ORIGINAL PAPER

Molecular evolution of a neurofibroma to malignant peripheralnerve sheath tumor (MPNST) in an NF1 patient: correlationbetween histopathological, clinical and molecular findings

Gill Spurlock • Samantha J. L. Knight • Nick Thomas •

Tim-Rasmus Kiehl • Abhijit Guha • Meena Upadhyaya

Received: 23 December 2009 / Accepted: 15 February 2010 / Published online: 15 March 2010

� Springer-Verlag 2010

Abstract

Objective Neurofibromatosis type 1 (NF1) patients have a

13% risk of developing a malignant peripheral nerve sheath

tumor (MPNST). Many MPNSTs are histopathologically

complex, with regions exhibiting features of the original

benign plexiform neurofibroma (PNF), of an atypical PNF,

or of MPNST showing varying degrees of de-differentiation.

This study analyzed the genetic alterations associated with

this pathological heterogeneity in order to identify the

genetic processes involved in transformation from a benign

to an aggressive malignant tumor.

Methods A histological and molecular analysis of a sin-

gle MPNST tumor that was subdivided into three histop-

athologically distinct regions, a benign PNF (region 1), an

atypical PNF (region 2), and a high-grade MPNST (region

3), was carried out. Tumor DNA from each region was

analyzed in conjunction with the patient’s lymphocyte

DNA. Somatic mutation analyses included loss-of hetero-

zygosity (LOH), MLPA analysis, NF1 gene sequencing,

and a microarray comparative genomic hybridisation (array

CGH) analysis.

Results The patient had a germline NF1 splice-site

mutation. The NF1-associated LOH analysis found that

LOH increased in the three tumor areas, with 9, 42, and

97% LOH evident in regions 1, 2, and 3, respectively.

Additional genetic changes, including losses of TP53, RB1,

CDKN2A, and of several oncogenes and cell-cycle genes,

were found only in the malignant MPNST (region 3). Array

CGH also identified genomic gains and losses in DNA

from region 3.

Discussion This is the first study that correlates the his-

tological and molecular changes associated with MPNST

development, confirming the significant cellular and

genetic heterogeneity that poses both diagnostic and thera-

peutic challenges.

Keywords Neurofibromatosis type 1 (NF1) �Plexiform neurofibroma (PNF) � Atypical neurofibroma �Malignant peripheral nervous system tumor (MPNST) �Tumor suppressor gene (TSG) � Germline mutation �Loss-of-heterozygosity

Introduction

Neurofibromatosis type 1 (NF1) is the most common

autosomal dominant neurogenetic disorder, affecting

about 1 in 3,000 individuals worldwide, resulting from

mutations of the NF1 gene located at 17q11.2. The dis-

ease is characterized by pigmentary skin changes and the

growth of a variety of benign and malignant tumors that

G. Spurlock � N. Thomas � M. Upadhyaya (&)

Institute of Medical Genetics, Cardiff University,

Heath Park, Cardiff CF14 4XN, Wales, UK

e-mail: Upadhyaya@cardiff.ac.uk

S. J. L. Knight

NIHR Biomedical Research Centre, Oxford,

Wellcome Trust Centre For Human Genetics,

Roosevelt Drive,

Oxford OX3 7BN, UK

T.-R. Kiehl

Department of Pathology, University Health Network,

Toronto, Ontario, Canada

A. Guha

The Arthur and Sonia Labatt Brain Tumor Research Centre,

The Hospital for Sick Children, Department of Surgery,

Division of Neurosurgery, Western Hospital,

University of Toronto, Toronto, Ontario, Canada

123

J Cancer Res Clin Oncol (2010) 136:1869–1880

DOI 10.1007/s00432-010-0846-3

develop in association with both the peripheral and central

nervous system (Upadhyaya and Cooper 1998; Bennett

et al. 2009). NF1 patients develop multiple benign dermal

neurofibromas that do not become malignant, while 30–

50% of NF1 patients also develop plexiform neurofibro-

mas (PNF) that usually grow along a major nerve. NF1

patients have a 8–13% lifetime risk of developing a

malignant peripheral nerve sheath tumor (MPNST), a

highly aggressive malignant neoplasm associated with

significant morbidity and mortality for affected patients

(Evans et al. 2002). MPNSTs usually grow out of either a

pre-existing PNF or a focal subcutaneous neurofibroma

(Brems et al. 2009a). Indeed, NF1 patients with internal

PNFs have a 20-fold increased risk of developing an

MPNST compared to NF1 individuals without such

tumors (Tucker et al. 2005).

MPNST in NF1 patients carries a poor prognosis, with a

5-year survival rate of only 21–41%. Early detection fol-

lowed by radical surgery in combination with chemo- and

radiotherapy is the current best standard of treatment

(Angelov et al. 1998). However, several obstacles do exist.

First, there are no radiological features on CT or MRI

which clearly differentiate a large benign PNFs from one

that has transformed into an MPNST. Second, pathological

diagnosis of an MPNST is also challenging, mainly related

to the current paucity of well-defined histological criteria, a

specific antigenic marker, and significant tissue heteroge-

neity. Much of the heterogeneity within MPNSTs is due to

their initial growth from a benign PNF, followed by pro-

gression of different regions independently into varying

grades of malignancy. A region of the MPNST may

therefore be designated as ‘atypical’, demonstrating histo-

logical features of both a benign PNFs and a low grade

MPNST, with hypercellularity and hyperchromatic nuclei,

however with relatively low mitotic activity. Atypical

PNFs may be asymptomatic or can be associated with pain

and neurological deficit. In regions of the tumor consistent

with frank MPNST, based on high cellularity, nuclear

pleomorphism, and mitotic activity, there can be further

pathological heterogeneity as to grade of MPNST as well

as degree of de-differentiation. One aspect of such

de-differentiation of a high-grade MPNST into skeletal

muscle, resembling rhabdomyosarcoma, is referred to as a

malignant triton tumor (MTT) (Stasik and Tawfik 2006).

Given the difficulties in the early clinical and pathological

diagnosis of MPNST, increased knowledge of the molec-

ular basis of NF1-associated malignant transformation is

clearly essential. Although some previous related studies

have examined the molecular genetics associated with

MPNST development in different patient cohorts, this is

the first study to characterize the molecular alterations

present in different regions of a single large histologically

heterogenous MPNST.

Materials and methods

Clinical details

This 25-year-old man with a family history of NF1 (on the

paternal side) had two unaffected siblings. Clinical features

consistent with a diagnosis of NF1 included bilateral

axillary freckling, cafe-au-lait spots (CALS), multiple

cutaneous neurofibromas, and significant learning disabil-

ity. The patient had a longstanding asymptomatic mass on

the medial-posterior aspect of his right calf, which on MRI

was consistent with a large PNF of the right posterior tibial

nerve. The patient presented with recent onset of balance

difficulty and associated mild episodic headaches. A brain

MRI revealed a new exophytic mass (*2 cm3) extending

from the right posterolateral aspect of the lower medulla

with resultant distortion of the IV ventricle and mild

hydrocephalus. The gadolinium-enhancing lesion was

removed, leading to resolution of his symptoms. Histo-

pathologic examination revealed the expectant pilocytic

astrocytoma.

Six months post-craniotomy, the patient complained of

growth of the PNF in his right calf and an associated

increase in pain. An MRI scan of the right leg was essen-

tially unchanged from his previous scans, with whole body

MRI revealing no other significant internal masses. Given

the evidence of recent increased growth and pain associ-

ated with the calf tumor, an open surgical examination was

recommended. Multiple tissue samples were taken from the

tumor for histological assessment (Fig. 1) and identified

complex histopathological changes in different regions of

this single contiguous tumor. One area of the tumor had

features of a benign PNF that lacked any of the overt

histological features of malignancy (region 1). A second

area had the morphology of an atypical neurofibroma

showing a trend toward low-grade malignancy (region 2).

In contrast, region 3 was an unequivocal MPNST, with

areas of necrosis and several scattered regions with rhabdo

myoblastic de-differentiation, features representing a

malignant triton tumor (MTT) (Fig. 1). Desmin and

myogenin staining, often associated with such rhabdoid/

skeletal muscle de-differentiation (Altmannsberger et al.

1985;Kumar et al. 2000), was focally positive in some

tumor cells, as well as in the skeletal muscle fibers

entrapped within the periphery of the tumor. While S-100

protein was also focally positive in some tumor cells, there

was no staining with the melanocytic markers HMB-45 and

MART-1 (Gowan et al. 1986;Kawakami et al.1994; Coulie

et al. 1994). The final pathological diagnosis was grade IV

MPNST exhibiting rhabdo myoblastic de-differentiation.

The patient underwent radiotherapy (3,500 rads) to the

tumor and 3 cm of adjacent soft tissue, with subsequent

attempt at radical limb-sparing surgery at 8 weeks

1870 J Cancer Res Clin Oncol (2010) 136:1869–1880

123

post-open biopsy. However, at the second surgery, the

vascular supply to the lower leg was compromised by

tumor invasion, so an above-knee surgical amputation was

performed. The patient remains well and fully ambulatory

four years post-diagnosis. However, a recent chest X-ray

suggests a solitary lung metastasis.

Patient samples

From our institutional research board approved tumor bank

(http://www.braintumorbank.ca/index), multiple samples

were taken from the three regions for quick section, for-

malin fixation (10%) plus paraffin embedding (FFPE), and

also flash frozen.. DNA and RNA were extracted from

tumor tissues as previously described (Upadhyaya et al.

2004, 2008a, b). These samples underwent immunohisto-

chemical and molecular analysis.

Tissue preparation and immunohistochemistry (IHC)

IHC was undertaken on 5-lm-thick FFPE sections, which

were de-paraffinized, rehydrated in xylene, and micro-

waved in 10 mmol/L citrate buffer. Positive controls

included tissues known to express the antigen, while

negative controls omitted the primary antibody. After

incubation with the primary antibodies (KI-67 (MIB-1-

Dako,Carpinteria, CA)-1:100; anti-Desmin (clone D33-

Dako)-1:200; anti-Myogenin (clone L026-Novocastra,

Newcastle,UK-1:50); anti-S-100 (predilutes-Ventana

Medical Systems, Tucson, AZ), detection was undertaken

with streptavidin/biotin peroxidase complex technique and

3,3-diaminobenzidine. All immunohistochemical studies

were performed in a Ventana automated stainer.

Molecular studies

A- NF1 mutation analysis

Our comprehensive NF1 gene mutation screen protocol

involves: (1) a direct DNA sequencing of all amplified

exons and cDNA fragments (ABI3730 sequencer, Applied

Biosystems, Warwick, UK); (2) an assessment of loss-

of-heterozygosity (LOH) and/or of genomic imbalance

across the NF1 gene region; and (3) a multiplex ligation-

dependent probe amplification (MLPA) analysis to identify

small deletions or duplications involving adjacent exons

(NF1 gene-specific MLPA Assay kits (P081/082 and P122,

MRC Holland, Amsterdam, NL) as previously reported

(Upadhyaya et al. 2004, 2008a).

Extracted tissue RNA was reverse-transcribed and the

entire NF1 coding region amplified in 24 overlapping

fragments (Upadhyaya et al. 2004). Any mutation identi-

fied at the cDNA level was subsequently confirmed by

direct-cycle sequencing of genomic DNA. The protocol

used to screen for NF1 somatic mutations in tumor DNA

has been described (Upadhyaya et al. 2008b). The Human

Genome Variation Society (www.hgvs.org/mutnomen)

recommended nomenclature was used to describe the

identified NF1 mutations. The nucleotide numbering is

based on the NF1 Ref Seq mRNA sequence (GenBank

accession number NM_000267.1) with the first nucleotide

being the A of the ATG translation initiation codon.

High Grade MPNST with Rhabdomyoblastic Features

PNF with high cellularity

Low Grade MPNST

Region 3

Region 1

Region 2

Myogenin

Fig. 1 Histopathological

examination of the complex

MPNST tumor from the NF1

patient. The three regions

biopsied represent the areas of

the tumor histologically

corresponding to (i) a plexiform

neurofibroma showing elevated

cellularity (region 1); (ii) an

atypical neurofibroma

exhibiting low-grade

malignancy (region 2); and (iii)

an unequivocal high-grade

MPNST, displaying both

necrosis and areas of

rhabdomyoblastic

differentiation (region 3

immunostain for myogenin).

Microscopy, 5-lm sections

stained with hematoxylin and

eosin

J Cancer Res Clin Oncol (2010) 136:1869–1880 1871

123

B- loss-of-heterozygosity (LOH) Studies

NF1 (17q11.2) LOH analysis across the entire NF1 gene

region used a panel of fluorescently labeled microsatellite

and polymorphic SNP-based markers spanning the geno-

mic region encompassing the NF1 gene region (Stewart

et al. 2008). A list of the markers included in this study is

given in Fig. 2.

TP53 (17p13) LOH analysis of the genomic region con-

taining the TP53 gene used four polymorphic markers: the

TP53 intragenic penta-nucleotide VNTR marker (Futreal

et al. 1991) and three flanking microsatellite markers for

D17S796, D17S804, D17S520 (Stewart et al. 2008).

CDKN2A (p16) (9p21) LOH analysis at the CDKN2A

locus used four microsatellite markers, including D9S304

(Hartmann et al. 2000), D9S1748, D9S1751 (Cairns et al.

1995), and D9S942 (Pollock et al.2001).

RB1 (13q14.2): Four microsatellite markers from across

the RB1 locus, D13S118, D13S119, D13S153, and B57

(Belchis et al. 1996) were used for the LOH analysis.

Array comparative genomic hybridization (aCGH) DNA

was extracted from the FFPE samples using a Qiagen kit.

DNA from the high-grade MPNST (region 3) was of suf-

ficient yield to permit microarray analysis. Because of

insufficient lymphocyte DNA available from this patient,

the DNA sample from region 3, together with a sex-

matched normal genomic DNA reference sample, was dif-

ferentially labeled (with ULS-Cy3 and ULS-Cy5) using the

Agilent Technologies ULS labeling kit according to

manufacturer’s instructions http://www.chem.agilent.com/

Library/usermanuals/Public/G4410-90020_CGH_ULS(FFPE).

pdf Publication Summary. aspx?whid = 53,004&liid =

3,786). The differentially labeled DNAs were co-hybridized

to a standard Agilent 244Koligonucleotide array for 48 h at

65�C in a rotating oven (Agilent Technologies). The

hybridized array was then washed according to Agilent

Technologies protocols, with the exception that the final

stabilization step was not performed. Immediately follow-

ing washing, the hybridized array was scanned at 5-lm

resolution in an Agilent Microarray Scanner. The image

data were extracted using Agilent Feature Extraction soft-

ware version 8.5 and the data analyzed using Agilent CGH

Analytics software version 3.4 (500 kb and 1 Mb windows

and a z-score method settings). All putative genomic

imbalances were recorded, and the approximate positions of

immediately proximal and distal oligonucleotides that

flanked each imbalance were noted (Human genome Build

35). For all identifiable genomic changes B2 Mb, the

Database of Genomic Variants (http://projects.tcag.ca/

variation/) was interrogated to determine whether any of

the identified variants had been previously found in control

individuals.

Results

Germline NF1 mutation

The initial NF1 germline mutation screen of the patient’s

genomic DNA failed to identify a causative mutation,

despite direct sequencing of the NF1 coding region. A

mutation analysis of total RNA extracted from the patient’s

dermal neurofibroma, however, identified a splicing

mutation in intron 12b (c.2002–14, C [ G). This mutation

was subsequently confirmed in DNA from the patient’s

lymphocytes and MPNST (Fig. 3).

13.3

13.2

12

13.1

11.2

11.2

11.1

12

21.1

21.2

21.31

21.3221.33

22

23.1

23.323.2

24.1

24.2

24.3

25.1

25.2

25.3

D17S520

EVI20

HHH202

EXON 5

I12b

J1/j2

IVS27AC2

IVS38GT5

3’NF1-1

D17S798

EW206

EW207

D17S796

TP53

D17S938

D17S804

Fig. 2 Idiogram of chromosome 17. Markers used for 17q11.2 and

17p13 analysis are indicated

1872 J Cancer Res Clin Oncol (2010) 136:1869–1880

123

Somatic NF1 mutations

A large (C1.4 Mb) genomic somatic NF1 deletion was

identified in the DNA from the high-grade region of the

MPNST by the MLPA and also in DNA from the patient’s

medullary pilocytic astrocytoma but not in DNA from a

dermal neurofibroma previously removed from the patient

(data not shown). There was also no evidence of LOH of

the NF1-region in this neurofibroma DNA.

Mutation analysis of tumor regions

As noted earlier and in Fig. 1, the pathologically heterogenous

MPNST was subdivided it into three regions and DNA was

extracted from FFPE tissue sections from each of these

regions. Pilocytic astrocytoma DNA was also available for

mutation analysis. Tumor DNA samples were initially

screened for LOH of the NF1-gene region (Fig. 4). As

expected, the patient’s lymphocyte DNA showed no evidence

of LOH, while a 42% level of NF1-associated LOH was seen

in the pilocytic astrocytoma DNA (Fig. 4). The levels of NF1-

LOH observed in the heterogenous tumor showed significant

regional variation, with little evidence of LOH from region 1,

about 40% LOH in the region 2, and almost complete LOH

(85 [ 97%) found in region 3 (Fig. 4).

In addition to LOH of the NF1-region in DNA from

region 3, LOH affecting the TP53, CDKN2A, and RB1 loci

was also found (Fig. 5). There was no evidence for LOH

involving these additional gene loci that have previously

been implicated in MPNST development (Brems et al.

2009a) in DNA either from the region 1 or from the pilo-

cytic astrocytoma. All LOH analyses were confirmed on

DNA isolated from three different tissue biopsies taken

from each region.

Microarray CGH (aCGH)

A lack of tumor DNA of sufficient quantity and quality

somewhat hindered this final part of the study, in which

most of the usable DNA was that isolated from the FFPE

sections of the high-grade MPNST (region 3). While suf-

ficient DNA was available from this tumor region, but not

from either region 1 or 2, the quality of such DNA was

often not optimal for aCGH analysis. However, while the

aCGH data obtained from this FFPE DNA did prove to be

‘noisy’, careful analysis of this data did reveal changes to

several genomic regions that exhibited gross imbalances

(Table 1). Comparison of the \2-Mb genomic changes

identified in the tumor DNA against entries in the Database

of Genomic Variants did identify two genomic regions that

Exon 13 Intron 12b

cDNA sequence of exon 12b and exon 13 with inclusion of 13 bp of intronic sequence:

GAAGGGAAAAGGGAACTCCTCTATGCCTTGACTCTCAGGATAGTGCAGCAGGATG

13 bp of intronic sequence included in final transcript with the new 'in-frame' nonsense mutation indicated

Exon 12b Exon 13

A

B

Fig. 3 Identification of germline NF1 mutation in mRNA and DNA

from the patient. a Demonstrates insertion of 13 bp of intron 12b

sequence between the mRNA sequence derived from exons 12b

and 13, with the new ‘in-frame’ nonsense mutation indicated.

b Confirmation in patient’s lymphocyte DNA that the c.2002-

14C[G mutation creates a new acceptor splice site in intron 12b and

results in the incorporation of the last 13 bp of intron 12b (yellow bar)

in the final NF1 transcript

J Cancer Res Clin Oncol (2010) 136:1869–1880 1873

123

overlapped with genomic regions reported to be variants in

the normal population, located at 8p23.3 and at 15q11.2.

Several of the genomic imbalances identified in the high-

grade tumor DNA did involve regions, and associated

genes, reported to be modified in MPNSTs. These included

gains at both the 5p13–p15 region, possibly involving the

NKD2, IRX4, IRX1, IRX2, and TRIO genes and gains at

17q21–q25 affecting the TOP2A, ETV4, HOXB7, BIRC5,

Pilocytic Astrocytoma

42% LOH

Patient's Blood DNANo LOH

Region 3 (Frozen Tissue)

85% LOH

Region 1

9% LOH

97% LOH

Region 3 (Paraffin section)

Region 2

40% LOH

Fig. 4 A LOH (loss-of-heterozygosity) or allelic imbalance analysis

of the NF1 gene region in DNA from the patient’s lymphocytes,

pilocytic astrocytoma, and from the three subregions of the large

complex MPNST tumor. The level of LOH is calculated from a

comparison of allele size for each informative marker. The results for

the NF1 30UTR marker are shown. The greatest levels of LOH are

associated with the DNA from region 3 (high-grade MPNST), with

decreasing LOH levels evident in the DNA from region 2 (atypical

neurofibroma) and region 1 (plexiform neurofibroma). Pilocytic

astrocytoma DNA shows an intermediate level of LOH at this marker

Fig. 5 Gene-specific LOH

analysis involving the TP53,

CDKN2A, and RB1 loci in DNA

from regions 1 and 3 of the

tumor. DNA from region 3

(high-grade MPNST)

demonstrates LOH at all three

loci, while that from region 1

(plexiform neurofibroma) shows

no LOH at any of these loci

1874 J Cancer Res Clin Oncol (2010) 136:1869–1880

123

Table 1 Summary of Gross Genome Imbalances Identified in MPNST DNA

ChromosomeRegion

Approximatestartpoint (bpposition)a

Approximateendpoint (bpposition)a

Overall sizeof genomicchange (bp)

Gain/loss

Previously reportedgene alterations

Previously reportedgenomic regionalterations

1p36.33–1p11.2 876656 120962972 120086316 Loss 1p35–33 loss, 1p21 loss

2q11.2 99864884 100149464 284580 Gain 2q11.2–q13 gain

4p16.1 5983561 10687582 4704021 Loss

4q25 110206691 110567249 360558 Gain

5p15.33–5p11b 75149 46136124 46060975 Gain NKD2,IRX4, IRX1, IRX2,TRIO

5p12

5q11.1–5q13.2 49595677 68834857 19239180 Loss 5q11.2 gain

5q13.3–5q14.1 74686092 77108298 2422206 Gain 5q14 gain

5q14.1–5q35.3b 77108298 180605972 103497674 Loss AP3B1, PDGFRB 5q21–q23, 5q31–q33, 5q34–q35.3 gain

6p25.1 5396503 6761132 1364629 Loss

6p24.3 8749335 10261966 1512631 Loss

6p24.1–6p23 12833204 13697123 863919 Loss

6p23–6p22.3 13697123 18810292 5113169 Gain 6p23(–p21) gain

6p22.3–6p21.33 19609545 31537963 11928418 Loss (6p23)–p21 gain

6p21.31–6p21.1 36666488 41475004 4808516 Loss (6p23)–p21 gain

6p21.1 41475004 44297052 2822048 Gain (6p23)–p21 gain

6p21.1–6q11.1 44297052 62040949 17743897 Loss (6p23)–p21 gain; 6p12 gain

7q21.11 80961488 81538684 577196 Gain HGF 7q21 gain; 7q11.23–q21.11gain

7q21.11 83102320 83938526 836206 Gain 7q21 gain; 7q11.23–q21.11gain

8p23.3b 176477 1519157 1342680 Loss

8q23.3 116490914 116726990 236076 Gain 8q22–24

9p24.3–9p23b,c 204367 14064145 13859778 Loss

9p22.3–9q21.3d 14309445 23405621 9096176 Loss CDKN2A, MTAP 9p21.3 loss, 9q21.32–q22.33gain

9q22.31 91668999 92245287 576288 Gain

10p12.1–10p11.21 25399466 35904502 10505036 Loss

10q11.22–10q26.3b 47970511 135395607 87425096 Loss 10q25 loss

11p15.5–11p15.2b 200300 14255898 14055598 Loss

11q14.1–11q24.1 78548953 123417159 44868206 Loss MMP12 11q22–23 loss

13q14.13–13q34b 45621017 114123908 68502891 Loss RB1 13q22.1–q22.2 gain

14q13.1–14q32.2 32578274 97692159 65113885 Loss 14q22

15q11.2 21577499 22607314 1029815 Loss

15p11.2–15q13.3(acrocentric’p-arm’)

13933266 29838578 15905312 Loss

15q14–15q22.1 35983162 58075527 22092365 Loss

15q22.2 58075527 60658969 2583442 Gain

15q22.31–15q24.2 62783123 73838909 11055786 Loss

15q25.1–15q26.3b 78583731 100282878 21699147 Loss

16 monosomye Loss 16p13.3–p13.2 gain;16p13.12–p13.11 gain

17p13.3–17p12b 29169 12832730 12803561 Loss TP53 Gain 17q11 loss

17p12–17q12 15629313 29622988 13993675 Loss

17q12–17q25.3b 29622988 78528704 48905716 Gain TOP2A, ETV4, HOXB7,BIRC5, SOX9

17q21–q25 gain

18q monosomyf 17825691 76111023 58285332 Loss

19p13.3–19p13.12 2363823 14464847 12101024 Loss 19p13.3–p13.2 gain

J Cancer Res Clin Oncol (2010) 136:1869–1880 1875

123

and SOX9 genes (Fig. 6) (Lothe et al. 1995, 1996,

2001;Storlazzi et al. 2006; Levy et al. 2004, 2007; Kresse

et al. 2008; Mantripragada et al. 2008, 2009; Miller et al.

2009). Deletions at the 15q25.2–q26.3 and 6p25.1 regions

were also identified. Interestingly, the 6p25.1-located

NRN1 gene is a hypoxia-induced gene that has been pro-

posed as being a potential target for therapeutic interven-

tion in cancer (Le Jan et al. 2006). A number of small

Table 1 continued

ChromosomeRegion

Approximatestartpoint (bpposition)a

Approximateendpoint (bpposition)a

Overall sizeof genomicchange (bp)

Gain/loss

Previously reportedgene alterations

Previously reportedgenomic regionalterations

20p12.3 6087725 7787480 1699755 Loss

20p12.2–12p11.1 9952852 28081029 18128177 Loss 20p12.2 loss

22q12.2–22q12.3 29135614 33741775 4606161 Loss

22q12.3 34459370 34788931 329561 Gain

Xq21.33 94342534 95255491 912957 Loss

a Human Genome Build 36b includes most telomeric regionc may be contiguous with region in next rowd may be contiguous with region in previous rowe whole chromosome appears deletedf whole q-arm may be deleted

Fig. 6 The array CGH results from analyzing DNA from the MPNST tumor (region 3 FFPE section) are illustrated. These array CGH images

represent the results for chromosome 5 (a) and chromosome 17 (b)

1876 J Cancer Res Clin Oncol (2010) 136:1869–1880

123

genomic imbalances that appear to involve only single

genes were also found, for example, the TRPS1 gene, at

8q24.12, has also been implicated in cancer (Chang et al.

2004).

Discussion

NF1 is a typical tumor suppressor gene syndrome in which

most NF1 germline mutations result in complete inactiva-

tion of neurofibromin, the NF1 gene product, with sub-

sequent somatic NF1 mutations inactivating the normal

NF1 allele and leading to tumorigenesis. Evidence for both

LOH of the NF1 gene region and somatic NF1 point

mutations has been found in the many different NF1-

associated tumors, such as cutaneous and plexiform

neurofibromas, pheochromocytomas, MPNSTs, myeloid

leukemias, GISTs, glomus tumors, gastric carcinoid, and

pseudarthrosis (Serra et al. 2000; Upadhyaya et al. 2004,

2008a, b; Bausch et al. 2007; Side et al. 1997; Maertens

et al. 2006, Stewart et al. 2007; Brems et al. 2009a, b;

Stevenson et al. 2006). The main functional role of neu-

rofibromin is to catalyze the inactivation of p21-Ras-GTP,

the biologically active form of the Ras protein; hence, the

loss of functional neurofibromin in most NF1-associated

tumors results in increased activated p21-Ras-GTP levels

and an associated increase in cell growth and proliferation

(Guha et al. 1996; Schubbert et al. 2007). The finding that a

similar loss of neurofibromin and p21-Ras hyperactivation

occurs in Schwann cells from both benign PNFs and

malignant MPNSTs clearly indicate that additional genetic

alterations are required to permit the transformation of cells

from a benign PNF into the highly malignant cells in an

MPNST (Carroll and Ratner 2008).

A number of genomic changes, including DNA ampli-

fication and the associated up-regulation of TOP2A and

EGFR, as well as genomic deletions and resultant inacti-

vation of theCDKN2A and TP53 genes, have been impli-

cated in MPNST development. The present study involved

a detailed histopathological and molecular examination of

an NF1 patient who presented with multiple cutaneous

neurofibromas, a pilocytic astrocytoma, and a large path-

ologically complex posterior tibial nerve tumor. Increasing

LOH of the NF1 gene region was associated with tumor

regions that were consistent with progression from a benign

PNF (region 1), through an atypical neurofibroma-like

stage (region 2), to the final MPNST (region 3).

A concurrent analysis of DNA isolated from MPNSTs

removed from three unrelated NF1 patients found evidence

for LOH involving the NF1-gene region in all three tumors

and identified the specific 1.4-Mb somatic NF1 deletion in

two tumor DNAs that was not present in DNA from either

the patients’ lymphocytes or benign dermal neurofibromas

(unpublished data). Thus, a somatic 1.4-Mb NF1 deletion

was present in 75% of the MPNSTs analyzed in this study,

confirming our previous data that such large genomic

deletions represent the major somatic NF1 mutation type in

MPNSTs (Upadhyaya et al. 2008a). No mitotic recombi-

nation was evident in these three MPNSTs, although this

was identified as the mutational mechanism in only a single

MPNST in the earlier study. These results confirm our

previous findings that NF1 patients affected by a consti-

tutional NF1 microdeletion never subsequently develop

similar large genomic deletions in their tumors (De Raedt

et al. 2006; Upadhyaya et al. 2008b).

In addition to genetic changes involving the NF1 locus,

MPNST development is also associated with molecular

alterations to a number of other genes, including the cell-

cycle-associated regulatory tumor suppressor genes, TP53,

CDKN2A, and RB1. Inactivation of both TP53 and CDKN2A

are involved in tumor progression, with CDKN2A encoding

for both p16ink4a and p14ARF that are known regulators of

the retinoblastoma and the p53 signal transduction pathways

(Brems et al. 2009a). Evidence from several NF1 mouse

models has also identified a role for Tp53 inactivation, in

conjunction with NF1 inactivation, in MPNST development

(Cichowski and Jacks 2001). A recent study of idiopathic

gliomas has shown that the combined inactivation of TP53,

NF1, and PTEN, both in embryonic and adult neural stem/

progenitor cells, is associated with the in vivo development

of malignant astrocytomas (Llaguno et al. 2009). Previous

cytogenetic studies of MPNST have often identified com-

plex karyotypes, demonstrating triploidy and tetraploidy,

that were also found in this MPNST, underlining the likely

similarity of cytogenetic and molecular aberrations in these

tumor types (Bridge et al. 2004; Fang et al. 2009).

The aCGH study of DNA from region 3 of the MPNST,

using the Agilent 244 K microarray platform, identified a

number of tumor-specific genomic gains and losses

(Table 1). Some of these genomic alterations confirm our

previous aCGH data which demonstrates a marked differ-

ence in the genomic profiles associated with MPNSTs

compared to those from benign PNFs (Mantripragada et al.

2008; 2009). These include genomic deletions involving

the HMMR, MMP13, MTOR, NF1, OSF2, CDKN2A,

PTCH2, RB1, and TP53 genes, and genomic insertions and

duplications involving cMET, HGF, PDGFRA, CCNE2,

SOX10, and TOP2A genes. A recent related study with the

32-K genome-wide microarray identified a consistent

9p21.3 deletion, a region that encompasses the CDKN2A,

CDKN2B, and MTAP genes (Mantripragada et al. 2009).

Amplified genomic regions involved both the

5p15.3region, containing the NKD2, IRX4, IRX1, IRX2,

TRIO, and HGF genes, and the 17q21–25 region that

includes the TOP2A, ETV4, HOXB7, BIRC, and SOX9

genes. Several of these amplified regions, and their

J Cancer Res Clin Oncol (2010) 136:1869–1880 1877

123

associated genes, have been implicated in MPNST devel-

opment (Storlazzi et al. 2006; Levy et al. 2007; Kresse

et al. 2008; Mantripragada et al. 2008, 2009; Miller et al.

2009). Similarly, several of the deleted genomic regions

and the related genes, including AP3B1 (5q14.1), PDGFR

B (5q31), CDKN2A, and MTAP (9p21), MMP12 (11q22.3),

RB1 (13q14.2), and TP53 (17p13.1), have also been

reported (Lothe et al. 1996, Mantripragada et al. 2009).

Several novel genomic changes were also identified in

DNA from the high-grade MPNST area of the tumor in the

present aCGH study and included deletions at 15q25.2–

q26.3 and at 6p25.1. The hypoxia-induced NRN1 gene,

located within the identified 6p25.1 deletion, has been

suggested to have prognostic significance and be a possible

target for therapeutic intervention in cancer (Le Jan et al.

2006). A number of genomic imbalances that apparently

involve only single genes were also found, these included

the TRPS1 gene (8q24.12) that has been implicated in

lymphoma development and in other cancers (Chang et al.

2004;Savinainen et al. 2004).

Most cancers are associated with the clonal acquisition

of an increasing number of genetic alterations, in which the

subsequent expansion of the growth-advantaged clones

explains the likely mechanism underlying the transforma-

tion from a benign PNF into malignant MPNST tumor.

Also, the likely presence of cancer stem cells in such solid

tumors needs consideration (Dalerba et al. 2007). These

cancer progenitor cells could possibly be involved in the

genesis of the initial benign PNF, in the cells of which

subsequent genetic alterations result in progression toward

MPNST development. These and other possible explana-

tions require further directed research and will most likely

involve the use of appropriate mouse NF1 model systems.

In summary, variable LOH was present, involving mul-

tiple genes in histopathologically different regions of this

complex tumor. This indicates that tumor development was

associated with a number of different genetic changes.

Variations in the type and degree of molecular alterations

correlated with different degrees of malignant transforma-

tion within the same posterior tibial nerve tumor. In con-

clusion, a correlation of the histological and molecular

changes associated with MPNST development has been

demonstrated, confirming the significant cellular and genetic

heterogeneity associated with these tumors. An expanded

study of a much larger tumor samples is clearly warranted.

The apparently MPNST tumor-type-specific molecular

alterations require confirmation and further characterization

to ascertain whether they may provide additional novel

diagnostic and therapeutic targets for the earlier identifica-

tion and treatment of these aggressive malignant tumors.

Acknowledgments We are grateful to Cancer Research, UK for 498

financial help and we thank K Mantripragada for technical assistance.

We are grateful to Dr Elham Sadighi-Akha for performing the aCGH

hybridizations. SJLK is supported by the NIHR Biomedical Research

Centre, Oxford. This work was also supported by the Wellcome Trust

[075491/Z/04].

References

Altmannsberger M, Weber K, Droste R, Osborn M (1985) Desmin is a

specific marker for rhabdomyosarcomas of human and rat origin.

Am J Pathol 118:85–95

Angelov L, Davis A, O’Sullivan B, Bell R, Guha A (1998)

Management of neurogenic sarcomas: a University of Toronto

experience. Neurosurgery 43:56–65

Bausch B, Borozdin W, Mautner VF, Hoffmann MM, Boehm D,

Robledo M, Cascon A, Harenberg T, Schiavi F, Pawlu C,

Peczkowska M, Letizia C, Calvieri S, Arnaldi G, Klingenberg-

Noftz RD, Reisch N, Fassina A, Brunaud L, Walter MA,

Mannelli M, MacGregor G, Palazzo FF, Barontini M, Walz MK,

Kremens B, Brabant G, Pfaffle R, Koschker AC, Lohoefner F,

Mohaupt M, Gimm O, Jarzab B, McWhinney SR, Opocher G,

Januszewicz A, Kohlhase J, Eng C, Neumann HP, European-

American Phaeochromocytoma Registry Study Group (2007)

Germline NF1 mutational spectra and loss-of-heterozygosity

analyses in patients with pheochromocytoma and neurofibroma-

tosis type 1. J Clin Endocrinol Metab 92:2784–2792

Belchis DA, Meece CA, Benko FA, Rogan PK, Williams RA, Gocke

CD (1996) Loss of heterozygosity and microsatellite instability

at the retinoblastoma locus in osteosarcomas. Diagn Mol Pathol

5:214–219

Bennett E, Thomas NS, Upadhyaya M (2009) Neurofibromatosis

type 1: its association with the Ras/MAPK pathway syndromes.

J Paeditr Neurol 7:1–5

Brems H, Beert E, de Ravel T, Legius E (2009a) Mechanisms in the

pathogenesis of malignant tumors in neurofibromatosis type 1.

Lancet Oncol 10:508–515

Brems H, Park C, Maertens O, Pemov A, Messiaen L, Upadhyaya M,

Claes K, Beert E, Peeters K, Mautner V, Sloan J, Yao L, Lee CR,

Sciot R, De Smet L, Legius E, Stewart D (2009b) Glomus

tumors in neurofibromatosis type 1: genetic, functional, and

clinical evidence of a novel association. Can Res 69:7393–7401

Bridge RS Jr, Bridge JA, Neff JR, Naumann S, Althof P, Bruch LA

(2004) Recurrent chromosomal imbalances and structurally abnor-

mal breakpoints within complex karyotypes of malignant periph-

eral nerve sheath tumor and malignant triton tumor: a cytogenetic

and molecular cytogenetic study. J Clin Pathol 57:1172–1178

Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo A, Mao L,

Herath J, Jenkins R, Westra W (1995) Frequency of homozygous

deletion at p16/CDKN2 in primary human tumors. Nat Genet

11:210–212

Carroll SL, Ratner N (2008) How does the Schwann cell lineage form

tumors in NF1? Glia 56:1590–1605

Chang GT, Jhamai M, van Weerden WM, Jenster G, Brinkmann AO

(2004) The TRPS1 transcription factor: androgenic regulation in

prostate cancer and high expression in breast cancer. Endocr

Relat Cancer 11:815–822

Cichowski K, Jacks T (2001) NF1 tumor suppressor gene function:

narrowing the GAP. Cell 104:593–604

Coulie PG, Brichard V, Van Pel A, Wolfel T, Schneider J, Traversari

C, Mattei S, De Plaen E, Lurquin C, Szikora JP, Renauld JC,

Boon T (1994) A new gene coding for a differentiation antigen

recognized by autologous cytolytic T lymphocytes on HLA-A2

melanomas. J Exp Med 180:35–42

Dalerba P, Cho RW, Clarke MF (2007) Cancer stem cells: models and

concepts. Ann Rev Med 58:267–284

1878 J Cancer Res Clin Oncol (2010) 136:1869–1880

123

De Raedt T, Maertens O, Chmara M, Brems H, Heyns I, Sciot R,

Majounie E, Upadhyaya M, De Schepper S, Speleman F,

Messiaen L, Vermeesch JR, Legius E (2006) Somatic loss of

wild type NF1 allele in neurofibromas: comparison of NF1

microdeletion and non-microdeletion patients. Genes Chromo-

somes Cancer 45:893–904

Evans DG, Baser ME, McGaughran J, Sharif S, Howard E, Moran A

(2002) Malignant peripheral nerve sheath tumors in neurofibro-

matosis 1. J Med Genet 39:311–314

Fang Y, Elahi A, Denley RC, Rao PH, Brennan MF, Jhanwar SC

(2009) Molecular characterization of permanent cell lines from

primary, metastatic and recurrent malignant peripheral nerve

sheath tumors (MPNST) with underlying neurofibromatosis-1.

Anticancer Res; 29:1255–1262

Futreal PA, Barrett JC, Wiseman RW (1991) An Alu polymorphism

intragenic to the TP53 gene. Nucleic Acids Res 19:6977

Gowan AM, Vogel AM, Hoak D, Gough F, McNutt MA (1986)

Monoclonal antibodies specific for melanocytic tumors distin-

guish subpopulations of melanocytes. Am J Pathol 123:195–203

Guha A, Lau N, Huvar I, Gutmann D, Provias J, Pawson T, Boss G

(1996) Ras-GTP levels are elevated in human NF1 peripheral

nerve tumors. Oncogene 12:507–513

Hartmann A, Rosner U, Schlake G, Dietmaier W, Zaak D, Hofstaed-

ter F, Knuechel R (2000) Clonality and genetic divergence in

multifocal low-grade superficial urothelial carcinoma as deter-

mined by chromosome 9 and p53 deletion analysis. Lab Invest

80:709–718

Kawakami Y, Eliyahu S, Delgado CH, Robbins PF, Rivoltini L,

Topalian SL, Miki T, Rosenberg SA (1994) Cloning of the gene

coding for a shared human melanoma antigen recognized by

autologous T cells infiltrating into tumor. Proc Natl Acad Sci

USA 91:3515–3519

Kresse SH, Skarn M, Ohnstad HO, Namløs HM, Bjerkehagen B,

Myklebost O, Meza-Zepeda LA (2008) DNA copy number

changes in high-grade malignant peripheral nerve sheath tumors

by array CGH. Mol Cancer 7:48

Kumar S, Perlman E, Harris CA, Raffeld M, Tsokos M (2000)

Myogenin is a specific marker for rhabdomyosarcoma: an

immunohistochemical study in paraffin-embedded tissues. Mod

Pathol 13:988–993

Le Jan S, Le Meur N, Cazes A, Philippe J, Le Cunff M, Leger J,

Corvol P, Germain S (2006) Characterization of the expression

of the hypoxia-induced genes Neuritin, TXNIP and IGFBP3 in

cancer. FEBS Lett 580:3395–3400

Levy P, Vidaud D, Leroy K, Laurendeau I, Wechsler J, Bolasco G,

Parfait B, Wolkenstein P, Vidaud M, Bieche I (2004) Molecular

profiling of malignant peripheral nerve sheath tumors associated

with neurofibromatosis type 1, based on large-scale real-time

RT-PCR. Mol Cancer 3:20

Levy P, Ripoche H, Laurendeau I, Lazar V, Ortonne N, Parfait B,

Leroy K, Wechsler J, Salmon I, Wolkenstein P, Dessen P,

Vidaud M, Vidaud D, Bieche I (2007) Microarray-based

identification of tenascin C and tenascin XB, genes possibly

involved in tumorigenesis associated with neurofibromatosis

type 1. Clin Cancer Res 13:398–407

Llaguno AS, Chen J, Kwon CH, Jackson EL, Li Y, Burns DK,

Alvarez-Buylla A, Parada LF (2009) Malignant astrocytomas

originate from neural stem/progenitor cells in a somatic tumor

suppressor mouse model. Cancer Cell 15:45–56 (Erratum:

Cancer Cell 15:240)

Lothe RA, Slettan A, Saeter G, Brøgger A, Børresen AL, Nesland JM

(1995) Alterations at chromosome 17 loci in peripheral nerve

sheath tumors. J Neuropathol Exp Neurol 54:65–73

Lothe RA, Karhu R, Mandahl N, Mertens F, Saeter G, Heim S,

Borresen-Dale AL, Kallioniemi OP (1996) Gain of 17q24-qter

detected by comparative genomic hybridization in malignant

tumors from patients with von Recklinghausen’s neurofibroma-

tosis. Cancer Res 56:4778–4781

Lothe RA, Smith-Sørensen B, Hektoen M, Stenwig AE, Mandahl N,

Saeter G, Mertens F (2001) Biallelic inactivation of TP53 rarely

contributes to the development of malignant peripheral nerve

sheath tumors. Genes Chromosomes Cancer 30:202–206

Maertens O, Prenen H, Debiec-Rychter M, Wozniak A, Sciot R,

Pauwels P, De Wever I, Vermeesch JR, de Raedt T, De Paepe A,

Speleman F, van Oosterom A, Messiaen L, Legius E (2006)

Molecular pathogenesis of multiple gastrointestinal stromal

tumors in NF1 patients. Hum Mol Genet 15:1015–1023

Mantripragada KK, Spurlock G, Kluwe L, Chuzhanova N, Ferner RE,

Frayling IM, Dumanski JP, Guha A, Mautner V, Upadhyaya M

(2008) High-resolution DNA copy number profiling of malig-

nant peripheral nerve sheath tumors using targeted microarray-

based comparative genomic hybridization. Clin Cancer Res

14:1015–1024

Mantripragada KK, de Stahl TD, Patridge C, Menzel U, Andersson R,

Chuzhanova N, Kluwe L, Guha A, Mautner V, Dumanski JP,

Upadhyaya M (2009) Genome-wide high-resolution analysis of

DNA copy number alterations in NF1-associated malignant

peripheral nerve sheath tumors using 32 K BAC array. Genes

Chromosomes Cancer 48:897–907

Miller S, Walter JJ, Mehta T, Hardiman A, Sites E, Kaiser S, Jegga A,

Li H, Upadhyaya M, Giovannini M, Muir D, Wallace M, Lopez

E, Serra E, Lazaro C, Stemmer-Rachamimov PageG, Aronow

BJ, Ratner N (2009) Integrative genomic analyses show altered

Schwann cell development in neurofibromatosis tumors and

implicate SOX9 as an addicting oncogene. EMBO Mol Med

1:236–248

Pollock PM, Stark MS, Palmer JM, Walters MK, Aitken JF, Martin

NG, Hayward NK (2001) Mutation analysis of the CDKN2A

promoter in Australian melanoma families. Genes Chromosomes

Cancer 32:89–94

Savinainen KJ, Linja MJ, Saramaki OR, Tammela TL, Chang GT,

Brinkmann AO, Visakorpi T (2004) Expression and copy

number analysis of TRPS1, EIF3S3 and MYC genes in breast

and prostate cancer. Br J Cancer 90:1041–1046

Schubbert S, Shannon K, Bollag G (2007) Hyperactive Ras in

developmental disorders and cancer. Nat Rev Cancer 7:295–308

Serra E, Rosenbaum T, Winner U, Aledo R, Ars E, Estivill X, Lenard

HG, Lazaro C (2000) Schwann cells harbor the somatic NF1

mutation in neurofibromas: evidence of two different Schwann

cell subpopulations. Hum Mol Genet 9:3055–3064

Side L, Taylor B, Cayouette M, Conner E, Thompson P, Luce M,

Shannon K (1997) Homozygous inactivation of the NF1 gene

in bone marrow cells from children with neurofibromatosis type 1

and malignant myeloid disorders. N Engl J Med 336:1713–1720

Stasik CJ, Tawfik O (2006) Malignant peripheral nerve sheath tumor

with rhabdomyosarcomatous differentiation (malignant triton

tumor). Arch Pathol Lab Med 130:1878–1881

Stevenson DA, Zhou H, Ashrafi S, Messiaen LM, Carey JC, D’Astous

JL, Santora SD, Viskochil DH (2006) Double inactivation of

NF1 in tibial pseudarthrosis. Am J Hum Genet 79:143–148

Stewart W, Traynor JP, Cooke A, Griffiths S, Onen NF, Balsitis M,

Shah AA, Upadhyaya M, Tobias ES (2007) Gastric carcinoid:

germline and somatic mutation of the neurofibromatosis type 1

gene. Fam Cancer 6:147–152

Stewart H, Bowker C, Edees S, Smalley S, Crocker M, Mechan D,

Forrester N, Spurlock G, Upadhyaya M (2008) Congenital

disseminated neurofibromatosis type 1: a clinical and molecular

case report. Am J Med Genet 146A:1444–1452

Storlazzi CT, Brekke HR, Mandahl N, Brosjo O, Smeland S, Lothe

RA, Mertens F (2006) Identification of a novel amplicon at distal

17q containing the BIRC5/SURVIVIN gene in malignant

peripheral nerve sheath tumors. J Pathol 209:492–500

J Cancer Res Clin Oncol (2010) 136:1869–1880 1879

123

Tucker T, Wolkenstein P, Revuz J, Zeller J, Friedman JM (2005)

Association between benign and malignant peripheral nerve

sheath tumors in NF1. Neurology 65:205–211

Upadhyaya M and Cooper DN. 1998. NeurofibromatosisType 1: From

Genotype to Phenotype Upadhyaya M, Cooper DN (Eds).

Oxford: BIOS Scientific Publishers

Upadhyaya M, Han S, Consoli C, Majounie E, Horan M, Thomas NS,

Potts C, Griffiths S, Ruggieri M, von Deimling A, Cooper DN

(2004) Characterization of the somatic mutational spectrum of

the neurofibromatosis type 1 (NF1) gene in neurofibromatosis

patients with benign and malignant tumors. Hum Mutat 23:134–

146

Upadhyaya M, Kluwe L, Spurlock G, Monem B, Majounie E,

Mantripragada K, Ruggieri M, Chuzhanova N, Evans DG,

Ferner R, Thomas N, Guha A, Mautner V (2008a) Germline and

somatic NF1 gene mutation spectrum in NF1-associated malig-

nant peripheral nerve sheath tumors (MPNSTs). Hum Mutat

29:74–82

Upadhyaya M, Spurlock G, Monem B, Thomas N, Friedrich RE,

Kluwe L, Mautner V (2008b) Germline and somatic NF1 gene

mutations in plexiform neurofibromas. Hum Mutat 29:E103–

E111

1880 J Cancer Res Clin Oncol (2010) 136:1869–1880

123