacks et al., 1994; Brannan et al., 1994) has not only provided new insights into the role of this gene in NF1 pathogenesis but has also revealed new and unexpected roles for NF1 gene in normal growth and development. Although the heterozygous mice appeared, in general, to be normal and did not display the classical symptoms of NF1 disease, Jacks et al. (1994) observed that, as these mice aged, they demonstrated a greater predisposition to a variety of tumor types, including lymphoma, lung adenocarcinoma, hepatoma, and fibrosarcoma. Furthermore, some of these mice developed tumors that are characteristically seen in human NF1, such as neurofibrosarcoma, pheochromocytoma, and myeloid leukemia. Cytological examination of tumor material showed that the wild-type Nf1 allele was lost in approximately half of the tumors, supporting the view that loss of heterozygosity (LOH) is required for tumorigenesis, thus making these mice a potential model for human disease. On the other hand, the homozygous mutant mice display a variety of developmental abnormalities. The most critical is a developmental cardiac defect that causes major circulatory dysfunction, leading to midgestation lethality (between days 12.5 and 14 of gestation). The mutant mice display generalized edema with systemic vascular congestion and signs of hemorrhage, all indicative of cardiac failure. Anatomical and histological analyses of the mutant hearts reveal many abnormalities, including double-outlet right ventricle (DORV) and ventricular septal defect (VSD) in addition to a dramatic enlargement of the endocardial cushions in the outflow tract in the atrioventricular region. The enlarged endocardial cushions most likely represent the critical abnormality, causing blockage of the inflow of circulating blood to the heart and causing the ultimate lethality. This endocardial cushion abnormality is caused by Ras up-regulation because of the absence of normal Nf1 gene product (Lakkis and Epstein, 1998), suggesting that this developmental process is regulated by the Nf1-GTPase activity directed towards Ras. This is surprising, because abnormalities of the heart are not a common clinical feature of NF1. There are, however, reports demonstrating the occurrence of cardiac abnormality in some NF1 patients (Lin and Garver, 1988; Tassabehji et al., 1993). In addition, there are several reports showing that visceral organs, including renal, hepatic, and skeletal muscle systems, are developmentally delayed (Brannan et al., 1994). They also consistently note hyperplasia of the prevertebral and paravertebral sympathetic ganglia cell, which originate from the neural crest cells. These observations suggested a new, previously unknown role for the NF1 gene during development and organogenesis. Additional examination of the brains of heterozygous mice reveals the presence of increased numbers of neural cells expressing high levels of glial fibrillary acidic protein (GFAP), a marker for astrocytes (Rizvi et al., 1999). Furthermore, a significant increase in GFAP-positive astrocytes is also observed in the hippocampus of about 60% of Nf1 heterozygous mice. The relevance of this observation stems from the finding that brains of human NF1 patients demonstrate increased expression of GFAP, indicative of astrocytosis, and the greater risk for benign astrocytic tumors (Nordlund et al., 1995). What is puzzling is that neither human nor mouse astrocytes normally express detectable amounts of neurofibromin, so the relationship of NF1 gene expression to astrocytosis is unclear (Daston et al., 1992; Nordlund et al., 1995). Moreover, mice expressing lower amounts of neurofibromin, as seen in Nf1 heterozygous mice, exhibit learning and memory difficulties. The Nf1-deficient mice do not perform as well as their wild-type littermates in the Morris water maze tests, a task of spatial learning (Silva et al., 1997). This suggests a role for neurofibromin in cognitive functions of the brain and is likely to be the correlate for the neurobehavioral and learning disabilities seen in about 30% of children with NF1 (Riccardi, 1981; North et al., 1995; Ozonoff, 1999). Other nervous system disorders in Nf1-deficient mouse embryos include cranial neural tube defect and exencephaly, both with incomplete penetrance (∼12.5%) and variable morphological expressivity. There are also visible brain structural defects in some nonexencephalic Nf1−/− embryos. These abnormalities show a gender bias, being more prevalent in female embryos (Lakkis et al., 1999). Gender bias of neurological disorders associated with NF1 has been reported in some cases. For example, in a clinical study of NF1 children with neurobehavioral abnormalities and learning disabilities, the neurobehavioral profile is indicative of dysfunction of frontal/subcortical brain regions. This frontal/subcortical neurobehavioral profile is more consistently observed in affected females. In addition, megalencephaly was observed among only NF1 females (Chapman et al., 1996), and morphometric analysis revealed significantly larger brain volumes, especially in the white matter of female NF1 patients (Said et al., 1996). The incidence and magnitude of neural tube defects in mice are increased in Nf1:p53 and Nf1:pax-3 double mutants (Vogel and Parada, 1998; Lakkis et al., 1999). Further in vitro studies performed on cells derived from the Nf1-deficient animals indicated that several cell types, including Schwann cells, fibroblasts, neurons, and myeloid cells, were abnormal. Nevertheless, some important questions in NF1 pathogenesis are whether Nf1 mutations that contribute to neurofibroma formation occur primarily in Schwann cells, fibroblasts, or in both cell types and whether the Ras pathway is involved in the pathogenesis. Studies on Schwann cells isolated from the Nf1−/− embryos showed that these cells behave very similarly to v-Ras-transformed rat Schwann cells in terms of morphological changes, abnormal growth, and response to β-neuregulin (GGF2). In addition, Schwann cells obtained from homozygous (Nf1−/−) and heterozygous (Nf1+/–) mice demonstrated elevated levels of Ras-GTP (Kim et al., 1995), implicating neurofibromin as a major Ras-GAP protein in Schwann cells. In another study (Kim et al., 1997), Schwann cells isolated from either homozygous or heterozygous for Nf1 mice displayed angiogenic and invasive properties similar to those of Schwann cells isolated from human neurofibromas. Nevertheless, Nf1 mutation in Schwann cells did not induce cell proliferation and hyperplasia, suggesting that other somatic events in addition to the Nf1 mutations are necessary for inducing Schwann cell proliferation and hyperplasia (Kim et al., 1995, 1997). Schwann cells from Nf1−/− mouse embryos as well as H-Ras-transformed Schwann cells show increased expression of the major myelin glycoprotein P0 (Rosenbaum et al., 1999; personal observations). This would indicate a role for neurofibromin in Schwann cell differentiation via Ras-signaling pathway (Gutmann et al., 1993; Rosenbaum et al., 1999). On the other hand, fibroblasts from Nf1-deficient mice also displayed abnormal behavior compared to their wild-type counterparts. Perineurial cells, a form of specialized fibroblasts that surround axon-Schwann cell groups, from Nf1-deficient mice exhibit abnormal properties that include increased proliferation and failure to form normal nerve fascicles in vitro (Rosenbaum et al., 1995). Moreover, Nf1-deficient fibroblasts showed cell-autonomous abnormalities with increased collagen deposition in vitro, a feature of tumors when up to 70% of neurofibromas are composed of collagen (Peltonen et al., 1986, 1988; Atit et al., 1999). In another study, skin fibroblasts in heterozygous Nf1 mice proliferated beyond their normal time during the maturation phase of wound healing, suggestive of the fact that the Nf1 gene product is an important regulator of the fibroblast response(s) following injury (Atit et al., 1999). This is in agreement with a previous study that demonstrated up-regulation of neurofibromin expression during normal human wound healing (Yla-Outinen et al., 1998). Extrapolation from these studies suggests that the lack of neurofibromin affects fibroblast behavior by increasing the proliferation capacity and collagen deposition in a manner seen in neurofibroma formation in NF1 patients. The effects on fibroblasts are most likely to occur in a Ras-independent manner (Atit et al., 1999). An additional biological role(s) for neurofibromin in neurons is suggested from results of primary neural crest- and placode-derived neuron cultures from Nf1-deficient mouse embryos. Unlike their wild-type counterpart, these neurons were able to survive and differentiate without exogenous neurotrophic factors (Vogel et al., 1995), indicating a role for neurofibromin as a negative regulator of neurotrophin-mediated signaling in survival and differentiation of peripheral neurons. Furthermore, in double-mutant mouse models of Nf1 and p53, it appears that both Nf1 and p53 act cooperatively to enhance the neurotrophin-independent neuron survival and proliferation (Vogel and Parada, 1998). In a different cell type, partial deficiency of Nf1 in hematopoietic cells causes aberrant growth of these cells, and myeloproliferative symptoms ensue as a result of increased sensitivity of the cells to granulocyte-macrophage colony-stimulating factor (GM-CSF; Bollag et al., 1996; Largaespada et al., 1996). Neurofibromin also plays a critical role in regulating the proliferation and survival of myeloid progenitor cells in response to cytokines such as interleukin (IL)-3, stem cell factor (SCF), and GM-CSF by modulating Ras signaling (Zhang et al., 1998). The relevance of these observations is that some patients with NF1 are predisposed to developing myeloproliferative diseases such as juvenile chronic myelogenous leukemia. Nevertheless, insofar as mice with heterozygous inactivation of the Nf1 gene did not develop the major symptoms of NF1 disease, especially neurofibroma, and the homozygous mice were embryonic lethal, the Nf1 knock-out mice were not considered a perfect model for the human disease. Therefore, another mouse model was established that appears to be a better representation of human NF1 disease with respect to development of more features of the NF1 disease, i.e., neurofibroma and neurofibrosarcoma (Cichowski et al., 1999). Chimeric mice with some Nf1−/− homozygous mutant cells were generated, overcoming the embryonic lethality of the germline mutation. These chimeric mice survive to birth, and all develop multiple neurofibromas. The neurofibromas are predominantly composed of Schwann cells that contain a homozygous mutation of Nf1 gene. From these and previously cited studies, it appears that LOH (homozygous mutation in both Nf1 alleles) is a prerequisite in the formation of neurofibroma development. They also provide good evidence that Schwann cells represent the primary cell type involved in development of neurofibroma. Another mouse model was also developed for malignant peripheral nerve sheath tumor (MPNST) and soft-tissue sarcoma by generating mice that carry combined mutations in both Nf1 and p53 genes in cis and trans configurations (Cichowski et al., 1999; Vogel et al., 1999). These mice develop soft-tissue sarcoma and classical MPNST in neural crest-derived tissues, characteristic of some NF1 patients. The tumors occur at an especially increased rate in mice carrying these mutations in cis configuration. These animal models suggest that a mutation in p53 gene, in addition to that in NF1, is required for the malignant transformation of cells of neural crest origin. Further insight into the function of the NF1 gene is indicated from studies of the Nf1 mutations in Drosophila melanogaster. The heterozygous flies appear normal; however, unlike Nf1-deficient mice, the homozygous Drosophila NF1 mutants are viable and fertile, although they are smaller than their wild-type counterpart at all developmental stages. Furthermore, these flies do not demonstrate any Ras-1 signaling abnormalities. The size defect in the mutant flies can be rescued by expression of activated cAMP-dependent protein kinase-A but not by manipulating Ras signaling pathway (The et al., 1997). This work on Drosophila suggests an alternative mechanism for Nf1 action in which NF1 and PKA interact in the same signaling pathway that regulates growth of the animal. Moreover, these flies display abnormal behavior compared to wild-type flies in that NF1 is essential for the cellular response to the neuropeptide pituitary adenylyl cyclase-activating polypeptide (PACAP38) at the neuromuscular junction. PACAP38 induces potassium channel currents through a unique pathway involving NF1 and PKA (Guo et al., 1997). The available animal models are not an exact representation of NF1 disease, because the animals do not exhibit the plethora of symptoms seen in NF1 disease. These models have increased our knowledge of the role of the NF1 gene product in the pathogenesis of NF1 and other related tumors; have helped us to understand the biology of Nf1 gene product; and have revealed new, previously unknown biological consequences of the alteration of the Nf1 gene product. Additional animal models such as conditional inactivation of the Nf1 gene in a tissue-specific manner should provide valuable information. We thank Lori Mendham for editorial comments. The investigators are supported by funds from the National Multiple Sclerosis Society and the National Institutes of Health (NINDS).
https://doi.org/10.1002/jnr.1142