Comparative pathology of nerve sheath tumors in mouse models and humans.

Abstract

Despite the progress made in our understanding of the biology of neurofibromatosis (NF), the long-term clinical outcome for affected patients has not changed significantly in the past decades, and both NF1 and NF2 are still associated with a significant morbidity and a decreased life span. A number of NF1 and NF2 murine models have been generated to aid in the study of NF tumor biology and in the development of targeted therapies for NF patients. A single, universal pathological classification of the lesions generated in these murine models is essential for the validation of the models, for their analysis and comparison with other models, and for their future effective use in preclinical treatment trials. For the formulation of a pathological classification of these lesions, the WHO classification of human tumors was used as a reference. However, it was not adopted for the classification of the GEM lesions because of some important differences between the human and murine lesions. A novel classification scheme for peripheral nerve sheath tumors in murine models was therefore devised.

Full text

Page 1

[CANCER RESEARCH 64, 3718–3724, May 15, 2004]
Meeting Report
Comparative Pathology of Nerve Sheath Tumors in Mouse Models and Humans
Anat O. Stemmer-Rachamimov,1David N. Louis,1,14Gunnlaugur P. Nielsen,1Cristina R. Antonescu,3
Alexander D. Borowsky,4Roderick T. Bronson,5Dennis K. Burns,6Pascale Cervera,8Margaret E. McLaughlin,9
Guido Reifenberger,10Michael C. Schmale,11Mia MacCollin,2Richard C. Chao,12Karen Cichowski,13
Michel Kalamarides,8Shanta M. Messerli,2Andrea I. McClatchey,14Michiko Niwa-Kawakita,8Nancy Ratner,15
Karlyne M. Reilly,16Yuan Zhu,7,17and Marco Giovannini8
Departments of1Pathology and2Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts;3Department of Pathology, Memorial Sloan-
Kettering Cancer Center, New York, New York;4Medical Pathology Center for Comparative Medicine, University of California at Davis, Davis, California;5Department of
Pathology, Harvard Medical School, Boston, Massachusetts;
Center, Dallas, Texas;8INSERM U434, Fondation Jean Dausset-CEPH, Paris, France;9Department of Pathology, Brigham and Women’s Hospital, Boston, Massachusetts;
10Department of Neuropathology, Heinrich-Heine-University, Duesseldorf, Germany;11Division of Marine Biology and Fisheries, Rosenstiel School of Marine and Atmospheric
Science, University of Miami, Miami, Florida;12Department of Medicine, Veterans’ Affairs Medical Center and University of California at San Francisco, San Francisco,
California;
Hospital Cancer Center, Harvard Medical School Department of Pathology, Charlestown, Massachusetts;15Department of Cell Biology, Neurobiology, and Anatomy, University
of Cincinnati College of Medicine, Cincinnati, Ohio;16Genetic Modifiers of Tumorigenesis Section, Mouse Cancer Genetics Program, National Cancer Institute at Frederick,
Frederick, Maryland; and17Division of Molecular Medicine and Genetics, University of Michigan Medical School, Ann Arbor, Michigan
6Department of Pathology and
7Center for Developmental Biology, University of Texas Southwestern Medical
13Genetics Division, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts;
14Massachusetts General
Abstract
Despite the progress made in our understanding of the biology of
neurofibromatosis (NF), the long-term clinical outcome for affected pa-
tients has not changed significantly in the past decades, and both NF1 and
NF2 are still associated with a significant morbidity and a decreased life
span. A number of NF1 and NF2 murine models have been generated to
aid in the study of NF tumor biology and in the development of targeted
therapies for NF patients. A single, universal pathological classification of
the lesions generated in these murine models is essential for the validation
of the models, for their analysis and comparison with other models, and
for their future effective use in preclinical treatment trials. For the for-
mulation of a pathological classification of these lesions, the WHO classi-
fication of human tumors was used as a reference. However, it was not
adopted for the classification of the GEM lesions because of some impor-
tant differences between the human and murine lesions. A novel classifi-
cation scheme for peripheral nerve sheath tumors in murine models was
therefore devised.
Introduction
The neurofibromatoses (NFs) NF1 and NF2 are autosomal domi-
nant tumor suppressor gene syndromes in which affected individuals
are predisposed to develop multiple tumors and hamartomas. NF1 and
NF2 are clinically and genetically distinct (1).
NF1affectsapproximately1in3000births.ThehallmarkofNF1isthe
development of multiple neurofibromas and plexiform neurofibromas.
Plexiform neurofibromas may undergo malignant transformation to ma-
lignant peripheral nerve sheath tumors (MPNSTs). Other neoplasms
associated with NF1 include pilocytic astrocytomas, diffuse astrocyto-
mas, leukemias, and pheochromocytomas. Nonneoplastic manifestations
include cafe ´ au lait macules, skin fold freckling, Lisch nodules, learning
disabilities, skeletal defects, and hypertension (1).
NF2 affects approximately 1 in 35,000 individuals, and its hallmark is
the development of bilateral vestibular schwannomas. Patients often
suffer from multiple schwannomas, meningiomas, and ependymomas as
well as a variety of clinically asymptomatic lesions, including Schwann
cell tumorlets, glial microhamartomas, and meningioangiomatosis (1, 2).
Despite the progress made in our understanding of the biology of
NF1 and NF2 in recent years, the mainstay of treatment remains local
tumor control by repeated surgeries, and the long-term clinical out-
come of these patients has not changed significantly (1, 3–5).
The development of several NF1 and NF2 murine models provides
a unique opportunity to gain insight into tumor initiation and progres-
sion in these syndromes as well as to develop and test different
prognostic and therapeutic tools. Furthermore, NF murine models
may aid in development of therapies for sporadic tumors because the
NF1 and NF2 genes are involved in the pathogenesis of some sporadic
MPNSTs and plexiform neurofibromas (6), meningiomas, schwanno-
mas, ependymomas, and mesotheliomas (7).
Biallelic inactivation of Nf1 (Nf1?/?) or Nf2 (Nf2?/?) in the
mouse is embryonic lethal (8–12), and although hemizygosity for one
of the murine Nf genes (Nf1?/? or Nf2?/?) is associated with
increased incidence of tumors, these are not peripheral nerve sheath
tumors [PNSTs (10, 13–18)]. Therefore, different strategies had to be
used to promote development of PNSTs, including the generation of
chimeric mice composed in part of ?/? cells for the relevant Nf gene
(19) or the generation of conditional knockout models with disruption
of the respective Nf gene only in specifically targeted cells, using
tissue-specific promoters and Cre-LoxP technology (14, 20–22). In
addition, models have been generated to identify functional domains
of the Nf1 and Nf2 proteins [including exon-specific knockout mice
(23), models with overexpression of the mutant gene in a selective cell
population (24), and models in which Nf mutants were crossed with
other knockout mice, enabling the evaluation of cooperation among
Nf1, Nf2, and other tumor-associated genes (such as p53 or EGFR) in
Schwann cells (19, 25, 26)].
Materials and Methods
A panel of 10 pathologists with different areas of expertise re-
viewed 56 lesions from 8 laboratories representing most published
and unpublished NF mouse models (Appendix A) and compared their
histological features with the corresponding human tumors. The aim
3718
Received 12/31/03; revised 3/9/04; accepted 3/11/04.
Grant support: United States Army Medical Research and Material Command Grant
DAMD17-00-0594.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
Note: A. O. Stemmer-Rachamimov and M. Giovannini were the meeting organizers.
We acknowledge pathologists that contributed to the paper (D. N. Louis and G. P.
Nielsen) and the other pathology panel members, who are listed alphabetically (C. R.
Antonescu, A. D. Borowsky, R. T. Bronson, D. K. Burns, P. Cervera, M. E. McLaughlin,
G. Reifenberger, and M. C. Schmale), and the neurologist, M. MacCollin. The mouse
modelers who presented and contributed pathology specimens are listed alphabetically
(R. C. Chao, K. Cichowski, M. Kalamarides, S. M. Messerli, A. I. McClatchey, M.
Niwa-Kawakita, N. Ratner, K. M. Reilly, and Y. Zhu).
Requests for reprints: Marco Giovannini, INSERM U434, Foundation Jean Dausset-
CEPH, 27 rue Juliette Dodu, 75010 Paris, France. E-mail:marco.giovannini@cephb.fr.

Page 2

of this workshop was to review all available PNSTs from the Nf1 and
Nf2 murine models generated to date and develop a pathological
classification to characterize these lesions in a uniform and consistent
manner. The meeting and pathological classification follow a format
successfully adopted by similar pathology workshops for murine
models of other organ systems (27, 28).
Results
A Separate GEM Classification Proposed. In formulating the
GEM classification of PNSTs, the panel referred to criteria and
terminology of the WHO classification of human nervous system
tumors (29). However, although the WHO classification served as a
useful point of reference, it was not adopted for the classification of
the GEM lesions because there were some important differences
between the human and murine lesions. The WHO classification of
human PNSTs is based on both histological and clinical characteris-
tics (and describes clinicopathological entities), whereas the GEM
classification must necessarily be based on histological features alone
because further work is needed to determine the clinical features of
murine PNSTs. Furthermore, although many of the histological fea-
tures displayed by GEM PNSTs were similar to those observed in
human tumors, some GEM lesions presented unique and distinct
morphology and did not fit into the human classification by histolog-
ical criteria.
For these reasons, a separate classification scheme for GEM is
proposed (Table 1), which is based on morphological criteria alone.
Specifically, grading in the GEM classification refers merely to the
presence or absence of histological features such as high cellularity,
necrosis, nuclear pleomorphism, and brisk mitotic activity. As op-
posed to the WHO classification, where grade is an indicator of
biological behavior, grading in the GEM classification is applied as a
histological descriptor only, and it has no predictable bearing on the
biological behavior of the tumors in the animals.
The WHO classification of PNSTs includes clinicopathological
entities that are benign (WHO grade I) and defined by their cell
composition as schwannomas, neurofibromas, and perineuriomas, and
one malignant entity (WHO grade III or IV), the MPNST (Appendix
B). In the GEM classification, grade I tumors are similarly defined by
their cell composition as GEM neurofibromas, GEM schwannomas,
and GEM perineuriomas. For the histologically malignant tumors
(GEM grade III), the term “GEM PNST” is preferred because the
tumors are composed of immature cells with some differentiation, and
the distinction among the three above-mentioned entities becomes
difficult (Table 2).
Comparative Pathology of Human and Murine Neurofibromas.
Human neurofibromas (WHO grade I) are defined as benign nerve
sheath tumors composed of a mixture of mature Schwann cells,
perineurial cells, and fibroblasts. Clinically, neurofibromas may be
dermal, nodular, or plexiform. Dermal and nodular neurofibromas are
superficial or deep lesions, respectively, that arise in small nerves and
may be well circumscribed or diffusely infiltrative. Plexiform neuro-
fibromas arise almost exclusively in the context of NF1, causing
diffuse enlargement of an affected plexus of large nerve and may
undergo malignant transformation. Histologically, human neurofibro-
mas are often hypocellular, composed of a mixed population of cells
in a myxoid background (Fig. 1A). The mixed cell composition is
important for diagnosis and can be confirmed by immunohistochem-
istry, which shows only a subset of S-100 protein-positive cells
(Schwann cells), or by electron microscopy (EM). The lesions are
often transversed by nerve axons and may contain varying amounts of
collagen fibers and mast cells.
GEM neurofibromas (grade I) are similarly defined as tumors
composed of a mixture of Schwann cells, perineurial cells, and fibro-
Table 1 GEM classification of PNSTsa
GEM neurofibroma, grade I
GEM schwannoma
Grade I
Grade II
GEM perineurioma, grade I
GEM PNST (NOS)
Grade II
Grade III
Grade III with divergent differentiation
GEM peripheral nerve hyperplasia
Diffuse, “global”
Schwann cell hyperplasia
aPNST, peripheral nerve sheath tumor; NOS, not otherwise specified.
Table 2 GEM nerve sheath tumor classification: diagnostic criteria and definitions
GEM nerve sheath tumor diagnostic criteria
A spindled cell tumor arising in a mouse model can be designated as a nerve sheath tumor when at least one of the following criteria is met:
The lesion is clearly originating from a nerve.
In case of a high-grade lesion (see below), the lesion is clearly originating in a definite low-grade nerve sheath tumor.
The tumor contains S-100 immunopositive cells and/or there is evidence on EMaof Schwannian and/or perineurial differentiation.
Tumors are described using 3 different parameters: cellular composition, grade, and pattern of growth and spread.
1. Cellular composition: as defined by histology and immunohistochemistry (S-100-positive cells are present) and/or EM.
GEM schwannoma: a tumor composed of mature Schwann cells, as shown by immunohistochemistry or EM.
GEM neurofibroma: a tumor composed of a mixed cell composition: mature Schwann cells, fibroblasts, and perineurial cells, as shown by immunohistochemistry or EM.
GEM perineurioma: a tumor composed of mature perineurial cells as shown by EM.
GEM PNST (not otherwise specified): a tumor arising from a peripheral nerve, with some evidence of Schwannian differentiation, by immunohistochemistry or EM, that
does not fit any of the above categories.
Divergent differentiation: differentiation along mesenchymal, epithelial, or neuroendocrine lines in a PNST.
2. Grade
Grade evaluation depends on morphological features only. These include cellularity, pleomorphism, mitotic activity, and necrosis.
Grade I: GEM tumor with low cellularity. Tumor cells have bland/uniform nuclear cytology. Mitoses are absent, and there is no necrosis.
Grade II: GEM tumor with increased cellularity, nuclear atypia, and some mitotic activity. These lesions have histological similarities to low-grade MPNSTs in humans but
may also include GEM cellular schwannomas (GEM schwannoma, grade II).
Grade III: GEM tumor with marked cellularity, nuclear atypia, and high rate of mitotic activity. May contain areas of necrosis or hemorrhage. These lesions histologically
correspond to MPNSTs in the WHO classification.
3. Stage (pattern of growth and spread)
In addition to the histological features of the tumors, the panel recommends describing the following features:
A) Interface of the tumor with adjacent normal tissues (circumscribed vs. infiltrative).
When infiltrative, the pattern of infiltration should be noted (destructive vs. infiltration with preservation of tissue).
B) Tumor size, measured as per greatest diameter.
C) Metastasis, evidence of metastasis in liver or lung. However, the panel recognizes that in some of the GEM models, the distinction between multiple primary tumors
arising in targeted cells and metastasis may be impossible.
aEM, electron microscopy; PNST, peripheral nerve sheath tumor; MPNST, malignant PNST.
3719
PATHOLOGY CLASSIFICATION OF PNSTS IN NF MOUSE MODELS

Page 3

blasts, as shown by immunohistochemistry and/or ultrastructural anal-
ysis. Histologically, grade I GEM neurofibromas are closely similar to
human neurofibromas. The tumors display low cellularity, minimal
nuclear atypia, and rare mitoses. They are composed of elongated, thin
nuclei in a loose, myxoid background (Fig. 2A) and contain collagen
bundles and mast cells. A definite case of plexiform neurofibroma was
not observed by the panel in the reviewed GEM lesions.
Comparative Pathology of Human and Murine Schwannomas.
The WHO defines human schwannomas as benign nerve sheath
tumors (WHO grade I) composed of mature Schwann cells. Schwan-
nomas are often encapsulated and do not infiltrate adjacent soft tissue.
Histologically, classic features include a biphasic growth pattern with
dense and loose areas (Antoni A and Antoni B, respectively) and the
presence of Verocay bodies (rows of palisaded Schwann cell nuclei
separated by a nuclear space; Fig. 1B). As opposed to neurofibromas,
the presence of necrosis, mitotic activity, or nuclear pleomorphism in
schwannomas does not necessarily suggest malignant transformation.
This is especially true in cellular schwannomas, a histological variant
with high cellularity, moderate nuclear pleomorphism, and scattered
mitotic figures (Fig. 1C). Despite all these “aggressive” histological
features, cellular schwannomas are clinically benign tumors and thus
have a different clinical behavior from neurofibromas displaying the
same features (which would be classified as MPNSTs, WHO grade
III). Confirmation, by S-100 protein immunohistochemistry or EM, of
a pure Schwann cell population is essential for diagnosis (Fig. 1D).
GEM schwannomas (grade I) are similarly defined as murine
tumors composed of mature Schwann cells, as shown by immunohis-
tochemistry (diffuse S-100 staining) and/or ultrastructural analysis.
Histologically, GEM schwannomas primarily display a dense pattern
of growth (Antoni A) composed of intersecting fascicles or storiform
designs. Verocay bodies are not common, but a similar pattern may be
identified in some tumors (Fig. 2B). As opposed to human schwan-
3720
Fig. 1. Pathology of human peripheral nerve sheath tumors.
A, neurofibroma composed of spindled and round cells and
collagen bundles in a loose, myxoid background. B, Schwan-
noma with prominent, characteristic formation of Verocay bod-
ies (arrows). C, cellular schwannoma, displaying moderate cel-
lularity and nuclear pleomorphism and the absence of Verocay
bodies. D, ultrastructural features of a schwannoma showing a
Verocay body composed of intertwining of cytoplasmic pro-
cesses surrounded by basal lamina. E, whorl formation and
fascicles of spindle-shaped cells in a soft tissue perineurioma. F,
diagnostic ultrastructural features of a perineurial cell include
elongated cytoplasmic processes, a basal lamina (long arrow),
and prominent pinocytotic vescicles (short arrow). G, malig-
nant peripheral nerve sheath tumor composed of atypical spin-
dle-shaped cells arranged in fascicles. H, large cells with abun-
dant eosinophilic cytoplasm (arrow) represent skeletal muscle
differentiation in a malignant peripheral nerve sheath tumor.
PATHOLOGY CLASSIFICATION OF PNSTS IN NF MOUSE MODELS

Page 4

nomas, GEM schwannomas are not encapsulated and often infiltrate
into adjacent soft tissues, in a nondestructive fashion, entrapping fat
and muscle (Fig. 2C). Other features commonly observed in human
schwannomas such as hyalinized vessels, vascular malformations,
thrombosed vessels, hemosiderin-laden macrophages, or hemorrhage,
were not observed in GEM schwannomas.
Grade II GEM schwannomas display moderate nuclear pleomor-
phism, increased cellularity, and scattered mitotic figures (Fig. 2D). It
has not yet been determined whether these lesions correspond in their
behavior to human cellular schwannomas or to human MPNSTs
(WHO grade III).
Comparative Pathology of Human and Murine Perineuriomas.
Human perineurioma is a benign PNST composed of mature perineur-
ial cells. There is an intraneural variant arising from a nerve, in which
the cells form whorls around nerve fibers in a pseudo-onion bulb
pattern, and a soft tissue variant, unassociated with nerve, that may
histologically mimic a meningioma or fibroma (Fig. 1E). Tumor cells
are epithelial membrane antigen immunopositive and S-100 protein
negative and have ultrastructural features of normal perineurial cells
(Fig. 1F).
GEM perineuriomas (grade I) are similarly defined as tumors
composed of perineurial cells. Unfortunately, epithelial membrane
antigen immunohistochemistry has been problematic in mouse tissues.
As a result, ultrastructural confirmation is essential for diagnosis
because the pseudo-onion bulb pattern is not specific and can be seen
in other tumor types. A definite case of GEM perineurioma was not
observed by the panel in this meeting. A GEM tumor composed of
thin elongated cells with tight whorls was reviewed, but electron
microscopic data were not available on this case.
The GEM classification includes two other histological entities,
which are not part of the WHO scheme: the GEM nerve sheath tumor
not otherwise specified; and the GEM nerve sheath tumor with diver-
3721
Fig. 2. Pathology of murine peripheral nerve sheath tumors.
A, GEM neurofibromas, grade I, composed of small stellate
cells in a loose, myxoid background. B, GEM schwannoma,
grade I, with vague formation of Verocay bodies. C, GEM
schwannoma widely infiltrating skeletal muscle. D, GEM schw-
annoma, grade II, a moderately cellular tumor with rare mitotic
figures. E, GEM peripheral nerve sheath tumor, grade III, a
cellular tumor composed of spindle-shaped cells with marked
nuclear atypia arranged in fascicles. F, large cells with abundant
cytoplasm (arrows) represent skeletal muscle differentiation in
a GEM peripheral nerve sheath tumor, grade III. G, GEM global
nerve hyperplasia. H, GEM diffuse Schwann cell hyperplasia.
PATHOLOGY CLASSIFICATION OF PNSTS IN NF MOUSE MODELS

Page 5

gent differentiation. The GEM nerve sheath tumor not otherwise
specified is defined as any tumor clearly arising from a nerve or
displaying Schwannian differentiation that does not fit into any of the
above-mentioned categories. The GEM nerve sheath tumor with di-
vergent differentiation is any GEM nerve sheath tumor in which there
is clear evidence of divergent differentiation by histology, immuno-
histochemical staining, or EM (for example muscle, bone, cartilage, or
epithelium). Divergent differentiation was seen in grade III GEM
nerve sheath tumors.
Comparative Pathology of Human and Murine MPNSTs. In the
WHO classification, MPNSTs are defined as malignant tumors arising
from peripheral nerve sheath cells (WHO grade III or IV). Histolog-
ically, low-grade MPNSTs are cellular lesions with occasional en-
larged hyperchromatic nuclei and scattered mitotic figures, whereas
high-grade MPNSTs exhibit high cellularity with marked nuclear
pleomorphism and brisk mitotic activity (Fig. 1G) and may display
areas of necrosis. In some of the tumors, mesenchymal differentiation
(including rhabdomyoblasts, cartilage, and bone) or epithelial and
neuroendocrine differentiation may be observed (Fig. 1H). Because of
the large spectrum of histological patterns displayed, the distinction of
MPNSTs from other malignant tumors of soft tissues may be difficult.
The WHO classification therefore requires that, to be classified as a
MPNST, a tumor has to either arise from a nerve or from a pre-
existing benign PNST or show perineurial or Schwann cell differen-
tiation by EM or immunohistochemistry.
Histologically, GEM grade III nerve sheath tumors are similar to
the human MPNSTs and are likewise defined as tumors that display
high cellularity, anaplasia, and frequent mitotic figures; that either
arise from a peripheral nerve or a lower grade PNST; or that show
differentiation along nerve sheath lines by immunohistochemistry or
EM (Fig. 2E). As in the human tumors, GEM grade III tumors may
display divergent differentiation (GEM PNST with divergent differ-
entiation, grade III) and histologically mimic other soft tissue tumors
(Fig. 2F). Therefore, as in the human MPNST with divergent differ-
entiation, demonstration of perineurial or Schwann cell differentiation
by EM or by immunohistochemistry is essential for diagnosis. The
term “GEM PNST, grade III” is preferred to “GEM MPNST” because
the word “malignant” is a clinical descriptor, and the clinical behavior
of these murine lesions is still unknown.
Murine Hyperplastic Lesions of the Peripheral Nerve. Human
hyperplastic peripheral nerve lesions consist of diffuse or nodular
enlargements of the nerve in which one or more of the normal
components are increased (30). Hyperplastic lesions are not included
in the WHO classification, which is limited to classification of neo-
plastic lesions.
The GEM hyperplastic lesions are classified according to the nor-
mal nerve component involved. Diffuse “global” nerve hyperplasia is
defined as diffuse enlargement of the involved nerve, which appears
histologically normal except for its increased size and thickened
endoneurium (Fig. 2G). Schwann cell hyperplasia is defined as a
multifocal, diffuse process, in which involved nerves are enlarged due
to increase of the Schwann cell component only (Fig. 2H).
Approach to GEM PNST Classification. Proper interpretation
and definition of murine lesions are dependent on high-quality and
uniform specimen processing (sampling, fixation, and staining) and
analysis (immunostaining and EM). To facilitate and standardize the
characterization of PNSTs in murine models, the panel developed a
practical approach to the handling of these specimens. These recom-
mendations do not replace but rather complement the previous rec-
ommendations made at the workshop that focused on central nervous
system GEM tumors (28).
Table 3 Summary of peripheral nerve sheath tumors in NF mouse models reviewed at the workshop
Tumor (GEM)Genotype Source/reference
GEM schwannomas
I
Nf2?/?:Nf1?/?
P0Cre; Nf2flox/flox; Nf1?/?
P0-Sch-?(39–121)
Nf2?/?:Nf1?/?
P0Cre; Nf2flox/flox; Nf1?/?
Nf2?/?:p53?/?cis
I. Saotome, C-H. Liu, M. Niwa-Kawakita, M. Giovannini, A. I. McClatchey, unpublished results
M. Niwa-Kawakita, M. Giovannini, unpublished results
24
I. Saotome, C-H. Liu, M. Niwa-Kawakita, M. Giovannini, A. I. McClatchey, unpublished results
M. Niwa-Kawakita, M. Giovannini, unpublished results
13
II
GEM neurofibromas
I
Nf1flox/KO;Krox20-cre?
Nf1?/? chimera
Nf1?/?:p53?/? cis
21
19
K. M. Reilly, D. A. Loisel, J. Ledger, R. T. Bronson, K. Cichowski, S. Shih, T. Jacks, unpublished resultsII
GEM PNSTa
III
Nf1?/?:p53?/?
P0-Sch-?(39–121)
p53K0/?,Nf1K0/flox;Krox20-cre?
19
24
L. F. Parada, F. Guignard, Y. Zhu, unpublished results
aPNST, peripheral nerve sheath tumor.
Table 4 WHO Classification for PNSTsa
Schwannoma (WHO grade I)
Cellular
Plexiform
Melanotic
Neurofibroma (WHO grade I)
Plexiform
Perineurioma
Inraneural perineurioma
Soft tissue perineurioma
MPNST (WHO grade III or IV)
Epitheloid
MPNST with divergent mesenchymal differentiation and/or epithelial differentiation
Melanotic
Melanotic psammomatous
aPNST, peripheral nerve sheath tumor; MPNST, malignant PNST.
Table 5 Recommendations for mouse work-up for PNSTsa
Grade I lesions: perform S-100:
Diffuse positivity 3 GEM schwannoma
Focal positivity 3 GEM neurofibroma
Negative 3 EM is essential.
With Schwann cell or perineurial differentiation on EM 3 GEM neurofibroma
No Schwann cell or/and perineurial cell differentiation on EM 3 tumor is not a
GEM PNST and should be classified according to EM features or markers as
fibrosarcoma, rhabdomyosarcoma, spindled cell carcinoma, etc.
Grade II or III lesions: perform S-100, desmin, actin:
Diffuse S-100 positivity 3 GEM schwannoma
With other markers positive 3 GEM schwannoma grade II or III with divergent
differentiation (Triton tumor)
With focal S-100 positivity 3 GEM PNST grade II or III
With other markers positive 3 GEM PNST grade II or III with divergent
differentiation.
Negative for S-100 3 EM (essential).
With Schwann cell or perineurial differentiation on EM 3 GEM PNST grade II or
III
Without Schwann cell or perineurial differentiation on EM 3 not a GEM PNST and
should be classified according to EM features or markers as fibrosarcoma,
rhabdomyosarcoma, spindled cell carcinoma, etc.
aPNST, peripheral nerve sheath tumor; EM, electron microscopy.
3722
PATHOLOGY CLASSIFICATION OF PNSTS IN NF MOUSE MODELS

Page 6

Peripheral nerve tumors in the mouse present as soft tissue tumors;
their origin from a peripheral nerve may not be easily recognized, and
it is therefore advisable that sampled lesions be removed with adjacent
structures to enable microscopic examination. The size of the tumor
and its location and relationship to adjacent structures (association
with a nerve, its infiltrating or sharp margin) should be carefully
noted.
Because the classification of GEM PNSTs is based on the cellular
differentiation of tumor cells (schwannian, fibroblastic, perineurial, or
divergent), immunohistochemistry and EM are essential for classifi-
cation. A tissue sample should be submitted in glutaraldehyde for EM
on every tumor. It is strongly recommended that in new, uncharac-
terized murine models, EM be performed on all lesions. Once a
murine model has been characterized, immunohistochemical analysis
may suffice for the analysis of grade I lesions, and EM may be
reserved only for the S-100 protein-negative cases (Appendix C).
In grade III GEM lesions, a panel of immunohistochemical stains is
recommended to highlight divergent differentiation and exclude other
types of soft tissue tumors. Immunostaining for S-100 protein
(Schwann cell) and actin and desmin (muscle) should be performed.
Lesions that are negative for these markers should be explored further
by ultrastructural studies. Additional immunohistochemical stains
may be necessary to exclude tumors when soft tissue tumors arise in
specific sites. Finally, to facilitate the review of immunostained slides
by pathologists and other scientists, either as panels in future work-
shops or as individual consultants, 3,3?-diaminobenzidine should be
used as a substrate of the detection system (rather than immunofluo-
rescence), and standard murine immunohistochemical protocols, as
are currently available at the Mouse Models of Human Cancer Con-
sortium website,18should be followed.
Discussion
Review of the available NF mouse models has illustrated that their
lesions have recapitulated, to various degrees, the human tumor coun-
terparts. These improved mouse models of NF now provide an in-
valuable platform for rigorous preclinical trials of innovative thera-
peutic approaches. Pathological validation of these tumors, an
essential step in the characterization of these models, is best achieved
by a panel, which provides the opportunity for pathologists with
different areas of expertise (neuropathology, soft tissue pathology,
EM, and veterinary pathology) to review and compare all available
models. This type of systematic and comprehensive evaluation is
valuable for formulating a classification with consistent and compre-
hensible terminology, as well as for providing a platform for commu-
nication between mouse modelers and pathologists, and for drafting
and disseminating recommendations for standardized specimen han-
dling and work-up procedures.
In the present review, some histological differences and similarities
were observed between murine and human PNSTs. For example, the
infiltrative growth pattern of GEM grade I schwannomas contrasts
with the well-circumscribed, pushing margin seen in human schwan-
nomas. On the other hand, some of the murine lesions showed close
histological similarities to the human counterparts (for example GEM
grade I neurofibromas and human neurofibromas and grade III GEM
PNSTs and human MPNSTs). The classification system proposed is a
descriptive, morphology-based classification that may evolve and
change as more information about the natural history of these murine
tumors becomes available. Also, as more NF mouse models emerge,
other histologically distinct lesions associated with specific genetic
alterations may become apparent. Eventually, the histology, biological
behavior, and genetic alterations could be correlated and integrated
into the GEM classification system and thus compared with the
genetics, biology, and histology of NF-associated human lesions and
corresponding sporadic human lesions. In this regard, it is recom-
mended that future central reviews be planned to validate and analyze
new models of GEM PNSTs, as well as other lesions in the NF mouse
models.
Acknowledgments
We are grateful to the following investigators for providing information and
materials for the slide sets: A. Berns; T. Jacks; L. Parada; E. Robanus-
Maandag; K. Shannon; and M. van der Valk. We thank our keynote speakers
A. Bernards and R. Fehon. We acknowledge and thank Dr. Judy Small, Jackie
Medina, Peter Bellermann, and the National Neurofibromatosis Foundation for
helping with the organization of the NF mouse modeling pathology meeting in
Boston in 2003.
Appendices
Appendix A: Summary of the NF Mouse Models Presented at the
Workshop (Table 3).
Appendix B: WHO Classification for Human PNSTs (Table 4).
Appendix C: Recommendations for Mouse Workup for PNSTs. Ini-
tially, until models have been well characterized, it is recommended that all
lesions be stained with H&E and S-100 immunostain and have EM analysis.
Immunohistochemical stains to rule out common tumors in specific sites have
been added (see Table 5).
References
1. Friedman JM, Gutmann DH, MacCollin M, Riccardi VM. Neurofibromatosis: phe-
notype, natural history, and pathogenesis, 3rd ed. Baltimore, MD: The Johns Hopkins
University Press; 1999.
2. Louis DN, Stemmer-Rachamimov AO, Wiestler OD. Neurofibromatosis type 2. In:
Kleihues P, Cavenee W. K., editors. Pathology and genetics of tumours of the nervous
system. Lyon, France: IARC; 2000. p. 219–22
3. Evans DG, Huson SM, Donnai D, et al. A clinical study of type 2 neurofibromatosis.
Q J Med 1992;84:603–18.
4. Rasmussen SA, Yang Q, Friedman JM. Mortality in neurofibromatosis 1: an analysis
using U. S. death certificates Am J Hum Genet 2001;68:1110–8.
5. Zoller M, Rembeck B, Akesson HO, Angervall L. Life expectancy, mortality and
prognostic factors in neurofibromatosis type 1. A twelve-year follow-up of an
epidemiological study in Goteborg, Sweden. Acta Derm Venereol 1995;75:136–40.
6. Perry A, Roth KA, Banerjee R, Fuller CE, Gutmann DH. NF1 deletions in S-100
protein-positive and negative cells of sporadic and neurofibromatosis 1 (NF1)-
associated plexiform neurofibromas and malignant peripheral nerve sheath tumors.
Am J Pathol 2001;159:57–61.
7. Baser ME, Evans DG, Gutmann DH. Neurofibromatosis 2. Curr Opin Neurol 2003;
16:27–33.
8. McClatchey AI, Saotome I, Ramesh V, Gusella JF, Jacks T. The Nf2 tumor suppres-
sor gene product is essential for extraembryonic development immediately prior to
gastrulation. Genes Dev 1997;11:1253–65.
9. Brannan CI, Perkins AS, Vogel KS, et al. Targeted disruption of the neurofibroma-
tosis type-1 gene leads to developmental abnormalities in heart and various neural
crest-derived tissues. Genes Dev 1994;8:1019–29.
10. Jacks T, Shih TS, Schmitt EM, et al. Tumour predisposition in mice heterozygous for
a targeted mutation in Nf1. Nat Genet 1994;7:353–61.
11. Lakkis MM, Epstein JA. Neurofibromin modulation of ras activity is required for
normal endocardial-mesenchymal transformation in the developing heart. Develop-
ment (Camb) 1998;125:4359–67.
12. Lakkis MM, Golden JA, O’Shea KS, Epstein JA. Neurofibromin deficiency in mice
causes exencephaly and is a modifier for Splotch neural tube defects. Dev Biol
1999;212:80–92.
13. McClatchey AI, Saotome I, Mercer K, et al. Mice heterozygous for a mutation at the
Nf2 tumor suppressor locus develop a range of highly metastatic tumors. Genes Dev
1998;12:1121–33.
14. Giovannini M, Robanus-Maandag E, van der Valk M, et al. Conditional biallelic Nf2
mutation in the mouse promotes manifestations of human neurofibromatosis type 2.
Genes Dev 2000;14:1617–30.
15. Bajenaru ML, Donahoe J, Corral T, et al. Neurofibromatosis 1 (NF1) heterozygosity
results in a cell-autonomous growth advantage for astrocytes. Glia 2001;33:314–23.
16. Gutmann DH, Loehr A, Zhang Y, et al. Haploinsufficiency for the neurofibromatosis
1 (NF1) tumor suppressor results in increased astrocyte proliferation. Oncogene
1999;18:4450–9.
17. Costa RM, Federov NB, Kogan JH, et al. Mechanism for the learning deficits in a
mouse model of neurofibromatosis type 1. Nature (Lond) 2002;415:526–30.
18. Silva AJ, Frankland PW, Marowitz Z, et al. A mouse model for the learning and
memory deficits associated with neurofibromatosis type I. Nat Genet 1997;15:281–4.
3723
18http://emice.nci.nih.gov/resources/protocols.
PATHOLOGY CLASSIFICATION OF PNSTS IN NF MOUSE MODELS

Page 7

19. Cichowski K, Shih TS, Schmitt E, et al. Mouse models of tumor development in
neurofibromatosis type 1. Science (Wash DC) 1999;286:2172–6.
20. Bajenaru ML, Zhu Y, Hedrick NM, et al. Astrocyte-specific inactivation of the
neurofibromatosis 1 gene (NF1) is insufficient for astrocytoma formation. Mol Cell
Biol 2002;22:5100–13.
21. Zhu Y, Ghosh P, Charnay P, Burns DK, Parada LF. Neurofibromas in NF1: Schwann
cell origin and role of tumor environment. Science (Wash DC) 2002;296:920–2.
22. Kalamarides M, Niwa-Kawakita M, Leblois H, et al. Nf2 gene inactivation in
arachnoidal cells is rate-limiting for meningioma development in the mouse. Genes
Dev 2002;16:1060–5.
23. Costa RM, Yang T, Huynh DP, et al. Learning deficits, but normal development
and tumor predisposition, in mice lacking exon 23a of Nf1. Nat Genet 2001;27:
399–405.
24. Giovannini M, Robanus-Maandag E, Niwa-Kawakita M, et al. Schwann cell hyper-
plasia and tumors in transgenic mice expressing a naturally occurring mutant NF2
protein. Genes Dev 1999;13:978–86.
25. Vogel KS, Klesse LJ, Velasco-Miguel S, et al. Mouse tumor model for neurofibro-
matosis type 1. Science (Wash DC) 1999;286:2176–9.
26. Reilly KM, Loisel DA, Bronson RT, McLaughlin ME, Jacks T. Nf1;Trp53 mutant
mice develop glioblastoma with evidence of strain-specific effects. Nat Genet 2000;
26:109–13.
27. Cardiff RD, Anver MR, Gusterson BA, et al. The mammary pathology of genetically
engineered mice: the consensus report and recommendations from the Annapolis
meeting. Oncogene 2000;19:968–88.
28. Weiss WA, Israel M, Cobbs C, et al. Neuropathology of genetically engineered mice:
consensus report and recommendations from an international forum. Oncogene 2002;
21:7453–63.
29. Kleihues P, Cavanee WK. Pathology and genetics. Tumors of the nervous system.
Lyon, France: IARC; 2000.
30. Scheithauer BW, Woodruff JM, Erlandson RA. Tumors of the peripheral nervous
system. Atlas of tumor pathology, vol. 24. Bethesda, MD: Armed Forces Institute of
Pathology; 1999.
3724
PATHOLOGY CLASSIFICATION OF PNSTS IN NF MOUSE MODELS