  |
 |
 |
 |
 |
|||
 |
|||||||
 |
|
|||||||||||||||
HuGE Review
|
This HuGE Review was published in the American Journal of Epidemiology 2000 Jan 1;151(1):33-40. NF1 Gene and Neurofibromatosis Type 1 by Sonja A. Rasmussen1 and J.M. Friedman2 (1)Centers for Disease Control and Prevention, Division of Birth Defects and Pediatric Genetics, Atlanta, GA Updated August 31, 1999 HuGE Review Internet Sites HuGE Review - NF1 Gene and Neurofibromatosis Type 1 Neurofibromatosis 1 (NF1), also known as von Recklinghausen disease, is caused by mutations of the NF1 gene, which is located at chromosome 17q11.2. The protein product, neurofibromin, has tumor suppressor activity and probably other functions as well. A wide variety of NF1 mutations have been found in patients with NF1. No frequently recurring mutation has been identified. No information is available currently on the distribution of NF1 mutations in different populations. Diagnosis of NF1 is based on established clinical criteria. The features used to make a diagnosis include those that are most frequent or distinctive: café-au-lait macules, discrete or plexiform neurofibromas, freckling in non-sun-exposed areas, optic glioma, Lisch nodules of the iris, sphenoid wing dysplasia, and long bone cortical thinning. Other important clinical manifestations of NF1 include learning disabilities, increased risk for certain malignancies, scoliosis, and hypertension. The presentation of many of these features is age-dependent. Average life expectancy of patients with NF1 is probably reduced by at least 10 to 15 years. Malignancy is the most common cause of death, but hypertension appears to be significantly associated with mortality. Primary prevention of NF1 complications is not currently possible. NF1 is an autosomal dominant condition. About 50% of cases result from new mutations. The NF1 mutation rate is among the highest observed in humans, with estimates ranging from about 1/7,800 to 1/23,000 gametes. Both somatic and germline mosaicism for NF1 mutations have been documented. NF1 is believed to be completely penetrant, but substantial variability in expression of features occurs, even in affected members of a single family. Most studies have not found an obvious relationship between particular mutations of the NF1 locus and the resulting clinical manifestations in a patient, but some possible allelic variants have been recognized. The wide variability of the NF1 phenotype, even in individuals with the same NF1 gene mutation, suggests that other factors are involved in determining the clinical manifestations, but the nature of these factors has not yet been determined. Laboratory testing for NF1 mutations is difficult. A protein truncation test is commercially available, but its sensitivity, specificity, and predictive value have not been established. Although mutation analysis or linkage can be used for prenatal diagnosis in some families, the value of such testing is limited because it cannot predict how severe the disease is likely to be. Presymptomatic molecular diagnosis is also possible in some instances but is rarely indicated clinically. No general population-based molecular studies of NF1 mutations have been performed. The prevalence of clinically-diagnosed NF1 ranges from 1/2,000 to 1/5,000 in most population-based studies. At this time, it appears that the benefits of population-based screening for clinical features of NF1 would not outweigh the costs of screening. The neurofibromatosis type 1 (NF1) gene is located at chromosome 17q11.2. NF1 and its protein product, neurofibromin, were characterized in 1990 (1, 2). The gene is large, spanning 350 kilobases of genomic DNA, and contains 60 exons (3). Neurofibromin belongs to a family of proteins known as GTPase activating proteins, which are negative regulators of the ras oncogene (4). Neurofibromin is believed to act as a tumor suppressor, but the protein has other functions as well. The proposed tumor suppressor function is supported by the findings of somatic "second hit" mutations of the NF1 gene in benign and malignant tumors from NF1 patients (5, 6). Of interest, three other genes (OMGP, EVI2B and EVI2A) lie within intron 27b of the NF1 gene and are transcribed in the opposite direction (7). The function of these embedded genes and their influence, if any, on the NF1 phenotype, are unknown. As of February 1999, the NF1 Genetic Analysis Consortium documented more than 240 different constitutional NF1 mutations in its database (http://www.nf.org/nf1gene/). Table 1 summarizes the types of mutations identified thus far. The majority of mutations lead to a truncated protein product; only about 10% involve amino acid substitutions and fewer than 2% are 3' untranslated region mutations. However, it should be noted that the types of mutations identified are largely dependent on the techniques used for mutation detection. This may result in an over-representation of mutation types that are more easily identified (i.e. large gene deletions) and an under-representation of those which may be more difficult to identify (i.e. mutations in the 3' untranslated region). None of the methods used for NF1 mutation detection are capable of identifying all mutation types. Mutations have been identified throughout the gene. While some mutations recur in different families, no true "hotspots" have been found in NF1. The most frequently recurring alteration is a nonsense mutation in exon 31 (R1947X) that accounts for 1-2% of NF1 mutations identified (14). At this time, no information is available on frequency of different mutations in different populations and ethnic groups. Clinical Features of NF1 Neurofibromatosis 1 (NF1), also known as von Recklinghausen disease, is the condition most commonly associated with NF1 gene mutations. Early discussions of NF1 referred to the condition as "neurofibromatosis" and included cases of the much less frequent condition, neurofibromatosis 2 (NF2). However, these conditions are both clinically and genetically distinct. The most characteristic lesions of NF2 are bilateral schwannomas on the vestibular portion of the eighth cranial nerve; such tumors are rarely seen in NF1 patients. NF2 results from mutations in the NF2 gene on chromosome 22. Despite advances in understanding of the molecular genetics of NF1, its diagnosis remains a clinical one, based on diagnostic criteria established by an NIH consensus conference (15, 16). A diagnosis of NF1 by these criteria requires the presence of two or more of the following: (a) six or more café-au-lait macules more than 5 millimeters (mm) in greatest diameter in prepubertal individuals and more than 15 mm in greatest diameter after puberty; (b) two or more neurofibromas of any type or one plexiform neurofibroma; (c) freckling in the axillary or inguinal regions; (d) an optic pathway tumor; (e) two or more Lisch nodules (iris hamartomas); (f) a distinctive osseous lesion, such as sphenoid wing dysplasia or thinning of the cortex of long bones (with or without pseudarthrosis); or (g) a first-degree relative (parent, sib, or child) with NF1 diagnosed by the above criteria. Some of these features, including café-au-lait spots, freckling in non-sun-exposed areas, and iris Lisch nodules, are not of clinical significance beyond their usefulness in making a diagnosis of NF1. Benign cutaneous and subcutaneous neurofibromas are present in nearly all patients with NF1 by adulthood, and their number in an individual varies widely from only a few to hundreds or more. While these lesions are primarily of cosmetic significance, they may be disfiguring and result in significant psychological distress. In contrast, about 15% of individuals with NF1 have plexiform neurofibromas (17). These tumors may extend into contiguous tissues, causing serious functional impairment and even death, and appear to be the site of malignant peripheral nerve sheath tumor development. Optic pathway tumors are observed in about 20 percent of children with NF1, but most do not cause ophthalmologic or other symptoms (18). Bony changes, such as pseudarthrosis, appear to occur in about 5 percent of cases (17). Often these changes are benign; however, some patients are severely affected, with long bone bowing leading to fracture, and in some cases, requiring amputation (19). Several other features are often associated with NF1, including macrocephaly, scoliosis, short stature, hypertension, and high-T2-signal-intensity lesions on magnetic resonance imaging of the brain (20). Most individuals with NF1 have normal intelligence, but 30% to 60% have learning disabilities (21). Individuals with NF1 also appear to be at increased risk for malignancy, but the magnitude of this increased risk is difficult to estimate, given the paucity of epidemiologic studies. In an investigation of a Danish cohort of 212 NF1 patients followed for 42 years, a relative risk of 4.0 (95% CI: 2.8-5.6) was observed for malignant neoplasms or benign central nervous system tumors among probands. Since the probands had been initially identified through hospitals and might represent a bias toward more severely affected cases, the relative risk was also determined for affected relatives; this risk was 1.5 (95% CI: 0.9-2.4). This risk was greater for females than for males (22). Certain types of cancers occur more frequently in individuals with NF1. Malignant peripheral nerve sheath tumors, often referred to as neurofibrosarcomas, are the most common malignancy occurring at increased frequency in NF1. These aggressive tumors are relatively resistant to therapy and often lethal (23). Central nervous system tumors, including optic pathway tumors, other astrocytomas, ependymomas, medulloblastomas as well as others, also occur more frequently in NF1 (24). In addition, individuals with NF1 have an increased risk for myeloid leukemias, with over a 200 fold relative risk for chronic myelomonocytic leukemia (25). The increased risk for malignancies in NF1 is compatible with the finding that the NF1 protein has GTPase activating protein activity, with resulting down-regulation of the ras oncogene (4). An increased risk for malignancy could be predicted to result from inactivation of this tumor suppressor function through NF1 mutation. The presentation of most NF1 features is age-dependent. Café-au-lait spots may be present at birth and increase in number in early childhood. Skin fold freckling is most often observed next. Neurofibromas frequently first appear or increase in number between the ages of 10 and 20. Lisch nodules of the iris are often not present in childhood but are seen in nearly all adults with NF1 (17). Prevalence of NF1 For several reasons, NF1 is a difficult condition for which to determine an accurate prevalence figure. First, the wide variability in expression means that mild cases may escape ascertainment in studies dependent on an affected individual coming to medical attention. Second, the age-dependent presentation of most NF1 features means that examination of children at a young age may miss cases that are truly affected with the condition. Third, the increased mortality seen in individuals with NF1 (see Morbidity and Mortality below) reduces the prevalence in later adulthood. Prevalence studies are summarized in Table 2 and suggest that NF1 is one of the most common autosomal dominant conditions. The prevalence does not appear to differ by gender. The wide variation in prevalence estimates may reflect differences in diagnostic criteria and methods of case ascertainment of the studies; however, the variation may also represent true differences between populations, perhaps due to a founder effect (particularly in smaller populations) or other factors. One study (30) has demonstrated differences in NF1 prevalence among different ethnic groups, with a higher prevalence in individuals of North African and Asian origin (1/522 and 1/1052, respectively), and a lower frequency among individuals of European and North American background (1/1562). These differences were statistically significant, and case ascertainment in this study was based on a mandatory physical examination for fitness for military service, suggesting that referral bias was not responsible for the observed differences. The question of the true prevalence of NF1 and whether it differs significantly between populations will require further study. Mortality of NF1 The best available mortality data are from a population-based study of NF1 patients living in Göteberg, Sweden (31). Adults (20 years and older) with NF1 were ascertained through multiple medical specialties. The average age at the time of ascertainment was 43.6 + 15.4 years for the 70 patients followed. Cases were followed for 12 years. Over this time period, 22 of these 70 NF1 patients died; 5.1 deaths were expected based on the general Swedish population. Of these 22 deaths, 13 were women and 9 were men, with 1.7 and 3.4 deaths expected in the populations, respectively, leading the authors to suggest that women may be more affected than men. The study showed a significantly reduced life expectancy in patients with NF1 (p < 0.001), with mean age at death of NF1 patients at 61.6 years, compared to a life expectancy in the general population of 75 years. Malignancy was the most common cause of death, occurring in 17 (24%) of the patients (31, 32); cardiovascular disease, hemorrhage, or embolism caused the death of 7 individuals. Some of these deaths were probably unrelated to the patients’ NF1. Hypertension was significantly associated with mortality -- 10 of 12 NF1 patients with high blood pressure died during the observation period. NF1 Risk Factors Paternal age has been shown to be significantly advanced in sporadic cases of several other autosomal dominant disorders, but whether paternal age is advanced in sporadic cases of NF1 is not clear. A study in Texas has recently addressed this question. Paternal age was obtained from birth certificates of cases (identified as NF1 patients seen in two specialty neurofibromatosis clinics), compared to that obtained from birth certificates of controls (two per case, chosen at random from the same year and county of birth). Fathers of NF1 patients were 1.5 years older than fathers of controls at the birth of the child, but this difference was not statistically significant (p=0.07) (33). It appears that the paternal age effect in sporadic cases of NF1 is either small or nonexistent. NF1 is the condition most commonly associated with NF1 gene mutations. The penetrance in NF1 is believed to be virtually 100% by adulthood (31); that is, individuals with an NF1 gene mutation have clinical manifestations of NF1, usually by the age of 6 years. Most studies have not found an obvious relationship between particular NF1 mutations and resulting clinical manifestations in a patient. However, attempts at genotype-phenotype correlation in NF1 are confounded by the effect of age, which increases the frequency of disease manifestations and the likelihood of serious complications in all patients. In addition, there is no consensus regarding how to define NF1 severity. Some studies of patients with large NF1 gene deletions indicate that they may have earlier onset of cutaneous neurofibromas and more often have dysmorphic facial features and mental retardation than most NF1 patients (13, 35, 36). However, not all NF1 patients with this phenotype have a large gene deletion (37), and some with large gene deletions have an unremarkable NF1 phenotype (38), raising questions about this genotype-phenotype relationship. The presence of a more severe phenotype may be a function of the amount of flanking DNA involved in the deletion, rather than NF1 gene deletion itself. Certain variants of NF1 have been associated either with specific NF1 mutations or with linkage to the NF1 gene, at least in some cases. These include Watson syndrome (characterized by pulmonic stenosis, café-au-lait spots, short stature, and cognitive impairment) (36, 37); familial multiple café-au-lait spots (without other NF1 features) (38-40); familial spinal neurofibromatosis (characterized by spinal tumors, and, sometimes café-au-lait spots, but not other features of NF1) (44, 45); and encephalocraniocutaneous lipomatosis (characterized by unilateral lipomatous growths, ipsilateral ophthalmologic and brain malformations, mental retardation, and seizures) (46). It appears that these variants may be allelic to NF1, at least in some families Patients with segmental neurofibromatosis have features of NF1 confined to a particular area of the body (e.g., one side of the body) (47). While it has been postulated that segmental neurofibromatosis results from a somatic mutation in the NF1 gene, this postulate has not yet been molecularly demonstrated. Somatic mosaicism for the NF1 gene has been reported in at least four cases (38, 48-50), but all these cases showed typical NF1, suggesting that the somatic mutation occurred early in embryonic development. Noonan syndrome is an autosomal dominant condition characterized by webbing of the neck, unusual facies, short stature, and congenital heart disease (often pulmonic stenosis). Features of Noonan syndrome, often without a cardiovascular malformation, have been observed in many patients with NF1. About 13% of patients with NF1 specifically examined for Noonan syndrome features had a Noonan syndrome phenotype (48); this frequency of co-occurrence seems unlikely if NF1 and Noonan syndrome are independent disorders. In some families, NF1 and Noonan syndrome have been shown to segregate as independent autosomal dominant traits, and Noonan syndrome is not linked to the NF1 locus in families without features of NF1. In other instances, features of both Noonan syndrome and NF1 appear to result from mutations of the NF1 gene, and both phenotypes segregate together in some families (51). It appears that the concurrence of NF1 and Noonan syndrome may have several different causes (52), but this question awaits further study. NF1 and the associated clinical presentations discussed above are the only conditions known to be caused by NF1 gene mutations. No studies of the NF1 gene in the general population have been performed. The wide variability of the NF1 phenotype, even in individuals with the same NF1 gene mutation, suggests that other factors are involved in determining clinical manifestations. These factors may include other genetic factors (modifying genes), environmental factors, and chance. Thus far, little is known about the relative contribution of these factors to the NF1 phenotype. A study of 175 individuals in 48 families, including six monozygotic twin pairs, evaluated variation of the NF1 phenotype with degree of relationship (53). The number of café-au-lait spots and the number of neurofibromas showed a high correlation between monozygotic twins, a lower correlation between first-degree relatives, and the lowest correlation among more distant relatives. The study also looked at the presence or absence of plexiform neurofibromas, optic gliomas, scoliosis, epilepsy, and referral for remedial education. With the exception of plexiform neurofibromas, these traits also showed familial clustering. The authors concluded that much of the phenotypic variation in NF1 is related to trait-specific "modifying genes". It has been suggested that environmental factors influence NF1 phenotype; however, no convincing evidence has been presented to support the involvement of any particular environmental factor. Riccardi has suggested that mechanical trauma (in the form of injury to the skin) may often precede the development of neurofibromas, but the evidence for involvement of this factor is anecdotal (54). The role of stochastic factors (chance) in the occurrence of some NF1 manifestations has also been hypothesized. Chance may be involved in determining which cells are affected by a somatic mutation and at what point in development somatic mutation occurs. Major questions remain about how the NF1 phenotype is determined, but it is likely that the NF1 genotype, modifying genes, environmental factors, and chance all play a role in the clinical manifestations of NF1 gene mutations. Laboratory testing for NF1 mutations is difficult. Although a variety of approaches has been used singly or in combination in research laboratories, none has been shown to be appropriate for routine clinical use. A protein truncation test is available commercially for NF1 mutation testing, but its sensitivity, specificity, and positive predictive value in a large group of patients have not been reported. In this test, RNA is reverse transcribed and the cDNA product is used to perform in vitro transcription and translation. Truncated neurofibromin proteins are identified by separating the protein products using an SDS-polyacrylamide gel (55). Mutations may then be confirmed by direct DNA sequencing. False positive results are possible when truncated proteins are not confirmed by sequencing (16). In addition, the protein truncation test cannot detect mutations that do not result in a truncated protein, such as missense mutations and large deletions, or mutations in which the RNA is unstable and thus unavailable for reverse transcription. The ability of the protein truncation test to detect mosaic mutations is unknown (10). However, it appears that the risk for both false positives (when a finding of a truncated protein is not confirmed by DNA sequencing) and false negatives may be significant with this test. Published studies of the sensitivity of the protein truncation test have been small; about 70% of cases of clinically diagnosed NF1 (14 of 20 cases in one study [55] and 11 of 15 cases in another [56]) had a positive result on the protein truncation test. 37 (77%) of 48 cases meeting NF1 diagnostic criteria referred for commercial testing are reported to have a positive protein truncation test result (Brown T, LabCorp, Research Triangle Park, NC, personal communication, 1999). No information is available on the specificity or positive predictive value of the protein truncation test. When the protein truncation test is negative, further studies may be helpful in identifying the mutation, but these studies are currently available only on a research basis. In familial NF1 cases (when two or more family members are affected), linkage analysis can be performed. The availability of intragenic microsatellite NF1 markers has increased the proportion of families in which linkage studies will be informative to an average of 90% and has also increased the diagnostic accuracy (57). Given that NF1 is easily diagnosed clinically in most affected individuals above the age of 6 years, the need for laboratory testing is limited to specific circumstances. One of these is for prenatal diagnosis when one of the parents is affected with NF1. If the causative mutation has been identified, direct testing for this specific mutation can be performed on chorionic villus or amniotic fluid samples; however, the severity of NF1 cannot be predicted prenatally, only the presence or absence of the mutation can be identified. Because of the wide variability in NF1 clinical expression, many families do not find prenatal diagnosis of NF1 acceptable (10). In families in which there are multiple affected relatives, linkage analysis can also be used for prenatal diagnosis. Once again, only presence or absence of the affected allele can be predicted, not the severity of the clinical manifestations. The other situation in which laboratory testing may be considered is in children at risk for NF1, before clinical diagnostic criteria are met. The child may be at risk because of a family history or because of having some features (typically café-au-lait spots), but not sufficient features to meet the established diagnostic criteria. While the ability to confirm or rule out the diagnosis with a laboratory test would be helpful, these children are at particular risk of possible stigmatization and unnecessary medical intervention if a false positive test results (16). Therefore, following the child on a regular basis for appearance of NF1 complications and sufficient clinical criteria to assure the diagnosis is likely to be a better option at this time. No general population-based studies using molecular testing to identify NF1 mutations have been performed. This type of study seems unnecessary since individuals over the age of six years with NF1 mutations can usually be identified by physical and ophthalmological examination. Clinical methods of NF1 ascertainment have been performed to estimate the prevalence of the condition in research studies in different populations (see Prevalence of NF1, above). However, population-based screening of individuals for clinical features of NF1 has not received substantial support. This is in part due to the difficulty of the effort: careful physical examination for NF1 features is time-consuming, unlike other population-based screening methods based on a simple laboratory test. In addition, since many NF1 features are age-dependent, diagnosis in a child under the age of 3 years of age is often challenging. However, most adult individuals with NF1 will have been identified as a result of a regular physical examination, even in absence of a screening program. An important question is whether an early NF1 diagnosis, as achieved through a screening program, would lead to prevention of NF1 complications. Since primary prevention of NF1 complications is not presently possible, this beneficial effect would be confined to the possibility that early recognition of complications may result in improved treatment. Several studies have assessed whether screening of individuals already known to have NF1 for complications is helpful. A recent paper suggests that the vast majority of abnormalities identified through a comprehensive screening program (consisting of ophthalmologic consultation with slit-lamp examination, chest radiograph, abdominal ultrasonography, neuroimaging and analysis of catecholamine levels) did not result in therapeutic action (54). Studies such as these have led many NF1 experts to suggest that a careful clinical evaluation for NF1 complications on an annual basis (or more often, if necessary) by a physician familiar with NF1 is optimal for affected individuals (16). Regular ophthalmological examination is also recommended in children with NF1 (18). Unfortunately, no studies are available to address the more general question whether an earlier NF1 diagnosis, made through a screening program, would lead to improved treatment. Another valid concern when considering a whether a population-based screening program may be beneficial is the effect that early diagnosis may have on family planning (avoidance of future pregnancies or utilization of prenatal diagnosis). In a recent survey, the majority of parents preferred a early diagnosis of NF1 in their child; however NF1 diagnosis did not usually result in avoidance of future pregnancies, and while prenatal diagnosis was viewed favorably, only a few parents said they would actually terminate an affected pregnancy (59). All of these issues will need to be taken into account in the discussion regarding population-based screening (whether using molecular methods or clinical methods); however, at this time, it appears that the benefits of early diagnosis do not outweigh the potential costs of a population-based screening program. Table 1. Summary of NF1 Mutation Types*.
*Reported to the NF1 Genetic Analysis Consortium, February, 1999 (http://www.nf.org/nf1gene/) Table 2. Studies of the prevalence of Neurofibromatosis type 1
*Estimated incidence at birth
General Resources Genetic Databases
Support Groups
Other Websites
|
