Article Text

Download PDFPDF

Clinical characterisation of 29 neurofibromatosis type-1 patients with molecularly ascertained 1.4 Mb type-1 NF1 deletions
  1. V-F Mautner1,
  2. L Kluwe1,
  3. R E Friedrich1,
  4. A C Roehl2,
  5. S Bammert2,
  6. J Högel2,
  7. H Spöri2,
  8. D N Cooper3,
  9. H Kehrer-Sawatzki2
  1. 1Department of Maxillofacial Surgery, University Medical Centre, Hamburg-Eppendorf, Germany
  2. 2Institute of Human Genetics, University of Ulm, Ulm, Germany
  3. 3Institute of Medical Genetics, School of Medicine, Cardiff University, Cardiff, UK
  1. Correspondence to Hildegard Kehrer-Sawatzki, Institute of Human Genetics, University of Ulm, Albert-Einstein-Allee 11, 89081 Ulm, Germany; hildegard.kehrer-sawatzki{at}uni-ulm.de

Abstract

Background Large deletions of the NF1 gene region occur in ∼5% of patients with neurofibromatosis type-1 (NF1) and are associated with particularly severe manifestations of the disease. However, until now, the genotype–phenotype relationship has not been comprehensively studied in patients harbouring large NF1 gene deletions of comparable extent (giving rise to haploinsufficiency of the same genes).

Method We have performed the most comprehensive clinical/neuropsychological characterisation so far undertaken in NF1 deletion patients, involving 29 patients with precisely determined type-1 NF1 (1.4 Mb) deletions.

Results Novel clinical features found to be associated with type-1 NF1 deletions included pes cavus (17% of patients), bone cysts (50%), attention deficit (73%), muscular hypotonia (45%) and speech difficulties (48%). Type-1 NF1 deletions were found to be disproportionately associated with facial dysmorphic features (90% of patients), tall stature (46%), large hands and feet (46%), scoliosis (43%), joint hyperflexibility (72%), delayed cognitive development and/or learning disabilities (93%) and mental retardation (IQ<70; 38%), as compared with the general NF1 patient population. Significantly increased frequencies (relative to the general NF1 population) of plexiform neurofibromas (76%), subcutaneous neurofibromas (76%), spinal neurofibromas (64%) and MPNSTs (21%) were also noted in the type-1 deletion patients. Further, 50% of the adult patients exhibited a very high burden of cutaneous neurofibromas (N≥1000).

Conclusion These findings emphasise the importance of deletion analysis in NF1 since frequent monitoring of tumour presence and growth could potentiate early surgical intervention thereby improving patient survival.

  • Neurofibromatosis type 1
  • NF1 microdeletions
  • genotype–phenotype correlation
  • hereditary cancer syndrome
View Full Text

Statistics from Altmetric.com

Approximately 5% of all patients with neurofibromatosis type 1 (NF1) possess large deletions in 17q11.2 that include both the NF1 gene and its flanking regions.1 Three types of NF1 deletion (type-1, type-2 and atypical) are distinguishable on the basis of their size and the locations of their respective breakpoints. The most common (type-1), accounting for 60–70% of all large NF1 deletions, are generated by non-allelic homologous recombination (NAHR) between segmental duplications (NF1-REP A and NF1-REP C). These recurrent type-1 deletions encompass 1.4 Mb and lead to the loss of 14 functional genes (figure 1).2–6 The less common type-2 NF1 deletions, accounting for ∼10–20% of all large NF1 deletions, span 1.2 Mb, but although similarly recurrent and generated by NAHR, they result in the loss of only 13 functional genes since LRRC37B is not deleted. The breakpoints of type-2 deletions are located within the SUZ12 gene and its pseudogene and are frequently associated with somatic mosaicism.7–9 By contrast, atypical NF1 deletions are of variable size and are characterised by non-recurrent breakpoints. Consequently, atypical NF1 deletions can differ in terms of the number of genes included within the deleted region.10–12 A particular characteristic of type-1 deletions is that the NAHR breakpoints cluster within two preferred regions, termed paralogous recombination sites 1 and 2 (PRS1 and PRS2) both of which are located within NF1-REPs A and C.5 6 From a clinical vantage point, large NF1 deletions are of considerable interest since they are frequently associated with more severe clinical manifestations than those observed in patients with intragenic NF1 gene mutations [13 14 and references therein]. More specifically, by comparison with other NF1 patients, those individuals harbouring NF1 deletions have been reported to have an increased risk of malignant peripheral nerve sheath tumours,15 lower average intelligence,16 connective tissue dysplasia, skeletal malformations and dysmorphic facial features13 14 and accelerated growth (height) and carpal bone age.17

Figure 1

Schematic representation of the NF1 gene region in 17q11.2. The 1.4 Mb-spanning type-1 deletions encompass a total of 14 genes, which are indicated by horizontal bars. Red arrows specify the relative positions of the multiplex ligation-dependent probe amplification (MLPA) within the type-1 deletion interval. Green arrows represent MLPA probes that were not located within the type-1 deletion region. From the results generated with the MLPA-kit P122 C1, it was not possible to ascertain whether or not the genes marked in yellow were included within the deleted region. Thus, discrimination between type-1 deletions with breakpoints located in the NF1-REPs, and atypical deletions which encompass the genes marked in yellow is not possible using the currently available MLPA-kit P122 version C1.

Although, as a group, there is a general tendency for patients possessing large NF1 deletions to display a more severe form of NF1 than patients with intragenic NF1 mutations, considerable clinical variability has been observed between individuals with large deletions.18 Hence, the general validity of reported genotype–phenotype correlations in association with large NF1 deletions remains unclear. One of the reasons for a lack of clarity could be that many studies have failed to demarcate the NF1 deletion breakpoints precisely enough for the number of genes residing within the deletion intervals to be accurately determined. Indeed, in the majority of NF1 deletion analyses performed to date, the deletions were characterised by fluorescent in situ hybridisation (FISH), a method which is insufficiently precise to distinguish between type-1 deletions and those atypical deletions with breakpoints located close to the NF1-REPs. Over the last few years, multiplex ligation-dependent probe amplification (MLPA) has been deployed in order to identify NF1 deletions.19 20 While the commercially available MLPA-kit (P122 version C1) can successfully distinguish type-1 from type-2 NF1 deletions, type-1 deletions cannot be reliably distinguished from atypical deletions which can also encompass the genes GOSR1, TBC1D29, RHOT1, RHBDL3 and C17orf75 in addition to the 14 genes included in the type-1 deletion interval (figure 1). This failing is due to the poor positioning of the probes in this MLPA-kit. FISH and/or MLPA are, therefore, unable to determine the NF1 deletion breakpoints with the degree of precision required to determine accurately the number of deleted genes. In order to improve deletion characterisation, customised oligonucleotide array comparative genomic hybridisation has recently been developed; this technique is able to differentiate unambiguously between all three types of NF1 deletion.12 The latter is a prerequisite for the meaningful analysis of the genotype–phenotype relationship in the context of NF1 deletions. To date, only Descheemaeker et al16 have employed precisely characterised NF1 deletions to explore genotype–phenotype relationships in NF1; these authors analysed 10 NF1 patients harbouring type-1 deletions with breakpoints located in the PRS1 and PRS2 hotspot regions. However, these authors confined their clinical investigation to the cognitive abilities of the patients and neglected to study other clinical manifestations of the disease. Thus, until now, no comprehensive study of the genotype–phenotype relationship in patients harbouring large precisely characterised NF1 gene deletions has been performed. Here, we analyse clinical data from 29 patients with precisely demarcated type-1 NF1 deletions to derive a comprehensive assessment of the clinical phenotype associated with this type of gross deletion.

Patients and methods

Patients

The 29 patients with type-1 NF1 deletions were identified by screening ∼800 NF1 patients, clinically examined in the Hamburg NF Centre (Department of Maxillofacial Surgery, University Medical Centre, Hamburg-Eppendorf), using micro-satellite markers as previously described.1 The deletion patients were not selected by clinical phenotype but rather were simply identified from among 800 sequential NF1 patients who came to the clinic to receive medical care, advice or genetic counselling. All patients (or their parents) gave written informed consent for the molecular studies to be performed and for the publication of photographs and clinical data. Two patients inherited the deletion from their mothers (patients 284 and 763), while the remaining 27 patients had de novo deletions, adjudged on the basis of the absence of a clinically affected parent. The mother of patient 284 (patient 1454-1), herself clinically affected as a result of a de novo type-1 NF1 deletion, was also included in the analysis. Among the 29 patients included in this study (12 men and 17 women) were nine children between the ages of 4 and 15.

Characterisation of the deletions

DNA was isolated from patient peripheral blood samples using the Qiamp kit (Qiagen, Valencia, California, USA). All 29 type-1 deletion patients were investigated by MLPA using the SALSA P122 C1 MLPA assay (MRC-Holland, Amsterdam, The Netherlands) according to the manufacturer's instructions and as described by Wimmer et al.19

In order to detect type-1 NF1 deletions with breakpoints located in the PRS1 and PRS2 hotspots, breakpoint-spanning PCRs were performed using the Expand Long Template PCR system (Roche, Penzberg, Germany) with the primers listed in additional table 1, as previously described.4 5 In patients where the breakpoints were not assigned to the PRS1 and PRS2 breakpoint regions, the extent of the deletion was determined by SNP or microsatellite marker analysis using PCR (primers listed in additional table 2).

FISH analysis

The type-1 deletions identified by the methods described above were also investigated by FISH as previously described7 8 to exclude mosaicism. A total of 30 metaphase spreads and at least 100 interphase nuclei were evaluated in order to exclude the presence of normal cells lacking the deletion.

Clinical investigation

The 29 patients were investigated according to a standardised protocol that included a comprehensive anamnestic evaluation as well as the assessment of various clinical parameters and NF1-associated manifestations at the University Medical Centre, Hamburg-Eppendorf, as outlined in additional text 1.

Statistical analysis

The χ2 goodness-of-fit test was used to assess whether the frequency of clinical symptoms differed between NF1 patients with large deletions and patients from the general NF1 population. Exact 95% confidence intervals were determined wherever possible for patients manifesting certain clinical symptoms. Confidence intervals for IQ scores were assigned under the assumption of a normal distribution.

Results

Characterisation of the deletions

The 28 large NF1 deletions investigated in this study were identified in 29 patients including patients 1454-1 and 284, who are mother and daughter. The deletions were ascertained by FISH and microsatellite marker analysis. Mosaicism with normal cells bearing two copies of the NF1 gene was not observed by FISH analysis of patient lymphocytes.

Breakpoint-spanning PCR analysis indicated that 24 of the 28 deletions had breakpoints within either the PRS1 or PRS2 recombination hotspot regions of the NF1-REPs (additional table 3). Thus, the deletion breakpoints were narrowed down to regions of length 3.5 kb (PRS2) and 2.9 kb (PRS1), respectively. The high-resolution breakpoint mapping in these 24 deletions indicated that they encompassed a total of 14 genes. However, four of the 29 patients did not generate a product in the PRS1 and PRS2 deletion breakpoint-spanning PCRs, even though the MLPA results were suggestive of a type-1 deletion. In these patients, the breakpoints could have been located either within the NF1-REPs (but not within the PRS1 and PRS2 regions) or proximal or distal to the NF1-REPs. In the latter case, the deletions might have been >1.4 Mb and, hence, could have included up to five additional genes (figure 1). MLPA was unable to provide any better resolution owing to the distance between the NF1-REPs and the next available MLPA probes (additional figures 1, 2). To determine whether or not these abovementioned genes flanking the NF1-REPs were present within the deletion intervals of the four patients who were negative for the PRS1/PRS2 breakpoint-spanning PCRs, flanking SNPs and a microsatellite marker were investigated. Heterozygosity patterns for these polymorphisms indicated that the breakpoints of the respective deletions were located within the NF1-REPs in all four cases (additional table 4). Hence, we may conclude that all 29 NF1 deletions in these patients encompass the same 14 functional genes.

Clinical phenotype associated with the type-1 NF1 deletions

In what follows, the clinical and neuropsychological features of 29 patients with type-1 NF1 deletions are described. Further, the frequencies of the observed features are compared with those reported previously in patients with large NF1 deletions. In this regard, we make particular reference to the studies of Venturin et al13 and Mensink et al,14 who investigated their own NF1 deletion patients and undertook a retrospective review of the literature on previously published patients with large NF1 deletions. Finally, the frequency of each clinical feature associated with type-1 NF1 deletions is compared with the available corresponding frequencies reported in the general NF1 population (additional table 5).

Dysmorphic facial features

Facial dysmorphism was noted in 26 of the 29 patients studied (90%; additional table 6; figures 2 and 3) and, hence, was present at a higher frequency than reported previously.13 14 Hypertelorism was the most common feature observed in 86% of our patients. Coarse facial appearance and facial asymmetry were also frequent. However, not all NF1 patients investigated here exhibited dysmorphic facial features that might immediately have raised suspicion of a large NF1 deletion. Patient 1338, for example, did not exhibit dysmorphic facial features (additional figure 3). Thus, although our study indicated that dysmorphic facial features are present in the majority of patients with type-1 NF1 deletions, facial dysmorphism is by no means a universal feature of type-1 deletions.

Figure 2

Facial appearance of female patients with type-1 NF1 deletions: (A) patient 752, (B) patient 1454,1, (C) patient 450, (D) patient 1333, (E) patient 1143, (F) patient 801. Hypertelorism was the most common feature.

Figure 3

Facial appearance of male patients with type-1 NF1 deletions: (A) patient 3028, (B) patient 521, (C) patient 270, (D) patient 1178, (E) patient 1547 and (F) patient 2284.

Overgrowth and unusual body habitus

Tall-for-age stature with height measurements at or above the 94th percentile, and large hands and feet were noted in 46% of the patients studied (additional table 7). By contrast, growth retardation and short stature are relatively common features in the general NF1 population.21–23 Accelerated height growth and carpal bone age have been previously reported to occur frequently in patients with large NF1 deletions, especially in pre-school children aged 2–6 years in whom overgrowth is most evident.17 In our study, two of the three children in the pre-school age were unusually tall (99th and 100th percentiles in height, respectively).

Macrocephaly, measured as occipital-frontal circumference at or above the 97th percentile, was ascertained in 39% of patients (additional table 7). Since macrocephaly is observed in 29–45% of all individuals with NF1,24–26 the frequency of this feature does not appear to be elevated in patients with type-1 deletions.

In five of the 29 patients studied, we observed pes cavus, a deformity of the feet that has not so far been described in patients with NF1 deletions (additional table 7 and figure 4).

CALM, freckling and Lisch nodules

Café-au-lait-macules (CALM) and Lisch nodules were observed in 27 of the 29 patients in our study (93%; additional table 8) whereas inter-triginous freckling was noted in 28 patients. The frequencies of these clinical features in deletion patients are concordant with their respective rates previously observed in the general NF1 population.27 28

Cognitive ability

A significant delay in cognitive development was ascertained in 14 of the 29 type-1 deletion patients (48%), whereas learning difficulties were recorded in 13 patients (45%; additional table 9). Learning difficulties have been previously noted in 30–60% of all NF1 children.29–34 A significant delay in cognitive development and/or learning difficulties were therefore observed in 93% of our patients, a rate markedly higher than that reported by Mensink et al (70%; additional table 5).

In 21 of our patients under study, a mean full-scale IQ (FSIQ) of 76.9 was ascertained (SD 14.6; 95% CI, 70.2 to 83.5). Seven of these 21 patients were children. Mental retardation (IQ<70) was evident in eight of the 21 patients (38%; four children and four adults) while borderline mental retardation (70<IQ<85) was noted in a further five of the 21 patients (24%). By contrast, an IQ<70 has been previously observed in only 6–8% of the general NF1 patient population.29

The mean FSIQ (76.9) observed in the 21 patients investigated in this study is similar to the mean FSIQ of 76.0 determined in 11 patients with type-1 NF1 deletions by Descheemaeker et al.16 These authors also ascertained the mean FSIQ in 106 NF1 individuals without an NF1 deletion to be 88.5. Thus, in general, the mean FSIQ in patients with type-1 NF1 deletions would appear to be significantly lower than the mean FSIQ in NF1 patients without a deletion. Although the average intelligence of type-1 NF1 deletion patients is generally lower than that of patients without deletions, it should be appreciated that there is a substantial overlap between the two groups.

Attention deficit hyperactivity disorder (ADHD)

Attention deficit hyperactivity disorder (ADHD) was diagnosed in eight of 24 deletion patients investigated (33%, three children and five adults; additional table 9). In a further eight of these 24 patients, the Test of Variables of Attention (T.O.V.A.) revealed abnormal results even though the strict definitional criteria for ADHD were not met. Taken together, 73% of the patients investigated in this study were considered to have attention difficulties. Among the eight children investigated, seven showed abnormal T.O.V.A. results, and three of these met the strict criteria for ADHD (37%). However, we noted ADHD in 49% of the children in the Hamburg NF1 patient cohort (N=∼800) unselected with respect to their mutation.35 Hence, ADHD does not appear to occur at an elevated frequency in children with large NF1 deletions as compared with children in the general NF1 population.

Skeletal anomalies

Anomalies of the skeletal system were detected in 22 of the 29 patients studied here (76%). The most frequent skeletal anomaly was scoliosis which was noted in 43% of the patients (scoliosis with a curve of ≥10°; additional table 10). Hence, scoliosis was encountered significantly more frequently in patients with type-1 NF1 deletions (χ2 test, p=0.005, 1 d.f.) than in the general NF1 population, in whom 10–20% have been reported to present with scoliosis.36 37 Severe scoliosis with a curve of ≥20° was observed in four of the 12 deletion patients with scoliosis.

Pectus excavatum was observed in nine of the 29 patients (31%), a frequency in accord with previous findings (33%) in patients with NF1 deletions.14 In the general NF1 population, pectus excavatum has been reported to occur at a frequency of 12% or even up to 50%.38 39

Bone cysts were noted to occur with a remarkably high frequency in our group of type-1 deletion patients (additional table 10). Indeed, eight of the 16 patients investigated by MRI had bone cysts. In five of these patients, the bone cysts were noted in association with fractures. Bone cysts have not previously been reported in patients with large NF1 deletions, probably because radiography and MRI investigations were not performed.

Connective tissue anomalies and heart defects

Hyperflexibility of joints was observed in 21 of the 29 patients (72%) with type-1 NF1 deletions (additional table 6). Mensink et al14 also reported a high frequency of this feature (58%) in patients with large deletions. These authors noticed that in addition to joint laxity, other connective tissue abnormalities were also frequently associated with large NF1 deletions, including soft skin on the palms and cardiovascular anomalies. In our own study, soft fleshy palms with an excess of connective tissue were observed in 50% of patients while congenital heart defects were observed in 21%. Previously, Venturin et al13 noted cardiovascular malformations in 18% of NF1 deletion patients. Various kinds of congenital heart disease have been repeatedly observed in NF1 patients, but data concerning their frequency in the general NF1 population are not congruent between different studies (the prevalence ranging from 2% to 27%).40–43

Muscular hypotonia and reduced speech intelligibility

Mild muscular hypotonia was documented in 13 of the 29 type-1 NF1 deletion patients (45%; additional table 11). This feature has not previously been noted in association with large NF1 deletions. Reduced speech intelligibility was observed in 14 patients (48%); this feature has not previously been reported in association with large NF1 deletions. Although the overall prevalence of speech disorders in NF1 is not yet known, speech difficulties generally appear to occur more frequently in patients with NF1 as compared to the normal population.44 45 Speech difficulties due to deficits in articulation occur in ∼25% of children with NF1.46

Subcutaneous and cutaneous neurofibromas

Subcutaneous neurofibromas were observed in 22 of the 29 patients (76%; additional table 12). If we consider only the 20 adult patients in our cohort, then all 20 patients had subcutaneous neurofibromas. Six of the 29 patients investigated here had ≥1000 subcutaneous neurofibromas (21%). The prevalence of subcutaneous neurofibromas appears to be significantly elevated in patients with type-1 NF1 deletions (χ2 test, p=0.008, 1 d.f.) because subcutaneous neurofibromas were ascertained in only 48% of 468 NF1 patients unselected by NF1 mutation type and hence probably reflecting the general NF1 population.47

Cutaneous neurofibromas were observed in 86% of our 29 type-1 deletion patients. Tucker et al47 observed cutaneous neurofibromas in 85% of 443 NF1 patients examined, a rate that may reflect the presence of these tumours in the general NF1 population. In our own study, a high burden of cutaneous neurofibromas (N≥1000) was noted in 10 patients (34%). If we consider only the 20 adult patients in our group, then 10 of these patients (50%) had ≥1000 neurofibromas. Eleven patients either lacked, or exhibited <10, cutaneous neurofibromas. However, eight of these 11 patients were children, while three were young adults (additional table 12).

In previous studies, it has been suggested that patients with large NF1 deletions frequently exhibit early onset growth of cutaneous neurofibromas.2 In our study, four of the nine children included lacked cutaneous neurofibromas, whereas the other five children had fewer than nine cutaneous neurofibromas. By contrast, subcutaneous neurofibromas (N≥20) were noted in four of the nine children in our study (44%). Thus, the early onset of subcutaneous neurofibromas would appear to be more pronounced in our own group of children with type-1 deletions than cutaneous neurofibromas.

Plexiform neurofibromas

Plexiform neurofibromas (PNF) were observed in 22 of the 29 patients studied here (76%; additional table 12). Thus, PNF appear to be more prevalent in our patients with precisely demarcated type-1 deletions than has been previously observed in studies of patients with less well-characterised NF1 deletions. The prevalence of PNF in the general NF1 population has been estimated by CT and MRI imaging to be 43–50%.48–50 Hence, the prevalence of PNF would appear to be significantly increased in patients with type-1 deletions as compared to all NF1 patients (χ2 test, p=0.008, 1 d.f.).

Multiple plexiform neurofibromas reportedly occur in 9–21% of patients with NF1.24 49 In our study, nine of the 29 type-1 deletion patients exhibited multiple superficial plexiform neurofibromas (31%).

Spinal neurofibromas

Paraspinal or spinal neurofibromas were observed in nine of 14 patients investigated by MRI of the spine (64%; additional table 12). In previous studies, spinal neurofibromas were observed much less frequently in patients with large deletions (6%; additional table 5), probably because spinal MRI was not performed. Based on the analysis of Tucker et al,47 it may be inferred that spinal neurofibromas occur in ∼30% of all NF1 patients. Thus, spinal neurofibromas appear to occur more frequently among patients with type-1 NF1 deletions than in all individuals with NF1 (χ2 test, p=0.008, 1 d.f.).

Optic pathway gliomas and structural brain abnormalities

Cerebral MRI revealed optic pathway gliomas in five of 27 patients (19%; additional table 12). The overall incidence of optic pathway gliomas in NF1 has been reported to be 15–19%.51 52 Thus, the prevalence of optic pathway gliomas is not significantly increased in patients with type-1 deletions as compared to the general NF1 population.

Structural brain abnormalities were observed, upon MRI investigation, in five of our 29 patients (17%; additional table 12). Korf et al53 previously observed structural brain abnormalities in three of five patients with large NF1 deletions. T2 hyper-intensities (T2H) are focal areas of high signal intensity on T2-weighted images that are thought to represent foci of neural dysplasia or dysmyelination.54 Here, T2H were observed in 13 of the 29 deletion patients investigated (45%). By contrast, all five patients with large NF1 deletions investigated by Korf et al53 had multiple regions of bright T2 signal intensity. However, discrete T2H are seen in an average of 64% of all individuals with NF1.37 54–57 Although T2H tend to be more common in children and to decrease with age,58 57% of NF1 patients between 16 and 30 years have been reported to exhibit T2 hyperintensities.56 Thus, we conclude that the prevalence of T2H in patients with type-1 NF1 deletions is not significantly elevated as compared to the general NF1 population.

Epilepsy

Epilepsy was noted in two of our patients (7%) and, hence, does not occur at an increased frequency in patients with type-1 deletions as compared with the general NF1 population, in whom epilepsy has been observed with a frequency of 3–7%.24 25 59

Malignant peripheral nerve sheath tumours

Six of the 29 patients investigated (21%) here had a malignant peripheral nerve sheath tumour (MPNST; additional table 12). In their review of the literature, Mensink et al14 reported MPNSTs in 13% of patients with NF1 deletions. Long-term clinical follow-up is essential in order to assess whether or not a patient has developed an MPNST. The long-term follow-up that we provided to our own patients may well be the factor responsible for the higher rate of MPNSTs detected in this study as compared with the review of the literature published by Mensink et al14 The prevalence of MPNSTs in all individuals with NF1 has been estimated to be 2–5%.60

Discussion

The first patient with a large NF1 gene deletion was described by Kayes et al.61 Since this report, a total of 166 patients with large deletions of the NF1 gene region have been identified and described in terms of their clinical phenotypes.13 14 Much information has been obtained from these studies regarding the delineation of the NF1 deletion-associated clinical phenotype. However, the establishment of genotype–phenotype correlations has been hampered by the incomplete clinical and molecular investigation of patients. Indeed, in most previous reports, the extent of the deletions was not precisely determined.

In this study, we have analysed 29 patients harbouring precisely defined type-1 NF1 deletions with breakpoints located within NF1-REPs A and C. Hence, all 29 NF1 deletions in these patients encompass the same 14 functional genes (figure 1). The clinical investigation of these 29 patients indicated that type-1 NF1 deletions are frequently associated with facial dysmorphic features, tall-for-age stature, large hands and feet, an excess of connective tissue in the hands and feet, hyperflexibility of the joints, cognitive impairment and scoliosis. These associations have already been noted in previous studies of large NF1 gene deletions even though these earlier studies often included less well-characterised patients.13 14 Our findings reported here serve to confirm that these clinical features are indeed frequent in patients with type-1 NF1 deletions. However, our study also indicates that even within this relatively homogeneous group of 29 patients (with comparable deletions leading to the loss of the same 14 functional genes), some variability is apparent in terms of the presence or absence of some of the abovementioned features.

The majority of clinical features observed in our type-1 deletion patients occurred at a higher frequency than has been previously reported in published reviews of the literature (additional table 5).13 14 This is probably due to the fact that these reviews included many different studies in which the clinical phenotypes of the NF1 patients had neither been comprehensively nor uniformly described and in which the sample sizes of analysed patients were often quite low. In this study, however, the patients were uniformly ascertained with only those patients harbouring clearly defined type-1 NF1 deletions being included.

Additional clinical features that we observed at high frequency in patients with type-1 NF1 deletions included pes cavus (noted in 17% of patients), bone cysts (50% of patients), attention deficit (73% of patients), muscular hypotonia (45%) and speech difficulties (48%). These features have not been reported to occur at increased frequency in previous studies of large NF1 deletions. This may be related to the comprehensive nature of our clinical investigation protocol and to the care that we took in selecting patients with comparable deletions. Remarkably, 73% of the deletion patients in our study exhibited an attention deficit as indicated by abnormal results in the T.O.V.A. It may be inferred from our observations that stimulant-based medication of attention deficit could help to improve both the learning capability and the social interactions of patients with type-1 NF1 deletions.

Clinical features observed at increased frequency in patients with large NF1 deletions

We noted that the prevalence of subcutaneous, plexiform and spinal neurofibromas as well as MPNSTs is significantly higher in patients with type-1 NF1 deletions than in the general NF1 population. Importantly, 76% of the patients with type-1 deletions investigated here had subcutaneous neurofibromas. The proportion of patients with type-1 NF1 deletions manifesting subcutaneous neurofibromas appears to be significantly higher than in the general NF1 population (p=0.008). This is important to consider because subcutaneous neurofibromas are associated with mortality in NF1.62 Patients with subcutaneous neurofibromas are roughly three times more likely to develop MPNSTs than individuals without subcutaneous neurofibromas.47 63 Consistent with this finding is that five of the six patients with MPNSTs in our study also had subcutaneous neurofibromas. The frequency of patients with MPNSTs in our study was 21%, thereby confirming the observation originally made by De Raedt et al15 that patients with NF1 deletions have a substantially higher risk for the development of MPNSTs as compared with NF1 individuals lacking large NF1 deletions. Thus, whereas the lifetime risk of an MPNST in all NF1 individuals is 8–13%,64 NF1 deletion patients have an estimated lifetime risk for an MPNST of 16–26%.15

The prevalence of plexiform neurofibromas in the 29 patients investigated here was also remarkably high (76%). Such a high prevalence has not been previously observed in studies of patients with large NF1 deletions. According to our findings, the proportion of patients with type-1 NF1 deletions manifesting plexiform neurofibromas is significantly higher (p=0.008) than in the general NF1 population. The high prevalence of plexiform neurofibromas reported here is eventually expected to translate into an increased risk of developing MPNSTs. Many plexiform neurofibromas are likely to be congenital lesions with the potential to undergo transformation into MPNSTs.49 Indeed, many NF1-associated MPNSTs arise in pre-existing plexiform neurofibromas.47 65 66

Spinal neurofibromas were observed in nine of 14 patients for whom spinal MRI data were available (64%). The high proportion of our patients manifesting spinal neurofibromas suggests that patients with NF1 deletions may have a high internal tumour load.

In summary, our study clearly indicates that the clinical phenotype associated with type-1 NF1 deletions is often severe and associated with additional complications which render special and more intensive clinical and psychological care necessary.

Acknowledgments

This work was funded by the Deutsche Forschungsgemeinschaft (DFG), grant FR1035/6-1, and the Deutsche Krebshilfe, grants 106982 and 108793. We would like to express our gratitude to the patients and their relatives who participated in this study.

References

View Abstract

Supplementary materials

Footnotes

  • Funding Other Funders: Deutsche Krebshilfe, grants 106982 and 108793; Deutsche Forschungsgemeinschaft grant FR1035/6-1.

  • Competing interests None.

  • Patient consent Patient or parental consent obtained.

  • Ethics approval This study was conducted with the approval of the University clinic of Ulm and the University Clinic of Hamburg.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.