JOBNAME: jnen 64#1 2005 PAGE: 1 OUTPUT: Tue January 18 18:20:30 2005
lww/jnen/90813/NEN18628
J Neuropathol Exp Neurol
Copyright Ó 2005 by the American Association of Neuropathologists, Inc.
Vol. 64, No. 1
January 2005
pp. 74–81
ORIGINAL ARTICLE
Expression Patterns of Cell Cycle Components in Sporadic and
Neurofibromatosis Type 1-Related Malignant Peripheral Nerve
Sheath Tumors
Trude Holmeide Ågesen, MSc, Vivi Ann Flørenes, MSc, PhD, Willemina M. Molenaar, MD, PhD,
Guro E. Lind, MSc, Jeanne-Marie Berner, MSc, PhD, Boudewijn E. C. Plaat, MD, PhD,
Rudy Komdeur, MD, PhD, Ola Myklebost, MSc, PhD, Eva van den Berg, MSc, PhD,
and Ragnhild A. Lothe, MSc, PhD
Abstract
The molecular biology underlying the development of highly
malignant peripheral nerve sheath tumors (MPNSTs) remains
mostly unknown. In the present study, the expression pattern of
10 selected cell cycle components is investigated in a series of
15 MPNSTs from patients with (n = 9) or without (n = 5)
neurofibromatosis type 1 (NF1). Thirteen tumors did not express
the cyclin-dependent kinase inhibitor, p16INK4A, an observation
that was related to homozygote gene deletions in three tumors,
heterozygote deletions in five, and gross gene rearrangements in
five. The absence of protein expression in the tumors with one
seemingly intact allele was not caused by promoter hypermethylation of p16INK4A or p14ARF. All tumor samples expressed
normal sized RB1, cyclin D3, CDK2, CDK4, p21CIP1, and p27KIP1
proteins, and only a single tumor showed an aberrant protein band
for one of these proteins, p21CIP1. Cyclin D1 was absent in four
tumors; all except one tumor showed expression of TP53 protein,
and three of nine MPNSTs had expression of normal-sized
MDM2. In conclusion, this study shows that the vast majority of
MPNSTs had gross rearrangements of the p16INK4A gene,
explaining the absence of the encoded protein in the same tumors.
The level of expression was equally distributed between the
familial (NF1) and sporadic cases, although it should be noted that
the 2 cases with p16INK4A expression were sporadic. The data
imply that the complete absence of p16INK4A is sufficient for
activation of the cell cycle in most MPNSTs; thus, it is not
From the Departments of Genetics (THA, GEL, RAL) and Tumor Biology
(J-MB, OM), Institute for Cancer Research, and Department of Pathology
(VAF), Norwegian Radium Hospital, Montebello, Oslo, Norway; Departments of Pathology (WMM, BECP), Surgical Oncology (RK), and
Clinical Genetics (EvdB), University Hospital of Groningen, the Netherlands; and Department of Molecular Biosciences (RAL), University of
Oslo, Oslo, Norway.
Send correspondence and reprint requests to: Professor Ragnhild A. Lothe, PhD,
Department of Genetics, Institute for Cancer Research, Norwegian Radium
Hospital, Montebello, 0310 Oslo, Norway. E-mail: [email protected]
Supported by grants from the Norwegian Cancer Society (RAL: A95068, OM,
VAF).
74
necessary for tumor proliferation to further stimulate the cycle
through alteration of other central components.
Key Words: Cell cycle, Malignant peripheral nerve sheath tumors,
Neurofibromatosis, NF1, p16INK4A, Western blot.
INTRODUCTION
Malignant peripheral nerve sheath tumors (MPNSTs)
are highly aggressive and arise sporadically or in patients with
the common autosomal dominant hereditary disorder, neurofibromatosis 1 (NF1) (1). The NF1 tumor suppressor gene
maps to chromosome band 17q11.2 and encodes the neurofibromin protein that functions as a suppressor of RASmediated signaling. Individuals with NF1 carry a germline
mutation in this gene. Several studies have shown loss of
heterozygosity of chromosome arm 17q sequences, including
the NF1 locus in MPNSTs (2–5). Thus, a complete inactivation of NF1 is assumed to contribute to development
of MPNSTs. However, somatic mutations in NF1 are also
reported in benign neurofibromas, indicating that additional
genetic events besides inactivation of NF1 are necessary for
malignant transformation (6, 7).
In contrast to several other soft tissue sarcomas, no
pathognomonic structural aberration has been found in MPNST
that often display complex karyotypes (8, 9). With comparative
genomic hybridization (CGH), recurrent chromosomal copy
number changes have been identified in MPNSTs, including losses from 9p and 13q (10–12). Based on the CGH data
and other studies reporting loss of heterozygosity of 17p
sequences, the tumor suppressor genes p16INK4A (map position
9p21), RB1 (13q14), and TP53 (17p13) have been suggested
to be target genes for these deletions.
Studies on the p16INK4A gene have revealed gene alterations in 50% to 75% of the MPNSTs analyzed (13–15).
Alterations in p16INK4A are not reported in neurofibromas, and
this emphasizes the importance of disrupted p16INK4A in the
progression of the disease. The cyclin-dependent kinase
inhibitor locus CDKN2A encodes p16INK4A (a-transcript) and
p14ARF (b-transcript), which are both important in the negative
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
JOBNAME: jnen 64#1 2005 PAGE: 2 OUTPUT: Tue January 18 18:20:31 2005
lww/jnen/90813/NEN18628
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
regulation of the cell cycle through their interaction with the
RB1 and TP53 pathways, respectively (16).
TP53 is frequently mutated in a number of cancer types,
whereas biallelic inactivation seems to be a rare event in
MPNSTs (17), even though a few point mutations in TP53
have been reported (18–20). Nevertheless, mouse models
show that the TP53 pathway is important for the development
of MPNSTs (21, 22).
In the present study, we used Western blot analysis to
further examine the protein levels of p16INK4, RB1, and TP53,
as well as seven additional cell cycle components in a series of
MPNSTs. Subsequently, molecular genetic and epigenetic analyses were performed on candidate genes, revealing interesting
results based on the protein expression study.
MATERIALS AND METHODS
Patients and Tumors
Samples from 18 MPNSTs were obtained from 16
patients admitted to the University Hospital of Groningen,
The Netherlands. All frozen tumor samples were sectioned,
hematoxylin and eosin-stained, and examined by an expert
pathologist. All diagnoses were confirmed by routine histology.
Three samples were excluded from the present study due to large
amounts of necrotic tissue, and hence poor protein and DNA
quality. The remaining samples all contained 80% to 100%
tumor cells. Clinical data from the 14 patients (15 MPNSTs)
included in the present study are summarized in Table 1. Ten
tumors were taken from 9 patients with NF1 diagnosis, 3
females and 6 males, whereas the other 5 tumors were collected from 1 female and 4 males, without a family history of
NF1. From 1 patient, 3 tumor samples (M1, M2, and M3) were
Cell Cycle Proteins in MPNST
analyzed, including 2 from different sites within the primary
tumor (PT) and 1 from a metastasis.
Western Blot Analyses of 10 Cell
Cycle Proteins
Frozen tissue samples were crushed into fine powder
using a pestle and a mortar filled with nitrogen (N2) prior to
lysis in cold lysis buffer (20 mmol/L Tris-HCl at pH 7.5,
137 mmol/L NaCl, 100 mmol/L NaF, 1 mmol/L Na2VO2, 10%
glycerol, 1% icopal (NP-40), 1 mmol/L phenylmethylsulfonyl
fluoride (PMSF), and 0.01 mg/mL each of aprotinin,
leupeptin, and pepstatin (all chemicals from Sigma, St. Louis,
MO). Lysates were sonicated and clarified by centrifugation.
Protein quantization was done by Bradford analysis and 50 to
80 mg protein/lane was resolved by SDS polyacrylamide gel
electrophoresis (SDS page). Following gel electrophoresis, proteins were blotted onto Immobilin-P membranes (Millipore,
Bedford, MA). Hybridization was performed as described by
St. Croix et al (23). Primary and secondary antibodies used in
the present study are listed in Table 2. As positive controls,
different cell lines were prepared as the tumor samples.
The human melanoma cell line, WM35, was used as a positive
control for detection of RB1, cyclin D1, cyclin D3, CDK2,
CDK4, p21CIP1, p27KIP1, and TP53 (24, 25). MFH1X, a malignant fibrous histiocytoma, and NCS2x, a nonclassified sarcoma, both established as xenograft cell lines, were used for its
expression of mutated TP53 and MDM2 proteins, respectively (26). A p16INK4A control was obtained commercially
(Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The signals of the protein bands were visually scored and graded
as follows: 2 = no expression; + = weak expression; ++ =
moderate expression; and +++ = strong expression. a-Tubulin
was used as a loading control. The intensity of the protein
TABLE 1. Clinicopathologic Data
Tumor
Sample
Gender
Age (year)
at Diagnosis
NF1
Status
Diagnosis
Tumor Type
Tumor Site
Tumor
Grade
B1
C*
D
E
F
G
H
J
K
M1‡
M2‡
M3‡
N
O
P
Q
M
M
M
F
M
F
M
M
F
M
M
M
M
M
M
F
28
27
74
25
22
52
41
61
32
26
26
26
41
53
29
42
Yes
Yes
No
Yes
Yes
Yes
No
No
No
Yes
Yes
Yes
No
Yes
Yes
Yes
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
MPNST
New PT?
Meta
PT
PT†
PT after TNF, presumably R
PT
R
PT after TNF
PT
PT biopsy
PT resection
R
PT
PT
PT
Synchronous meta or multiple PT
Proximal jejunal
Lung
Mediastinum
Neck
Thigh
Mediastinum
Axilla
Knee
Tigh
Thoracic wall
Thoracic wall
Thoracic wall left
Tigh
Shoulder
Sacral area
Neck
NK
NK
III
III
III
III
I
III
II
III
III
II
III
III
III
III
PT, primary tumor; Meta, metastasis; R, recurrence; NK, not known.
*, Published by Molenaar et al (64).
†, Triton tumor.
‡, M1, M2, M3; 3 tumor samples from 1 patient. M1 and M2 are from the same tumor (biopsy and resection).
q 2005 American Association of Neuropathologists, Inc.
75
JOBNAME: jnen 64#1 2005 PAGE: 3 OUTPUT: Tue January 18 18:20:32 2005
lww/jnen/90813/NEN18628
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
Ågesen et al
TABLE 2. Antibodies Used in the Western Blot Experiments
Antibody
Antibody Specification
Manufacturer
Concentration Used
in Western Blot
a-tubulin
RBI
p16INK4A
cyclin D1
cyclin D1
cyclin D3
CDK2
CDK4
p21CIP1
p27KIP1
TP53
MDM2
W401B
W402B
Monoclonal Mouse Anti-Human a-tubulin (clone 57)
Monoclonal Mouse Anti-Human Retinoblastoma (RBI)
Polyclonal Rabbit Anti-Human p16 (C-20)
Monoclonal Mouse Anti-Human Cyclin D1 (clone DCS-6)
Monoclonal Mouse Anti-Human Cyclin D1 (clone DCS-6)
Monoclonal Mouse Anti-Human Cyclin D3 (clone DCS-22)
Polyclonal Rabbit Anti-Human CDK2 (M2)
Polyclonal Rabbit Anti-Human CDK4 (C-22)
Polyclonal Rabbit Anti-Human p21 (C-19)
Monoclonal Mouse Anti-Human p27 (clone 57)
Monoclonal Mouse Anti-Human p53 (DQ-1)
Monoclonal Mouse Anti-Human MDM2 (AB1)
Anti-Rabbit IgG HRP
Anti-Mouse IgG HRP
Oncogene Research Products, Boston, MA
Pharmingen, San Diego, CA
Santa Cruz Biotechnology, Inc., Santa Cruz, CA
Oncogene Research Products
Neomarkers, Fremont, CA
DAKO A/S, Glostrup, Denmark
Santa Cruz Biotechnology, Inc.
Santa Cruz Biotechnology, Inc.
Santa Cruz Biotechnology, Inc.
BD Biosciences, San Jose, CA
Santa Cruz Biotechnology, Inc.
Oncogene Research Products
Promega, Madison, WI
Promega, Madison, WI
0.5–1.0 mg/mL
2–2.5 mg/mL
2–3 mg/mL
2 mg/mL
1–2 mg/mL
3.4 mg/mL
1 mg/mL
1 mg/mL
1–2 mg/mL
1 mg/mL
5 mg/mL
0.6 mg/mL
1:5,000
1:5,000
bands in each sample was visually compared with the expression of a-tubulin. The relative intensity among all samples
was finally decided based on joint evaluation of the different
runs and different exposures. The visual scorings were done by
two of the authors (THÅ and VAF).
Gene Rearrangements of CDKN2A by
Southern Blot Analysis
One previously analyzed MPNST (1: 650) and 2 normal
blood samples were included as positive and negative controls
for p16INK4A pattern in the Southern blot experiments (13).
Tissue samples were digested with proteinase K, and genomic
DNA extracted semi-automatically by phenol/chloroform
followed by ethanol precipitation as described by the manufacturer (Nucleic Acid Extractor, Applied Biosystems, Foster
City, CA). DNA (3.5 mg) from each sample was digested to
completion by BamHI (Amersham Biosciences Europe
GmbH, Oslo, Norway), and then separated by electrophoresis
through a 0.8% agarose gel. The separated fragments were
transferred onto a Hybond-N+ membrane (Amersham Biosciences) and DNA was fixed to the membrane by baking at
80°C for 2 hours. A cDNA fragment containing the human
p16INK4A sequence was used as probe (provided by Dr. D.
Beach) (27), and radioactively labeled with 32P by ‘‘random
priming’’ (28). The hybridization was carried out in 0.5 mol/L
sodium phosphate (pH 7.2), 7% SDS, and 1 mmol/L sodium
EDTA overnight at 65°C. After hybridization, excess probe
was removed in several washing steps with 40 mmol/L NaP
(pH 7.2) and 1% SDS. To correct for an unequal loading of
DNA, the membranes were rehybridized with a control probe,
APOB, localized to chromosome 2. Quantization of signal
intensities was achieved by two-dimensional densitometry
using a Molecular Dynamics (Sunnyvale, CA) laser densitometer. The net signals (integrated optical density) of
CDKN2A bands were calculated relative to the signal obtained
with the APOB control probe, i.e. CDKN2A/APOB of the
sample divided by the average CDKN2A/APOB for the
samples scored as normal. Signals weaker than 25% or 75%
were scored as a homozygote or heterozygote deletion,
76
respectively. This conservative cutoff was chosen since some
contribution of signal is expected from the presence of normal
cells in the tumor and the fact that the nature of Southern
blotting does not provide exact quantitative results. Rehybridization of the Southern blot with the p16INK4A probe was performed and the signal intensities measured again, confirming
the initial results.
Epigenetic Analysis of CDKN2A Locus
DNA samples submitted to methylation analyses were
modified according to the protocol of the CpGenome DNA
modification kit (InterGen, Boston, MA) as previously described (29). Briefly, 1 mg DNA was used as starting amount
and each treated sample was resuspended in 50 mL 0.5 3 TE,
pH 7.5. DNA methylation status of CDKN2A (a- and b-transcript) promoters were determined by subsequent methylation
specific PCR (MSP) (30), using different primer sets specific
for methylated and unmethylated CpG sites. Previously described primer sets were used for amplification of p14ARF and
p16INK4A (30). All fragments were amplified with 25 pmol of
each primer. The reactions contained from 1.2 to 1.5 mmol/L
MgCl2, and 0.25 to 2 mL bisulphite modified DNA solution as
template. The enzyme used for the MSP reactions was 0.625
unit HotStart (Qiagen Inc., Valencia, CA). The annealing
temperatures were 64°C (unmethylated a-transcript), 63°C
(methylated a-transcript), and 59°C for both b-transcript. In
vitro SssI (New England BioLabs, Beverly, MA) methylated
Placenta DNA (Sigma) was used as the positive control for the
methylated reactions, whereas DNA from normal lymphocytes
was used as the positive control for the unmethylated reactions. Water replacing template in the MSP was the negative
control for both reactions. The PCR products were analyzed by
running 7.5% polyacrylamide gel and stained with ethidium
bromide before photographing using an UV transilluminator.
RESULTS
The Western blot results for the 10 selected cell cycle
proteins are summarized in Table 3.
q 2005 American Association of Neuropathologists, Inc.
JOBNAME: jnen 64#1 2005 PAGE: 4 OUTPUT: Tue January 18 18:20:34 2005
lww/jnen/90813/NEN18628
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
Cell Cycle Proteins in MPNST
TABLE 3. Protein Expression of Cell Cycle Components in MPNSTs
Expression Levels of Cell Cycle Regulators Observed by Western Blot Analyses
Tumor Sample
RBI
(13q14.2)
P16INK4A
(9p21)
cyclin D1
(11q13.3)
Cyclin D3
(6p21)
CDK2
(12q13)
CDK4
(12q14)
p21CIP1
(6p21.2)
p27KIP1
(12p13)
TP53
(17p13.1)
MDM2
(12q14.3-q15)
B1
C
D
E
F
G
H
J
K
M1
M2
M3
N
O
P
Q
Positively stained
MPNST number analyzed
++
++
+
++
+
++
++
++
++
+
++
+++
++
+++
++
+++
15/15
(100%)
2
2
+++
2
2
2
+
2
2
2
2
2
2
2
2
2
2/15
(13%)
+
+
2
+
2
++
++
+
+
+
+
2
+
+
2
+
11/15
(73%)
+++
+++
+
++
++
++
+++
++
++
++
++
++
+
++
++
++
15/15
(100%)
++
++
++
++
+
+
+
+
++
++
++
++
++
++
++
++
15/15
(100%)
++
++
++
++
++
++
++
++
++
++
++
+
++
++
++
++
15/15
(100%
++
++
+
+
+
++
+++*
++
++
++
+++
+
+
+
+++
+
15/15
(100%)
++
++
+++
++
++
++
++
++
+++
++
++
++
+
++
++
+
15/15
(100%)
++
++
++
++
+
+
++
++
++
++
++
++
+
2
++
++
14/15
(93%)
na
na
na
na
+++
na
na
2
2
2
+
2
2
2
+++
2
3/9
(30%)
Note: Proteins are indicated on the top of each column with the chromosome localization in parentheses.
*, In addition to the expected protein band of 21 kDa, a single band with a lower molecular weight was seen. Protein expression level is indicated as follows: 2, no expression;
+, weak; ++, moderate; +++, strong; na, not analyzed.
p16INK4A was only detected in 2 of 15 (13%) MPNSTs.
One tumor showed weak protein expression, whereas a strong
expression of p16INK4A was detected in the second (Fig. 1A).
Thirteen of 15 (87%) MPNSTs showed gene alterations in
p16INK4A by Southern blot (Fig. 1B). Of the two MPNSTs with
p16INK4A protein expression, only tumor D with a strong
expression showed a normal gene pattern in the Southern blot
experiment. In five samples, the signal intensity was evaluated
to be weaker than between 25% and 75% compared with normal samples, indicating a heterozygote deletion of p16INK4A.
Three MPNSTs had an intensity that corresponds to a homozygote deletion of the gene. Five tumors had gene rearrangements, seen as restriction fragments of abnormal mobility. In
two of these tumors, the normal band was detectable in
addition to the abnormal restriction fragment (Fig. 1B).
The CDKN2A locus encoding both p16INK4A and
ARF
p14
revealed no hypermethylation in either of the two
promoters. For p16INK4A, 13 of 15 tumors were successfully
analyzed, whereas the success rate for p14ARF was 11 of 15.
All tumor samples expressed RB1, cyclin D3, CDK2,
CDK4, p21CIP1, and p27KIP1 proteins. Examples related to the
level of a-tubulin are illustrated in Figure 2. The antibody
used to detect RB1 hybridizes to both phosphorylated and
unphosphorylated proteins, and seemingly most of the RB1
protein was hypophosphorylated. RB1, cyclin D3, CDK2,
CDK4, and the protein CDK inhibitor p27KIP1 typically
showed a moderate or strong expression level, whereas p21CIP1
ranged from a low to strong protein level. One tumor (H)
revealed a protein with reduced size in addition to the wildtype fragment for p21CIP1 (Fig. 2).
q 2005 American Association of Neuropathologists, Inc.
TP53 expression was seen in all but one tumor, typically
at a moderate level. Eleven of 15 tumors (73%) expressed
cyclin D1, although in most cases only weakly. Three of nine
MPNSTs expressed MDM2 protein and two of these at
a strong level.
The expression levels of the cell cycle proteins analyzed
were similarly distributed between sporadic MPNSTs and
NF1-associated MPNSTs. However, for p16INK4A, both cases
with expression (D and H) were sporadic MPNSTs, whereas
both MDM2-positive tumors were from NF1 cases.
DISCUSSION
The molecular mechanisms underlying the development
of MPNSTs are only partly understood. Several lines of evidence
indicate that cell cycle regulators are involved in the progression of this disease. Previous reports concerned with expression
of cell cycle components in MPNSTs have most commonly
used immunohistochemistry for in situ analyses of tissue sections, with no information on the protein size. In the present
study of MPNSTs, we used Western blot analysis to obtain information about possible alterations in the protein sizes caused by
gene alterations or regulation at the expression level such as
phosphorylations and dephosphorylations of proteins. From
the expression pattern of 10 cell cycle regulators used in this
study, the most striking observation was made for p16INK4A.
Only 2 of 15 (13%) MPNSTs revealed expression of p16INK4A,
which is in agreement with previous in situ expression studies
(15, 31, 32). These results and the fact that most of the precursor lesions, neurofibromas, and plexiform neurofibromas
77
JOBNAME: jnen 64#1 2005 PAGE: 5 OUTPUT: Tue January 18 18:20:36 2005
lww/jnen/90813/NEN18628
Ågesen et al
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
FIGURE 1. Protein expression and DNA rearrangement of p16INK4A in MPNSTs. (A) p16INK4A was detected in 2 samples (D and H). In
sample H, weak protein expression was seen, whereas sample D revealed high level of p16INK4A. a-Tubulin was used as a protein
loading control. (B) The Southern blot picture shows the results of all tumor samples run in the present study. N, normal; D,
homozygote deletion; r, restriction fragments with abnormal mobility as well as the normal fragment; R, restriction fragments with
abnormal mobility and only residual amounts of the normal restriction fragments; d, heterozygote deletion. 650 and 651 are
a previously analyzed MPNST and the corresponding blood sample, respectively, and used as controls in our experiments (13). M1
and M2 are 2 samples from the same tumor. ni, sample not included in the study.
express p16INK4A, suggest that lack of functional p16INK4A is
associated with malignant progression (15, 31, 32).
The p16INK4A gene, located at chromosome band 9p21,
is found inactivated by deletions, mutations, and methylation
over a wide range of malignancies (33). The main mechanism
for lack of p16INK4A protein in MPNST has been shown to be
gross alterations of the gene, including homozygote deletions
(13–15). This was later confirmed by interphase FISH and PCR
analyses (34, 35). In this study, we found, as expected, gene
alterations in p16INK4A in most tumor samples. Six of the
15 tumors with gene rearrangement showed a complete inactivation of p16INK4A, confirming that lack of protein expression
is caused by genomic alterations. Three MPNSTs showed a
restriction fragment of abnormal mobility where only residual
amounts of the normal band were seen, whereas homozygote
deletions were observed in three tumors. In our initial study,
we found homozygote deletion in one of 12 MPNSTs (13).
In total, this suggests that 15% of MPNST are homozygously deleted for p16INK4A, which is approximately twice the
78
frequency of sarcomas, generally reported to be 7% (36). Only
one (sample H) of the MPNSTs with heterozygote deletions of
the gene had p16INK4A protein expression. The absence of protein seen in most tumors might be due to complete inactivation
of the gene, and the remaining second allele could be inactivated by mechanisms not detectable by Southern blot analysis. The p16INK4A and p14ARF, both encoded by the CDKN2A
locus and known to be inactivated by promoter hypermethylation
in a wide range of tumors (33, 37), were, however, not methylated. This observation is in keeping with our initial findings
in which none of 12 MPNSTs was methylated at the CDKN2A
locus (13). On the other hand, methylation of these genes has
recently been reported in a few MPNSTs (35). Taking all of the
data together, we conclude that promoter methylation does not
seem to be a major mechanism explaining the absence of
CDKN2A protein. Only one of the two MPNSTs with observed
p16INK4A protein expression had normal gene restriction fragments when hybridized with a p16INK4A probe, indicating that
although alteration of this gene and its encoded protein is
q 2005 American Association of Neuropathologists, Inc.
JOBNAME: jnen 64#1 2005 PAGE: 6 OUTPUT: Tue January 18 18:20:50 2005
lww/jnen/90813/NEN18628
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
Cell Cycle Proteins in MPNST
FIGURE 2. CDK4 protein expression for selected MPNSTs observed by Western blot analysis is shown top left in the figure.
Expression of p21CIP1 from two different runs is shown top right. In one MPNST (sample H), we observed a fragment with a lower
molecular weight (17–18 kDa) in addition to the expected protein band at 21 kDa. a-Tubulin was used as a protein loading control
and is shown below of its respective membrane. Weak protein staining was seen on this membrane for tumor p, and the sample
was reanalyzed and interpreted on a second film as ++ (moderate expression). ni, sample not included in the study.
frequent and most likely important for the progression of the majority of MPNST, it is not a necessity for MPNST development.
Loss of RB1 function is frequently found in a number of
cancers such as retinoblastoma, osteosarcoma, breast cancer,
and small cell lung and bladder carcinomas (16, 38). Previous
studies have investigated the protein level of RB1 in MPNSTs,
as well as in other sarcomas using immunohistochemistry (31,
39–41). RB1 was expressed in all MPNSTs, and by CGH the
same tumors showed a loss of 13q chromosome material in
seven of the samples also expressing RB1 (unpublished data).
These findings might indicate a heterozygote deletion of RB1
in these MPNSTs, but in order to identify possible gene
mutations or other gene rearrangements in the remaining allele
the samples should be submitted to more detailed gene analysis. However, tumor cells showing loss of functional p16INK4A
tend to retain wild-type RB1, whereas cells with no RB1
expression generally express wild-type p16INK4A (42).
The D-type cyclins play a critical role in regulating the G1
restriction point through binding to CDK4 or CDK6, leading to
a subsequent phosphorylation and inactivation of RB1. Cyclin
D1 was, in most cases, weakly expressed in our tumor samples,
which may be explained by the labile nature of the D-type cyclins.
A relatively low frequency of cyclin D1-immunoreactive
MPNSTs has also been reported by other investigators using
immunohistochemistry analysis (39, 40). Amplifications of the
cyclin D1 gene have been detected in different types of
sarcoma (36), but in a previous study we found amplification
of cyclin D2 in only one of 12 MPNSTs (13). No amplification
was detected of cyclin D1 and cyclin D3 in any of the 12
q 2005 American Association of Neuropathologists, Inc.
MPNSTs, showing that amplifications of the D-type cyclins
are not frequent in these tumors (13). In the present study,
cyclin D3 was found to be more intensely expressed compared
with cyclin D1. It is assumed that cyclin D3 is the most widely
expressed D-type cyclin and that it may have other cellular
roles distinct from, but associated with cell cycle regulation,
such as promoting differentiation (43). However, we cannot
exclude that this difference was caused by technical variations
in, for instance, antibody specificity and sensitivity.
CDK4 and CDK2 protein expression was observed in all
MPNSTs. In agreement with our results, Yoo et al reported
expression of CDK4 in 93% of the MPNSTs analyzed as well
as similar findings in other soft tissue sarcomas included in the
study (40). In sharp contrast to these data, Birindelli et al
detected CDK4 expression in only one of 26 MPNSTs analyzed (31). Expression of CDK4 and CDK2 has been described in other sarcomas as well as in carcinomas (44–47). To
our knowledge, the present study describes, for the first time,
CDK2 expression in MPNSTs. An overexpression of CDK2
and CDK4 could be an indication of accelerated cell cycle.
However, the kinases need, in addition to association with
corresponding cyclins, the proper phosphorylations and dephosphorylations at conserved threonine residues to accelerate
the cell cycle (48). Overexpression could be due to amplification of the gene, but we have shown previously that CDK4
amplification is rare in MPNST (13). One may also argue that
CDK4/2 amplification and overexpression of CDK4 or CDK2
would not be a matter of necessity for the tumor development
when p16INK4A is nonfunctional.
79
JOBNAME: jnen 64#1 2005 PAGE: 7 OUTPUT: Tue January 18 18:20:59 2005
lww/jnen/90813/NEN18628
Ågesen et al
The CDK inhibitors, p21CIP1 and p27KIP1, are members of
the same Cip/Kip family with inhibition of cyclin (D, E, and
A)/CDK (4, 6, and 2) complexes as their major function.
p27KIP1 is suggested to be involved in the malignant transformation of neurofibromas based on the observation of decreased
immunostaining of nuclear p27KIP1 in high-grade MPNSTs
versus neurofibromas, and a strong cytoplasmic p27KIP1 staining
in the MPNSTs (32, 39). In our study, p27KIP1 was seen in all
tumors, in most cases at a moderate to high level, but the
difference between nuclear and cytoplasmic staining could not
be determined due to the nature of Western blot analysis. It has
been suggested that an increased expression of p27KIP1 may
also function to protect the cells from the toxic effect of any
high levels of cyclin E and cyclin D1 (49). Supporting these
findings, Kourea et al found a significant relationship between
the high levels of nuclear cyclin E and cytoplasmic p27KIP1 in
MPNST (39). Because p27KIP1 was sequestered to the cytoplasm, the protein would be prevented from inhibiting the
cyclin E/CDK2 complexes and the cancer cells would be
protected from the inhibitory effects of p27KIP1 (49).
It is well documented that TP53 activates p21CIP1
transcription in response to DNA damage and cellular stress
(16). Kourea et al have reported positive p21CIP1 immunoreactivity in 16 of 35 (46%) MPNSTs (39). We observed p21CIP1
protein expression in all tumors; however, in 8 of 15 (53%)
MPNSTs, a moderate to high level was seen. It has been
demonstrated that a certain amount of p21CIP1 is actually
necessary to promote association of CDK4 and the D-type
cyclins and target the complex to the nucleus, but at high
concentration p21CIP1 inhibits the CDK4/cyclin D activity and
arrests the cell cycle (50). In one sample (H), two protein
bands, one normal sized and one with a lower molecular
weight, were seen on the Western blot. Recent studies have
shown that both caspase 3 and proteinase 3 can mediate
p21CIP1 cleavage during early stage of apoptosis (51, 52). Our
band has, however, a slightly larger molecular size than the
protein band observed in the former studies (17 to 18 kDa vs
14 kDa and 10 kDa, respectively), suggesting that other
mechanisms might have caused truncation of p21CIP1 in this
tumor. Even though somatic mutations are not a common
mechanism to inactivate p21 CIP1 (53, 54), a nonsense mutation
that generates a truncated protein cannot be excluded.
The TP53 protein level is normally low due to its short
half-life, and a rise in TP53 is seen as a response to different
types of DNA damages. Accumulation of TP53 is commonly
seen in tumor samples, usually as a consequence of gene mutations, which have been shown to be the major cause of stabilization and inactivation of the protein (55). Expression of TP53,
detected by immunohistochemistry, has been observed in sarcomas including MPNSTs ranging from 29% to 100% in MPNSTs
(31, 32, 39, 40, 56–58). We observed TP53 expression in all but
one MPNST. Supporting the importance of TP53 in MPNST
tumor progression is the finding of low or no levels of TP53 in
neurofibromas (31, 39, 58). A few mutations have been reported
in MPNSTs (18–20, 31), but biallelic inactivation seems to be
rare (17); thus other mechanisms (e.g. binding to viral or cellular oncoproteins) may cause stabilization of the protein (59).
The transcription of MDM2 is activated by a functional
TP53, and is one of the central components in the negative
80
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
regulation of the TP53 protein level in cells. MDM2 interacts
with and prevents TP53 from further stimulating transcription
of downstream genes, and targets TP53 for degradation (60).
MDM2 activity itself is inhibited by p14ARF, one of the two
transcripts derived from the CDKN2A locus (61–63). An
alteration of the CDKN2A locus, as observed in the majority
of the MPNSTs in this study, most likely also alters the
expression of p14ARF, and the absence of p14ARF would
therefore predict high levels of MDM2. However, we found
MDM2 expressed in only three of nine (33%) MPNSTs
analyzed, and one of the three had a seemingly normal
CDKN2A gene locus. Taken together, it is likely that p14ARF
expression is not completely absent in most MPNSTs, and
other regulation mechanisms not studied here further control
the expression of these proteins. Previous studies have also
demonstrated MDM2 protein expression in MPNST in
agreement with our data (31, 40).
The present study describes the protein expression level
of a set of cell cycle components in the same MPNSTs.
Strikingly, absence of the cyclin-dependent kinase inhibitor
p16 INK4A was seen in most cases and found associated with
gross gene alterations of the CDKN2A locus. In contrast to in
situ immunohistochemistry, we could evaluate the protein sizes of
the 10 cell cycle components analyzed. With the exception of
p16INK4, only one antibody showed an abnormally sized protein in a single tumor. The expression levels of the proteins
were overall not skewed among the tumors and were equally
distributed between NF1-associated cases and sporadic ones.
These data suggest that p16 INK4A inactivation is enough for
the abnormal stimulation of the cell cycle, and altered expression of other central components is usually not necessary in
MPNST.
REFERENCES
1. Riccardi VM. Neurofibromatosis: Phenotype, Natural History and
Pathogenesis, 2nd ed. Baltimore: John Hopkins University Press, 2002
2. Legius E, Marchuk DA, Collins FS, et al. Somatic deletion of the
neurofibromatosis type 1 gene in a neurofibrosarcoma supports a tumor
suppressor gene hypothesis. Nat Genet 1993;3:122–26
3. Lothe RA, Slettan A, Saeter G, et al. Alterations at chromosome 17 loci in
peripheral nerve sheath tumors. J Neuropathol Exp Neurol 1995;54:65–73
4. Skuse GR, Kosciolek BA, Rowley PT. Molecular genetic analysis of
tumors in von Recklinghausen neurofibromatosis: Loss of heterozygosity
for chromosome 17. Genes Chromosomes Cancer 1989;1:36–41
5. Glover TW, Stein CK, Legius E, et al. Molecular and cytogenetic analysis
of tumors in von Recklinghausen neurofibromatosis. Genes Chromosomes Cancer 1991;3:62–70
6. Colman SD, Williams CA, Wallace MR. Benign neurofibromas in type 1
neurofibromatosis (NF1) show somatic deletions of the NF1 gene. Nat
Genet 1995;11:90–92
7. Sawada S, Florell S, Purandare SM, et al. Identification of NF1 mutations
in both alleles of a dermal neurofibroma. Nat Genet 1996;14:110–12
8. Mertens F, Rydholm A, Bauer HF, et al. Cytogenetic findings in malignant
peripheral nerve sheath tumors. Int J Cancer 1995;61:793–98
9. Mertens F, Dal Cin P, De Wever I, et al. Cytogenetic characterization of
peripheral nerve sheath tumors: a report of the CHAMP study group.
J Pathol 2000;190:31–38
10. Lothe RA, Karhu R, Mandahl N, et al. Gain of 17q24-qter detected by
comparative genomic hybridization in malignant tumors from patients with
von Recklinghausen’s neurofibromatosis. Cancer Res 1996;56:4778–81
11. Mechtersheimer G, Otano-Joos M, Ohl S, et al. Analysis of chromosomal
imbalances in sporadic and NF1-associated peripheral nerve sheath tumors
by comparative genomic hybridization. Genes Chromosomes Cancer
1999;25:362–69
q 2005 American Association of Neuropathologists, Inc.
JOBNAME: jnen 64#1 2005 PAGE: 8 OUTPUT: Tue January 18 18:21:00 2005
lww/jnen/90813/NEN18628
J Neuropathol Exp Neurol Volume 64, Number 1, January 2005
12. Schmidt H, Taubert H, Meye A, et al. Gains in chromosomes 7, 8q, 15q
and 17q are characteristic changes in malignant but not in benign
peripheral nerve sheath tumors from patients with Recklinghausen’s
disease. Cancer Lett 2000;155:181–90
13. Berner JM, Sorlie T, Mertens F, et al. Chromosome band 9p21 is
frequently altered in malignant peripheral nerve sheath tumors: studies of
CDKN2A and other genes of the pRB pathway. Genes Chromosomes
Cancer 1999;26:151–60
14. Kourea HP, Orlow I, Scheithauer BW, et al. Deletions of the INK4A gene
occur in malignant peripheral nerve sheath tumors but not in neurofibromas. Am J Pathol 1999;155:1855–60
15. Nielsen GP, Stemmer-Rachamimov AO, Ino Y, et al. Malignant transformation of neurofibromas in neurofibromatosis 1 is associated with
CDKN2A/p16 inactivation. Am J Pathol 1999;155:1879–84
16. Sherr CJ. Cancer cell cycles. Science 1996;274:1672–77
17. Lothe RA, Smith-Sorensen B, Hektoen M, et al. Biallelic inactivation
of TP53 rarely contributes to the development of malignant peripheral
nerve sheath tumors. Genes Chromosomes Cancer 2001;30:202–206
18. Nigro JM, Baker SJ, Preisinger AC, et al. Mutations in the p53 gene occur
in diverse human tumor types. Nature 1989;342:705–708
19. Menon AG, Anderson KM, Riccardi VM, et al. Chromosome 17p deletions and p53 gene mutations associated with the formation of malignant
neurofibrosarcomas in von Recklinghausen neurofibromatosis. Proc Natl
Acad Sci USA 1990;87:5435–39
20. Legius E, Dierick H, Wu R, et al. TP53 mutations are frequent in malignant
NF1 tumors. Genes Chromosomes Cancer 1994;10:250–55
21. Cichowski K, Shih TS, Schmitt E, et al. Mouse models of tumor development in neurofibromatosis type 1. Science 1999;286:2172–76
22. Vogel KS, Klesse LJ, Velasco-Miguel S, et al. Mouse tumor model for
neurofibromatosis type 1. Science 1999;286:2176–79
23. St. Croix B, Florenes VA, Rak JW, et al. Impact of the cyclin-dependent
kinase inhibitor p27Kip1 on resistance of tumor cells to anticancer agents.
Nat Med 1996;2:1204–10
24. Florenes VA, Lu C, Bhattacharya N, et al. Interleukin-6 dependent induction of the cyclin dependent kinase inhibitor p21WAF1/CIP1 is lost during
progression of human malignant melanoma. Oncogene 1999;18:1023–32
25. Florenes VA, Bhattacharya N, Bani MR, et al. TGF-beta mediated G1 arrest
in a human melanoma cell line lacking p15INK4B: Evidence for cooperation between p21Cip1/WAF1 and p27Kip1. Oncogene 1996;13:2447–57
26. Florenes VA, Maelandsmo GM, Forus A, et al. MDM2 gene amplification
and transcript levels in human sarcomas: Relationship to TP53 gene
status. J Natl Cancer Inst 1994;86:1297–302
27. Serrano M, Hannon GJ, Beach D. A new regulatory motif in cell-cycle control
causing specific inhibition of cyclin D/CDK4. Nature 1993;366:704–707
28. Feinberg AP, Vogelstein B. A technique for radiolabeling DNA restriction
endonuclease fragments to high specific activity. Anal Biochem 1983;132:6–13
29. Smith-Sorensen B, Lind GE, Skotheim RI, et al. Frequent promoter hypermethylation of the O6-methylguanine-DNA methyltransferase (MGMT)
gene in testicular cancer. Oncogene 2002;21:8878–84
30. Herman JG, Graff JR, Myohanen S, et al. Methylation-specific PCR:
a novel PCR assay for methylation status of CpG islands. Proc Natl Acad
Sci USA 1996;93:9821–26
31. Birindelli S, Perrone F, Oggionni M, et al. Rb and TP53 pathway alterations in sporadic and NF1-related malignant peripheral nerve sheath
tumors. Lab Invest 2001;81:833–44
32. Zhou H, Coffin CM, Perkins SL, et al. Malignant peripheral nerve sheath
tumor: a comparison of grade, immunophenotype, and cell cycle/growth
activation marker expression in sporadic and neurofibromatosis 1-related
lesions. Am J Surg Pathol 2003;27:1337–45
33. Liggett WH Jr, Sidransky D. Role of the p16 tumor suppressor gene in
cancer. J Clin Oncol 1998;16:1197–1206
34. Perry A, Kunz SN, Fuller CE, et al. Differential NF1, p16, and EGFR
patterns by interphase cytogenetics (FISH) in malignant peripheral nerve
sheath tumor (MPNST) and morphologically similar spindle cell
neoplasms. J Neuropathol Exp Neurol 2002;61:702–709
35. Perrone F, Tabano S, Colombo F, et al. p15INK4b, p14ARF, and
p16INK4a inactivation in sporadic and neurofibromatosis type 1-related
malignant peripheral nerve sheath tumors. Clin Cancer Res 2003;9: 4132–38
36. Maelandsmo GM, Berner JM, Florenes VA, et al. Homozygous deletion
frequency and expression levels of the CDKN2 gene in human sarcomas:
relationship to amplification and mRNA levels of CDK4 and CCND1. Br
J Cancer 1995;72:393–98
q 2005 American Association of Neuropathologists, Inc.
Cell Cycle Proteins in MPNST
37. Esteller M, Corn PG, Baylin SB, et al. A gene hypermethylation profile of
human cancer. Cancer Res 2001;61:3225–29
38. Weinberg RA. Tumor suppressor genes. Science 1991;254:1138–46
39. Kourea HP, Cordon-Cardo C, Dudas M, et al. Expression of p27(kip) and
other cell cycle regulators in malignant peripheral nerve sheath tumors and
neurofibromas: The emerging role of p27(kip) in malignant transformation of neurofibromas. Am J Pathol 1999;155:1885–91
40. Yoo J, Park SY, Kang SJ, et al. Altered expression of G1 regulatory proteins
in human soft tissue sarcomas. Arch Pathol Lab Med 2002;126: 567–73
41. Karpeh MS, Brennan MF, Cance WG, et al. Altered patterns of retinoblastoma gene product expression in adult soft-tissue sarcomas. Br J Cancer
1995;72:986–91
42. Sherr CJ, Roberts JM. Inhibitors of mammalian G1 cyclin-dependent
kinases. Genes Dev 1995;9:1149–63
43. Bartkova J, Lukas J, Strauss M, et al. Cyclin D3: requirement for G1/S
transition and high abundance in quiescent tissues suggest a dual role in
proliferation and differentiation. Oncogene 1998;17:1027–37
44. Sasano H, Frost AR, Saitoh R, et al. Immunolocalization of cyclins D and
E and cyclin dependent kinase (cdk) 2 and 4 in human breast carcinoma.
Anticancer Res 1997;17:3685–90
45. Skomedal H, Kristensen GB, Lie AK, et al. Aberrant expression of the cell
cycle associated proteins TP53, MDM2, p21, p27, cdk4, cyclin D1, RB,
and EGFR in cervical carcinomas. Gynecol Oncol 1999;73:223–28
46. Ito K, Sasano H, Yoshida Y, et al. Immunohistochemical study of cyclins
D and E and cyclin dependent kinase (cdk) 2 and 4 in human endometrial
carcinoma. Anticancer Res 1998;18:1661–64
47. Benassi MS, Molendini L, Gamberi G, et al. Alteration of pRb/p16/cdk4 regulation in human osteosarcoma. Int J Cancer 1999;84:489–93
48. Morgan DO. Principles of CDK regulation. Nature 1995;374:131–34
49. Weinstein IB. Disorders in cell circuitry during multistage carcinogenesis:
the role of homeostasis. Carcinogenesis 2000;21:857–64
50. LaBaer J, Garrett MD, Stevenson LF, et al. New functional activities for
the p21 family of CDK inhibitors. Genes Dev 1997;11:847–62
51. Jin YH, Yoo KJ, Lee YH, et al. Caspase 3-mediated cleavage of
p21WAF1/CIP1 associated with the cyclin A-cyclin-dependent kinase 2
complex is a prerequisite for apoptosis in SK-HEP-1 cells. J Biol Chem
2000;275:30256–63
52. Pendergraft WF III, Rudolph EH, Falk RJ, et al. Proteinase 3 sidesteps
caspases and cleaves p21(Waf1/Cip1/Sdi1) to induce endothelial cell
apoptosis. Kidney Int 2004;65:75–84
53. Watanabe H, Fukuchi K, Takagi Y, et al. Molecular analysis of the
Cip1/Waf1 (p21) gene in diverse types of human tumors. Biochim
Biophys Acta 1995;1263:275–80
54. Mousses S, Ozcelik H, Lee PD, et al. Two variants of the CIP1/WAF1 gene
occur together and are associated with human cancer. Hum Mol Genet
1995;4:1089–92
55. Levine AJ. p53, the cellular gatekeeper for growth and division. Cell
1997;88:323–31
56. Watanabe T, Oda Y, Tamiya S, et al. Malignant peripheral nerve sheath
tumors: High Ki67 labeling index is the significant prognostic indicator.
Histopathology 2001;39:187–97
57. Kawai A, Noguchi M, Beppu Y, et al. Nuclear immunoreaction of p53
protein in soft tissue sarcomas: A possible prognostic factor. Cancer 1994;
73:2499–2505
58. Halling KC, Scheithauer BW, Halling AC, et al. p53 expression in neurofibroma and malignant peripheral nerve sheath tumor: an immunohistochemical study of sporadic and NF1-associated tumors. Am J Clin Pathol
1996;106:282–88
59. Blagosklonny MV. Loss of function and p53 protein stabilization. Oncogene 1997;15:1889–93
60. Woods DB, Vousden KH. Regulation of p53 function. Exp Cell Res
2001;264:56–66
61. Funk JO. Cancer cell cycle control. Anticancer Res 1999;19:4772–80
62. Pomerantz J, Schreiber-Agus N, Liegeois NJ, et al. The Ink4a tumor
suppressor gene product, p19Arf, interacts with MDM2 and neutralizes
MDM2’s inhibition of p53. Cell 1998;92:713–23
63. Zhang Y, Xiong Y, Yarbrough WG. ARF promotes MDM2 degradation
and stabilizes p53: ARF-INK4a locus deletion impairs both the Rb and
p53 tumor suppression pathways. Cell 1998;92:725–34
64. Molenaar WM, Ladde BE, Schraffordt KH, et al. Two epithelioid
malignant schwannomas in a patient with neurofibromatosis: cytology,
histology and DNA-flow-cytometry. Pathol Res Pract 1989;184:529–34
81