lable at ScienceDirect

ARTICLE IN PRESS

European Journal of Medical Genetics xxx (2009) 1–6

Contents lists avai

European Journal of Medical Genetics

journal homepage: ht tp: / /www.elsevier .com/locate/e jmg

Original article

Two sporadic spinal neurofibromatosis patients with malignant peripheralnerve sheath tumour

C. Fauth a, H. Kehrer-Sawatzki b, A. Zatkova c, S. Machherndl-Spandl d, L. Messiaen e, G. Amann f,J.A. Hainfellner g, K. Wimmer a,c,*

a Clinical Genetics Section, Department of Medical Genetics, Molecular and Clinical Pharmacology, Medical University Innsbruck, Austriab Institute of Human Genetics, University of Ulm, Germanyc Department of Medical Genetics, Medical University Vienna, Austriad Krankenhaus der Elisabethinen Linz, Austriae Department of Genetics, Medical Genomics Laboratory, University of Alabama at Birmingham, Al, USAf Clinical Institute of Pathology, Medical University of Vienna, Austriag Institute of Neurology, Medical University Vienna, Vienna, Austria

a r t i c l e i n f o

Article history:Received 10 May 2009Accepted 1 August 2009Available online xxx

Keywords:Neurofibromatosis type 1Spinal neurofibromaMalignant peripheralnerve sheath tumourSplicing mutationLOHTP53

* Corresponding author at: Clinical Genetics SectGenetics, Molecular and Clinical Pharmacology, MSchopfstr 41, 6020 Innsbruck, Austria. Tel.: þ43 512 9073510.

E-mail address: katharina.wimmer@i-med.ac.at (K

1769-7212/$ – see front matter � 2009 Elsevier Massdoi:10.1016/j.ejmg.2009.08.001

Please cite this article in press as: C. Fauthtumour, European Journal of Medical Genet

a b s t r a c t

We analyzed two unrelated male patients in whom neurofibromatosis type 1 (NF1) was not suspecteduntil they presented with malignant peripheral nerve sheath tumours (MPNSTs) in their thirties andforties, respectively. Patient A presented with progressive peroneus paresis due to a rapidly growingMPNST in the thigh. MRI examination revealed multiple symmetrical spinal neurofibromas in thispatient as well as in patient B who presented at the age of 42 with paraparesis and an MPNST at spinallevel L4. Dermal features in both patients were strikingly mild, therefore both patients were consideredbelonging to the NF1-subform of spinal neurofibromatosis (SNF). The novel NF1 mutations identified, i.e.splice mutation, c.7675þ1G > A, in patient A and two alterations, p.Cys1016Arg and p.2711delVal, locatedin trans in patient B support the notion that the phenotype of SNF may be related to mutations withpossible residual functionality. The MPNSTs of both patients showed LOH affecting chromosome 17including the NF1 locus. Furthermore, a truncating TP53 mutation was identified in the tumour of patientA. Both alterations are frequent findings in NF1-associated MPNSTs. To our knowledge these are the firstMPNST patients with the clinical phenotype of SNF. The clinical course observed in these two patientssuggests that nodular plexiform neurofibromas and spinal-nerve-root neurofibromas which may beasymptomatic for a long time and, hence, unrecognized in SNF patients bear the risk for malignanttransformation.

� 2009 Elsevier Masson SAS. All rights reserved.

1. Introduction

Neurofibromatosis type 1 (NF1) is characterized by a broadphenotypic variability even among family members carrying thesame NF1 mutation [1]. Nevertheless, almost all adult patients showdermal features, i.e. cafe-au-lait spots (CLS), freckling and cutaneousand/or subcutaneous neurofibromas, allowing for unequivocalclinical diagnosis according to the NIH consensus criteria [2–4].Exceptions are patients with spinal neurofibromatosis (SNF). The

ion, Department of Medicaledical University Innsbruck,03 70513; fax: þ43 512 9003

. Wimmer).

on SAS. All rights reserved.

et al., Two sporadic spinal nics (2009), doi:10.1016/j.ejmg

phenotype of SNF is characterized by multiple symmetric spinalneurofibromas and a strikingly mild dermal involvement, which isfrequently limited to CLS. Dermal neurofibromas are not observedeven in adults with SNF, which makes the clinical diagnosis ofNF1 difficult [5]. The phenotype of SNF has been reported to beconsistent in a few multi-generation familial cases [5–9] and tosegregate with NF1 mutations [6–8,10]. However, the geneticmechanisms leading to this peculiar phenotypic expression arelargely unknown.

We report here two presumably sporadic patients who pre-sented with the phenotype of SNF. The patients were diagnosedfirst with NF1 in their thirties and forties, respectively, when theydeveloped malignant peripheral nerve sheath tumours (MPNSTs).To our knowledge these are the first patients with SNF reported todevelop malignancies.

eurofibromatosis patients with malignant peripheral nerve sheath.2009.08.001

Table 1

Feature Patient A Patient B

Age at examination 31 years 42 yearsCafe-au-lait macules 7 � (>1.5 cm Ø) NoneFreckling None NoneLisch nodules None UnknownCutaneous

neurofibromasNone 2–6 (subcutaneous)

Plexiformneurofibromas

N. peroneus None

Spinal neurofibromas C5/6–Th3/4, Th7/8–L5

extending sacralC1, C4-C6, L2/3

Malignancies MPNST (right distal thigh,n. peroneus)

MPNST L3(m. psoas region)

Other NF1-associatedtumours

Asymptomaticpheochromocytomaand optic glioma

None

NF-associated bonelesions

Scoliosis None

Intellectual abilities Normal NormalOther features Femur cyst or intraosseous

tumourNone

NIH criteria forclinical NF1diagnosis fulfilled

Yes, (>6 CLS, >2 neurofibromasof any type, optic glioma)

No (only multipleneurofibromas, but noother criteria fulfilled)

C. Fauth et al. / European Journal of Medical Genetics xxx (2009) 1–62

ARTICLE IN PRESS

2. Patients and methods

2.1. Patient A

Patient A initially presented at the age of 30 years with a sensorydeficit and pain of the right lower leg. Subsequently he developeda progressive peroneal paresis. A few months later a rapidlygrowing malignant peripheral nerve sheath tumour (MPNST) of theright distal thigh was diagnosed.

Upon clinical examination the patient had 7 cafe-au-lait spots(>15 mm diameter) and thoracic scoliosis. He did not have anycutaneous neurofibromas, axillary freckling or Lisch nodules.

Magnetic resonance imaging (MRI) of the spine detectedmultiple bilateral paraspinal neurofibromas in the cervical, thoracicand lumbosacral region (C5/6–Th3/4, Th7/8–L5, extending into thesacral region). Neurofibromas of the sacral region were localizedintra- and presacrally and were growing into the sacral plexus(Fig. 1A). MRI of the legs indicated nodular plexiform neurofi-bromas of the right ischiadic, peroneal and tibial nerve and of thefemoral nerves on both sides (Fig. 1B).

The MPNST developed into a palpable tumour with the size ofa tennis ball (14 � 5.6 � 9.8 cm) within a few weeks. It mostprobably arose in the nodular plexiform neurofibroma of theproximal tibial and peroneal nerve of the right leg. The tumour wassurgically removed and the patient underwent several cycles ofchemotherapy with epirubicine and ifosfamide. Histopathologyshowed a myxoid variant of an MPNST.

In addition, the patient had an asymptomatic pheochromocy-toma and a cyst in the distal right femur. The pheochromocytomawas removed by laparoscopy. MRI of the brain showed an asymp-tomatic glioma of the left optic nerve and an arachnoid cyst of theright temporopolar region.

Despite the very mild cutaneous phenotype of the patient, NIHcriteria for clinical diagnosis of neurofibromatosis type I (NF1) werefulfilled with >6 cafe-au-lait spots, >2 neurofibromas of any typeand an optic glioma (see Table 1).

The patient’s parents are 58 and 63 years old. Reportedly, bothhave no signs of NF1.

2.2. Patient B

Patient B presented at the age of 42 years with nocturnal low backpain radiating into the left groin and upper thigh. Clinical examina-tion revealed that the patient had 2–6 nodules thought to be -albeithistologically unconfirmed- subcutaneous neurofibromas. Heshowed no cafe-au-lait spots and no axillary freckling. Hence, the

Fig. 1. MRI scans of patient A showing the sacral region of the spine with pre- and intrasneurofibromas of the peroneal and tibial nerves (B).

Please cite this article in press as: C. Fauth et al., Two sporadic spinaltumour, European Journal of Medical Genetics (2009), doi:10.1016/j.ejmg

criteria for clinical diagnosis of NF1 were not fulfilled in this patient.However, an opthalmological examination in order to evaluate thepresence of Lisch nodules was not performed (Table 1).

MRI showed a paravertebral tumour mass in the left psoasregion (L2–L3), intraspinal neurofibromas in the cervical (C1,C4–C6) and lumbar region (L2–S1) and diffuse enlargement ofnerve roots in all segments of the spine.

The paravertebral tumour in the psoas region was removedsurgically. Histopathology showed a malignant peripheral nervesheath tumour grade II according FNCLCC nomenclature(Fig. 2A).

The intraspinal tumour at C5/C6 led to severe spinal cordcompression with myelopathy. Therefore, laminectomy wasperformed. The tissue consisted mainly of spindle like cells ina collagen matrix with S100 positivity in about 50% of the cells anda low proliferation index, all features typical for benign neurofi-bromas (Fig. 2B). A few areas of this neurofibroma showed anincreased cell density and hyperchromatic nuclei, but low prolif-eration index, reminiscent of a ‘‘cellular’’ [11] or ‘‘atypical’’ [12]neurofibroma (Fig. 2C).

Due to relapse of the MPNST a few months later the patientunderwent several courses of chemotherapy. Two and a half years

acrally growing neurofibromas (A) and the right distal thigh with nodular plexiform

neurofibromatosis patients with malignant peripheral nerve sheath.2009.08.001

Fig. 2. Hematoxilin and eosin staining of tumors removed from patient B: (A) MPNST showing epitheloid cells with polymorphic nuclei and mitotic cells (black arrows). (B) An areaof the intraspinal tumor removed from C5/6 showing spindle shaped cells in a collagen matrix typical for neurofibroma tissue and (C) an area with higher cellularity reminiscent ofa cellular or atypical neurofibroma.

C. Fauth et al. / European Journal of Medical Genetics xxx (2009) 1–6 3

ARTICLE IN PRESS

after the initial diagnosis he died of disseminated metastasis of theMPNST.

2.3. NF1 mutation analysis

NF1 mutation analysis was performed as previously described[13]. Briefly, we amplified the entire NF1 coding region in five partlyoverlapping long-range RT-PCR-products using cDNA generatedfrom RNA that has been extracted from short-term lymphocytecultures treated with puromycin prior to RNA extraction. The entireRT-PCR-products were subsequently sequenced using 18 primers[14,15].

NF1 exons were amplified using standard PCR-protocols withthe following primer pairs: exon 18 forward: 50-CCAGAAGTTGTGTACGTTCTTTTCT-30 and reverse: 50-GGGAAATTTGAGAAGAGATGTAGAG-30, exon 43 forward: 50-CTCCAGGGATGTATTAGAGC-30 reverse:50-CAAGATAACTGTCTTGAATGG-30 and exon 48 forward: 50-GTCTATAAATATTACTCCACTCCC-30 reverse: CAAAAGTCAAGTCAGTTACAAGG-30.

Automated cycle sequencing was performed using BigDyeTerminator Cycle Sequencing Kit (Applied Biosystems, Foster City,CA) and sequencing products were run on an ABI PRISMs3100-XLGenetic Analyzer (Applied Biosystems).

2.4. TP53 mutation analysis

TP53 exons 2–11 with short flanking intron sequences wereamplified in 7 reactions using standard PCR conditions andsequenced by automated cycle sequencing as described above.Primer sequences were kindly provided by the Research GroupHuman Genetics, Division of Medical Genetics UKBB, Departmentof Biomedicine, University of Basel.

2.5. LOH analysis

LOH analysis was performed using microsatellite markers fromchromosome 17 as described [16]. The sequences of the primersand their precise genomic positions, are given in SupplementaryTable 1. PCR products were analyzed on an ABI 3100 sequencerusing a standard fragment analysis protocol. All markers wereanalyzed at least twice by independent PCR experiments. LOH ina tumour sample was scored when the height ratio of the twomarker alleles was at least 20% different from the allele ratioobserved in patient peripheral blood DNA.

2.6. Somatic cell hybrids containing a single chromosome 17

Hybrid cell lines containing either of both chromosomes 17 ofpatient B were generated by fusion of thymidine kinase-deficient

Please cite this article in press as: C. Fauth et al., Two sporadic spinal ntumour, European Journal of Medical Genetics (2009), doi:10.1016/j.ejmg

mouse cell line B82 to lymphoblastoid cells of the patient bypolyethylene-glycol treatment. After selection for several weeksin HAT medium (Hypoxanthine Aminopterin Thymidinemedium) clones containing only one chromosome 17 wereisolated and genotyped using microsatellite markers (Supple-mentary Table 2).

3. Results

3.1. Mutation analysis in patients A and B

Sequencing of the RT-PCR product comprising NF1 exons 35–49of patient A revealed two splicing alterations, both caused bynucleotide change c.7675 þ 1G > A leading to loss of the genuine50ss of intron 43. The majority of the aberrantly spliced transcriptsshowed inclusion of 39 nucleotides of the intronic sequencefollowing exon 43. These transcripts resulting from the use ofa cryptic 50ss contained the mutation and lead to a premature stopcodon. Skipping of exon 43 was observed in a smaller fraction ofaberrantly spliced transcripts. The mutation, c.7675 þ 1G > A, wasconfirmed by genomic DNA sequencing of exon 43 and the flankingintronic sequences. This mutation has already been reported in ourprevious study [13].

In patient B, we identified two sequence alterations. The tran-sition c.3046T > C in NF1 exon 18, leads to the exchange of argininefor a cystein (p.Cys1016Arg). The second alteration is a 3-bpdeletion, c.8131_8133delGTT, in NF1 exon 48 which leads to loss ofone amino acid, p.2711delVal. p.Cys1016Arg has so far been foundonly once in a young sporadic patient, where the mutation wasproven to be de novo, whereas c.8131_8133delGTT has neither beenfound as pathogenic mutation nor as benign polymorphism in morethan 3000 individuals tested for NF1 mutations (L. Messiaenunpublished results). By the analysis of human/mouse hybrid celllines containing the different chromosomes 17 of patient B, weshowed that both mutations are in trans. As the parents were notavailable for testing we could not verify whether either one or bothof the mutations occurred de novo in the patient.

3.2. Analysis of tumour tissues from patients A and B

Paraffin embedded tumour tissue was available from bothpatients. In the MPNST of patient A, all informative markersanalyzed showed LOH extending at least from the distal q-arm tothe p-arm and most likely affecting the entire chromosome(Fig. 3A). Sequence analysis of exon 43 and the flanking intronicsequences showed that LOH affected the wild-type NF1 allele.Furthermore, a heterozygous stop mutation (p.Y126X; c.378C > G)in TP53 exon 5 was found in patient A (Fig. 3B).

eurofibromatosis patients with malignant peripheral nerve sheath.2009.08.001

Fig. 3. (A) LOH analysis in the MPNST of patient A using microsatellite markers. Underlined markers are located within the NF1 gene. Black circles next to the marker indicate LOH.(B) Sequencing of the 50 splice site of the mutation carrying NF1 exon 43 in lymphocytes and the MPNST (tumor tissue) of patient A shows LOH of the wild-type allele andsequencing of the TP53 exon 5 reveals a heterozygous nonsense mutation, p.Y126X, in the MPNST of the patient.

C. Fauth et al. / European Journal of Medical Genetics xxx (2009) 1–64

ARTICLE IN PRESS

Sequence analysis of the exons 18 and 48 clearly showed LOH ofthe exon 18 wild-type sequence in the MPNST of patient B whereasin exon 48 no LOH was observed (Fig. 4B). Analysis of microsatellitemarkers confirmed that LOH in this tumour was affecting the exon18 wild-type allele and restricted to the centromeric region of thelong arm of chromosome 17 with the proximal breakpoint betweenmarker D17S1873 and NF1 exon 18 and the distal breakpointbetween marker IVS27AC28.4 and exon 48 (Fig. 4B; SupplementaryTable 3). We did not find a pathogenic TP53 alteration in the MPNSTof patient B. In contrast to the MPNST, a spinal neurofibromaremoved at C5/C6 did not show LOH (Supplementary Table 3).

4. Discussion

NF1 patients have a 8–13% lifetime risk to develop MPNSTs [17].MPNSTs are thought to arise by malignant transformation ofSchwann cells within a pre-existing plexiform neurofibroma.

Fig. 4. (A) LOH analysis in the MPNST of patient B using microsatellite markers. Underlinedwhite circles no LOH and gray filled circles not analyzed. (B) Sequencing of the alteration-carof patient B shows LOH of the wild-type exon 18 allele in tumor tissue.

Please cite this article in press as: C. Fauth et al., Two sporadic spinaltumour, European Journal of Medical Genetics (2009), doi:10.1016/j.ejmg

Individuals with internal plexiform neurofibromas were 20-timesmore likely to have MPNSTs than individuals without plexiformneurofibromas [18]. To our knowledge we report here for the firsttime the development of MPNSTs in patients with spinal neurofi-bromatosis. The clinical course observed in these two patientssuggests that even asymptomatic and, hence, unrecognized,nodular plexiform neurofibromas and spinal-nerve-root neurofi-bromas in SNF patients bear the risk for malignant transformation.Pain and neuropathy were the first signs indicative of malignanttransformation in both patients. Drouet et al. [19] describe 4/18patients with neuropathies that developed MPNSTs and reporta higher incidence of MPNSTs in patients with neuropathy than inthe general NF1 population. Neuropathy was associated with focalor diffuse paraspinal neurofibromas in 14/18 cases. At least one ofthe patients in this study had no cutaneous neurofibromas at theage of 28, but suffered from diffuse nerve root tumours at multiplelevels of the spine. Most likely this patient also belongs to the

markers are located within the NF1 gene. Black circles next to the marker indicate LOH,rying NF1 exons 18 (ex 18) and 48 (ex 48) in lymphocytes and the MPNST (tumor tissue)

neurofibromatosis patients with malignant peripheral nerve sheath.2009.08.001

C. Fauth et al. / European Journal of Medical Genetics xxx (2009) 1–6 5

ARTICLE IN PRESS

clinical category of SNF. Hence, it is possible that MPNSTs are notuncommon in NF1 patients with multiple spinal neurofibromasand in particular in patients with SNF. Moreover, Drouet et al. [19]identified an association between peripheral neuropathy andsubcutaneous neurofibromas which may also be regarded asnodular neurofibromas along the peripheral nerves. Nodularneurofibromas have been observed also in our SNF patients as wellas in other SNF patients (Bruce Korf personal communications).

Analysis of the MPNST of patient A uncovered LOH most likelyaffecting the entire chromosome 17. This may indicate loss of theentire chromosome with or without reduplication of the remainingchromosome. Further a heterozygous TP53 mutation was found.Heterozygosity of this mutation may result from heterogeneitywithin the malignant tissue with only a proportion of the cellscarrying a hemi- or homozygous TP53 mutation. Alternatively, itmay indicate that LOH in this tumour is caused by loss and redu-plication of the entire chromosome 17 with a TP53 mutation thathas subsequently occurred on one of the chromosomes only. Sincethe former possibility would be in accordance with biallelic TP53inactivation but the latter with monoallelic TP53 loss, it remains tobe elucidated whether biallelic inactivation is necessary or whetherhaploinsufficiency of TP53 may be sufficient to promote malignantprogression [20]. LOH affecting the NF1 locus has been observedmore frequently in MPNSTs than in benign plexiform neurofi-bromas [21]. Whether this indeed indicates that LOH is a morefrequent mutational mechanism for somatic NF1 inactivation inMPNSTs than in benign tumours remains to be elucidated. LOHresulting either from mitotic recombination or deletions isa frequent mechanism for inactivating somatic NF1 mutations inplexiform neurofibromas from which MPNSTs usually derive.However, in these benign tumours LOH is usually restricted to thelong arm of chromosome 17 [16]. Hence, LOH affecting the entirechromosome 17 including the p-arm in the MPNST of our patient Amay reflect chromosomal instability of the malignant tissue ratherthan the second-hit NF1 mutation. In contrast to patient A, theMPNST of patient B exerted LOH only in a limited region ofchromosome arm 17q. Hence, LOH in this tumour may representthe somatic NF1-inactivating second-hit mutation.

The molecular mechanisms that lead to the peculiar phenotypeof SNF are largely unknown. Nevertheless, it is striking that the NF1mutations identified in our patients fit into the mutational spec-trum of SNF patients [22,23] which appears to be overrepresentedfor missense and underrepresented for nonsense and frameshiftmutations when compared to the mutational spectrum of thegeneral NF1 population [14].

Mutation c.7675 þ 1G > A identified in patient A affects splicingof the rarely mutated NF1 exon 43. Although, in cultivated bloodlymphocytes, the majority of transcripts derived from the mutatedallele contain a premature stop codon and may therefore be subjectto nonsense mediated decay, a smaller proportion of the transcriptsfrom the mutated allele skip exon 43. Skipping of the in frame exon43 has also been shown in a significant proportion of transcripts ofsome tissues of healthy individuals [24]. Therefore, it is conceivablethat the protein products of NF1 transcripts lacking exon 43 retainsome functionality. Due to the lack of relevant tissue of the patientit is not possible to test whether the splicing pattern caused by thismutation differs in other tissues, but this is possible, since it hasbeen shown that the aberrant transcripts resulting from splicingmutations may vary in different tissues of NF1 patients [25] as inpatients with other genetic disorders [26].

In patient B we identified two novel non-truncating alterationsin trans. Their effect on neurofibromin function is unknown sincethey are both located outside the functionally characterizeddomains of this large protein. The dramatic chemical change froma sulphur-containing to a basic amino acid and the fact that the

Please cite this article in press as: C. Fauth et al., Two sporadic spinal ntumour, European Journal of Medical Genetics (2009), doi:10.1016/j.ejmg

affected highly conserved (down to Danio rerio and Drosophilamelanogaster) amino acid Cys1016 has been found as a de novoalteration in a 4-year old girl with pigmentary signs of NF1 (L.Messiaen unreported results) strongly suggests that the alterationp.Cys1016Arg is pathogenic. PolyPhen a web-based tool to predictthe functional effect of SNPs also rates this amino acid change asprobably damaging. Furthermore, loss of the exon 18 wild-typeallele in the MPNST of this patient argues in favour of the patho-genicity of this misssense alteration. The pathogenicity of thesecond alteration, p.2711delVal, is questionable. The finding thatthe benign spinal neurofibroma of the patient did not show LOHstill allows for the speculation that p.2711delVal affecting a highlyconserved amino acid leads to a gene product with reduced butresidual function. Nevertheless, we currently assume thatp.2711delVal is a most likely benign unclassified variant andp.Cys1016Arg represents the pathogenic mutation responsible forthe phenotype of patient B that resembles SNF without CLS ormultiple isolated neurofibromas or a combination of these twoalternate forms of neurofibromatosis that have been proposed torepresent a spectrum of manifestations of one NF1 form [27].

Taken together, the nature and localisation of the identified NF1mutations in our patients support the notion that the peculiarlymild cutaneous phenotype of SNF may be related to a possibleresidual function of the mutated gene product. Currently, however,it is difficult to conceive how the mutation type would lead to thestriking spinal involvement that has been reported to be consistentin familial SNF cases [5,6,9], but can also be observed in someclassical NF patients [28,29]. It has been proposed that plexiform-including paraspinal plexiform neurofibromas and dermalneurofibromas may originate from distinct progenitor cells [30].Hence, it may be speculated that the mutations identified in SNFpatients impair neurofibromin function to different extent in thesedistinct progenitor cells leading to the plexiform paraspinalphenotype in the absence of cutaneous neurofibromas.

The identification of NF1 mutations in our SNF patients hasbroader implications also for affected family members. Due to theextremely mild cutaneous presentation of NF1 both were suspectedto suffer from NF1 only when they developed MPNSTs in theirthirties and fourties, respectively. Asymptomatic or oligosympto-matic younger siblings or children with the mutation may alsodevelop symptomatic spinal or peripheral nodular neurofibromasthat bear the risk of malignant transformation. Therefore, appro-priate counselling and, if desired, genetic testing should be offeredto those relatives.

Acknowledgements

We thank Drs. B. Scheithauer, B. Korf and M. Coulibaly-Wimmerfor helpful discussions on pathological, clinical and radiologicalfindings. We are grateful to Dr. R. Silye for providing us with paraffintissue of tumours removed from patient B. The authors would alsolike to thank Dr. Eric Legius for helpful discussions and for orga-nizing workshops on the molecular biology of NF1 that were fundedby FWO-Flanders grant no. WO.027.09 and gave the collaborators ofthis publication the opportunity for fruitful discussions.

Appendix. Supplementary material

Supplementary data associated with this article can be found inthe online version at doi:10.1016/j.ejmg.2009.08.001.

References

[1] D. Easton, M. Ponder, S. Huson, B. Ponder, An analysis of variation in expres-sion of neurofibromatosis (NF) type 1 (NF1): evidence for modifying genes,Am. J. Hum. Genet. 53 (1993) 305–313.

eurofibromatosis patients with malignant peripheral nerve sheath.2009.08.001

C. Fauth et al. / European Journal of Medical Genetics xxx (2009) 1–66

ARTICLE IN PRESS

[2] D.H. Gutmann, A. Aylsworth, J.C. Carey, B. Korf, J. Marks, R.E. Pyeritz,A. Rubenstein, D. Viskochil, The diagnostic evaluation and multidisciplinarymanagement of neurofibromatosis 1 and neurofibromatosis 2, JAMA 278(1997) 51–57.

[3] S.M. Huson, The neurofibromatosis: classification, clinical features andgenetic counselling, in: D. Kaufmann (Ed.), Neurofibromatosis, Karger, Basel,2008, pp. 1–20.

[4] D. Stumpf, J. Alksne, J. Annegers, S. Brown, P. Conneally, D. Housman,M. Leppert, J. Miller, M. Moss, A. Pileggi, I. Rapin, R. Strohman, L. Swason,A. Zimmerman, Neurofibromatosis, Arch. Neurol. 45 (1988) 575–578.

[5] S. Pulst, V. Riccardi, P. Fain, J. Korenberg, Familial spinal neurofibromatosis:clinical and DNA linkage analysis, Neurology 41 (1991) 1923–1927.

[6] E. Ars, H. Kruyer, A. Gaona, P. Casquero, J. Rosell, V. Volpini, E. Serra, C. Lazaro,X. Estivill, A clinical variant of neurofibromatosis type 1: familial spinalneurofibromatosis with a frameshift mutation in the NF1 gene, Am. J. Hum.Genet. 62 (1998) 834–841.

[7] D. Kaufmann, R. Muller, B. Bartelt, M. Wolf, K. Kunzi-Rapp, C.O. Hanemann,R. Fahsold, C. Hein, W. Vogel, G. Assum, Spinal neurofibromatosis withoutcafe-au-lait macules in two families with null mutations of the NF1 gene, Am.J. Hum. Genet. 69 (2001) 1395–1400.

[8] I. Pascual-Castroviejo, S.I. Pascual-Pascual, R. Velazquez-Fragua, P. Botella,J. Viano, Familial spinal neurofibromatosis, Neuropediatrics 38 (2007) 105–108.

[9] M. Poyhonen, E. Leisti, S. Kytola, J. Leisti, Hereditary spinal neurofibromatosis:a rare form of NF1? J. Med. Genet. 34 (1997) 184–187.

[10] L. Messiaen, V. Riccardi, J. Peltonen, O. Maertens, T. Callens, S.L. Karvonen,E.L. Leisti, J. Koivunen, I. Vandenbroucke, K. Stephens, M. Poyhonen, Inde-pendent NF1 mutations in two large families with spinal neurofibromatosis,J. Med. Genet. 40 (2003) 122–126.

[11] B.W. Scheithauer, C. Giannini, J.M. Woodruff, et al., Perineurinoma, in:P. Kleihues., W.K. Cavenee (Eds.), World Health Organization Classification ofTumours. Pathology and Genetics: Tumours of the Nervous System, Interna-tional Agency for Research and Cancer (IRAC), Lyon, 2000, pp. 169–171.

[12] G.P. Nielsen, A.O. Stemmer-Rachamimov, Y. Ino, M.B. Moller, A.E. Rosenberg,D.N. Louis, Malignant transformation of neurofibromas in neurofibromatosis 1 isassociated with CDKN2A/p16 inactivation, Am. J. Pathol. 155 (1999) 1879–1884.

[13] K. Wimmer, X. Roca, H. Beiglbock, T. Callens, J. Etzler, A.R. Rao, A.R. Krainer,C. Fonatsch, L. Messiaen, Extensive in silico analysis of NF1 splicing defectsuncovers determinants for splicing outcome upon 5’ splice-site disruption,Hum. Mutat. 28 (2007) 599–612.

[14] L.M. Messiaen, K. Wimmer, NF1 mutational spectrum, in: D. Kaufmann (Ed.),Neurofibromatosis, vol. 16, Karger, Basel, 2008, pp. 63–77.

[15] I. Vandenbroucke, R. van Doorn, T. Callens, J.M. Cobben, T.M. Starink,L. Messiaen, Genetic and clinical mosaicism in a patient with neurofibroma-tosis type 1, Hum. Genet. 114 (2004) 284–290.

[16] K. Steinmann, L. Kluwe, R.E. Friedrich, V.F. Mautner, D.N. Cooper, H. Kehrer-Sawatzki, Mechanisms of loss of heterozygosity in neurofibromatosis type 1-associated plexiform neurofibromas, J. Invest. Dermatol. (2008).

[17] D.G. Evans, M.E. Baser, J. McGaughran, S. Sharif, E. Howard, A. Moran, Malig-nant peripheral nerve sheath tumours in neurofibromatosis 1, J. Med. Genet.39 (2002) 311–314.

Please cite this article in press as: C. Fauth et al., Two sporadic spinaltumour, European Journal of Medical Genetics (2009), doi:10.1016/j.ejmg

[18] T. Tucker, P. Wolkenstein, J. Revuz, J. Zeller, J.M. Friedman, Association betweenbenign and malignant peripheral nerve sheath tumors in NF1, Neurology 65(2005) 205–211.

[19] A. Drouet, P. Wolkenstein, J.P. Lefaucheur, S. Pinson, P. Combemale,R.K. Gherardi, P. Brugieres, J. Salama, P. Ehre, P. Decq, A. Creange, Neuro-fibromatosis 1-associated neuropathies: a reappraisal, Brain 127 (2004)1993–2009.

[20] R.A. Lothe, B. Smith-Sorensen, M. Hektoen, A.E. Stenwig, N. Mandahl, G. Saeter,F. Mertens, Biallelic inactivation of TP53 rarely contributes to the developmentof malignant peripheral nerve sheath tumors, Genes Chromosomes Cancer 30(2001) 202–206.

[21] M. Upadhyaya, L. Kluwe, G. Spurlock, B. Monem, E. Majounie, K. Mantripragada,M. Ruggieri, N. Chuzhanova, D.G. Evans, R. Ferner, N. Thomas, A. Guha,V. Mautner, Germline and somatic NF1 gene mutation spectrum in NF1-associated malignant peripheral nerve sheath tumors (MPNSTs), Hum. Mutat.29 (2008) 74–82.

[22] L. Messiaen, T. Callens, J.B. Williams, D. Babovic-Vuksanovic, S.M. Huson,E. Legius, R. Mac Gardner, I. Pascual-Castroviejo, S. Plotkin,G.B. Schaefer, M. Wilson, B. Korf, Genotype–Phenotype Correlations inSpinal NF, The American Society of Human Genetics, San Diego,California, 2007.

[23] M. Upadhyaya, G. Spurlock, L. Kluwe, N. Chuzhanova, E. Bennett, N. Thomas,A. Guha, V. Mautner, The spectrum of somatic and germline NF1 mutations inNF1 patients with spinal neurofibromas, Neurogenetics (2009).

[24] I. Vandenbroucke, J. Vandesompele, A. De Paepe, L. Messiaen, Quantification ofNF1 transcripts reveals novel highly expressed splice variants, FEBS Lett. 522(2002) 71–76.

[25] H.V. Leskela, T. Kuorilehto, J. Risteli, J. Koivunen, M. Nissinen,S. Peltonen, P. Kinnunen, L. Messiaen, P. Lehenkari, J. Peltonen,Congenital pseudarthrosis of neurofibromatosis type 1: impaired osteo-blast differentiation and function and altered NF1 gene expression, Bone44 (2009) 243–250.

[26] O. Chiba-Falek, R.B. Parad, E. Kerem, B. Kerem, Variable levels of normal RNAin different fetal organs carrying a cystic fibrosis transmembrane conduc-tance regulator splicing mutation, Am. J. Respir. Crit. Care Med. 159 (1999)1998–2002.

[27] B.R. Korf, J.W. Henson, A. Stemmer-Rachamimov, Case records of the Massa-chusetts General Hospital. Case 13-2005. A 48-year-old man with weakness ofthe limbs and multiple tumors of spinal nerves, N Engl. J. Med. 352 (2005)1800–1808.

[28] L. Kluwe, M. Tatagiba, C. Funsterer, V.F. Mautner, NF1 mutations and clinicalspectrum in patients with spinal neurofibromas, J. Med. Genet. 40 (2003)368–371.

[29] K. Wimmer, M. Muhlbauer, M. Eckart, T. Callens, H. Rehder, T. Birkner,J.G. Leroy, C. Fonatsch, L. Messiaen, A patient severely affected by spinalneurofibromas carries a recurrent splice site mutation in the NF1 gene, Eur.J. Hum. Genet. 10 (2002) 334–338.

[30] L.Q. Le, T. Shipman, D.K. Burns, L.F. Parada, Cell of origin and microenviron-ment contribution for NF1-associated dermal neurofibromas, Cell Stem Cell 4(2009) 453–463.

neurofibromatosis patients with malignant peripheral nerve sheath.2009.08.001