Journal of Histochemistry & Cytochemistry http://jhc.sagepub.com/

Contribution of immunohistochemistry to diagnostic problems of human cerebral tumors. L F Eng and L J Rubinstein J Histochem Cytochem 1978 26: 513 DOI: 10.1177/26.7.357640 The online version of this article can be found at: http://jhc.sagepub.com/content/26/7/513.citation

Published by: http://www.sagepublications.com

On behalf of:

Official Journal of The Histochemical Society

Additional services and information for Journal of Histochemistry & Cytochemistry can be found at: Email Alerts: http://jhc.sagepub.com/cgi/alerts Subscriptions: http://jhc.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

>> Version of Record - Jul 1, 1978 What is This?

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

0022- l554/78/2607-0513/$02.00/0 THE

JOURNAL

OF HISTOCHEMISTRY

© 1978 by The

Copyright

AND

Histochemical

CONTRIBUTION

OF

Inc.

IMMUNOHISTOCHEMISTRY

PROBLEMS

OF

LAWRENCE Department

of Pathology

HUMAN

F. ENG

(Neuropathology),

94305

and

Vol. 26, No. 7, pp. 513-522, 1978 Printed in USA.

CYTOCHEMISTRY Society,

Veterans

TO

CEREBRAL

AND

LUCIEN

Stanford Administration

J. RUBINSTEIN

University

School

Hospital,

(MS

DIAGNOSTIC

TUMORS’

Palo

of Medicine,

Alto,

Stanford,

California

94304

California

78- 142)

Immunocytochemical localization of the glial fibrillary acidic (GFA) protein (an astrocyte-associated protein) in formalin-fixed, paraffin-embedded tissue sections by the peroxidase-antiperoxidase method of Sternberger is currently being employed to assist in the diagnosis of human brain tumors. Positive reaction to GFA protein antibody has been demonstrated in the following tumors: astrocytomas, astroblastomas, ghioblastomas, ependymomas, subependymomas, and the astrocytic cells of mixed ghiomas. Negative reaction to GFA protein antibody is found in primitive neuroepithehial cells, ganghion cells, ohigodendroglia, choroid plexus epithehium, vascular endothehium, meningeal cells, fibroblasts, and other mesenchymal elements. The method combines a high degree of specificity with extreme sensitivity, which renders it superior to even the best results obtained so far with histologic stains specific to neuroghial fibers, such as Mallory’s phosphotungstic acid-hematoxyhin. It is proving of particular diagnostic importance in the following problems: 1) the demonstration of astrocytic cellular differentiation in either primitive or highly anaplastic CNS tumors; 2) the study of mixed tumors of the central nervous system, i.e. mixed ghiomas, mixed ghiomas and sarcomas, and gangliogliomas; 3) the demonstration of the ghial nature of tumors, such as astroblastomas, in which fibril formation is either scant or absent; 4) the diagnosis of ghiomas invading or metastasizing into the meninges and in extraneural sites; and 5) the exclusion of nonghial tumors that may superficially resemble astrocytomas. We believe that immunohistochemical determination of the GFA protein will soon be regarded as an essential tool for the more precise identification of diagnostically difficult brain tumors. The

development

oxidase

(PAP)

of the

peroxidase-antiper-

immunocytochemical

lipid

technique

(29), the demonstration of its applicability to formalin-fixed, paraffin-embedded tissue (30), and the availability of monospecific antiserum to purified in sensitivity pathology.

brain amide

proteins have offered new dimensions and specificity in diagnostic tissue A water-soluble protein which has

brains

by

aqueous

extraction

and

purified

by

the

Veterans

MRIS 2390, and National CA 11689 and NS 11632.

Administration

Institutes

of Health

The

migrates 41,000,

protein

isolated

as two bands and that from

47,000, as determined gel electrophoresis

with m. normal

by disc polyacrylin sodium dodecyl sul-

Isolation of the water-soluble normal brain by conventional sulfate fractionation,

iso-

electric precipitation, and column chromatography-has been hampered by its low solubiity and tendency to self-aggregate and to coaggregate with other acidic proteins in the initial

and 23, MS

crude extract. From normal only about 20% of the total

by

ammonium sulfate precipitation and isoelectnc precipitation. The GFA protein contains high contents of aspartic and gbutamic acids, alanine, and leucine; and negligible amounts of cysteine, ‘Supported

carbohydrate.

fate (SDS) (13,31). GFA protein from methods-ammonium

been designated the glial fibrillary acidic (GFA) protein has been isolated from human pathologic tissue rich in fibrous astrocytes (i.e., old plaques from multiple sclerosis (MS) brains) shown to be present in astrocytes (1, 13-19, 25, 28, 31). The protein is isolated from

and

from MS plaques w. of 47,000 and

nologic

activity

buffer,

pH

can

8. The

rat and mouse GFA protein

be extracted majority

with of the

can be solubiized with solutions kosyl, SDS, or 4 M urea (23). the

recently

published

for the water-soluble Dahi, and coworkers

grant

grants

properties

(2-9,

20,

513

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

33)

protein 8), and of such

phosphate

GFA

containing For this

purification GFA (3,

brain, immuprotein Sarreason,

procedures by Bignami, the unusual “native”

GFA

ENG

AND

of the

study

514 protein

must

of Liem that the 80-90%

be re-examined

in light

RUBINSTEIN dymal and the subependymal glial layers, foci of cerebral white matter gliosis, and areas of cortical cerebellar gliosis. In addition to the stains mentioned above, the Hoizer stain was also used as a special

and Shelanski (24) who have reported Dahl and Bignami preparations contain microtubular protein with small

amounts

of GFA

protein

copurifying.

made DahI

a similar and Bignami

conclusion data with

of the

microtubular

protein

by others Chemical,

We

staining

have

published

by us and

(14).

disc

immunological,

and

immunocyto-

chemical studies of the GFA protein, its presence in astrocytes as the major protein component of gbial filaments, and its distinction from lamentous protein have been reviewed

neurofirecently

(14). Immunohistochemical protein by Sternberger’s dase ded

localization of peroxidase-antiperoxi-

method on formalin-fixed, tissue sections has been

laboratory diagnosis

GFA

paraffin-embedreported from

our

(10, 11) and by others (12, 26) in the of human cerebral tumors. The tech-

nique is currently being used in our laboratory to help resolve some of the more difficult diagnostic problems that may beset their interpretation. system surgery review.

Our experience (CNS) tumor or at autopsy

to date material forms

MATERIALS

AND

Materials: We have munoperoxidase method mate total of 100 tumors paraffin-embedded

on central nervous obtained either at the basis of this

METHODS

to date examined by the imfor GFA protein an approxiinvolving the CNS, by using

tissue

obtained

and post mortem, and sectioned nesses varying from S to 7 micra. embedding material

and

processing

included

and

submitted

our

own

were

routine.

indifferently from

diagnostic

other services.

both

The

operation

at thickof fixation,

Our

tumors centers

at

routinely Methods

source

of

processed

or originating material

in from

examined

includes primary and metastatic neoplasms involving the CNS. Primary CNS tumors include both glial and nonglial neoplasms. The ghial tumors that have up until now been examined include astrocytomas, astroblastomas,

ghioblastomas,

ependymomas,

subepen-

dymomas, ohigodendroghiomas, ganghioghiomas, primary pineal parenchymal tumors, and a variety of mixed gliomas and mixed gliomas and sarcomas. In addition

to

routine

hematoxylin

and

eosin

stain,

spe-

included iron-hematoxylin and van Gieson, the Gordon-Sweets’ silver method for reticuhin fibers, and Mallory’s phosphotungstic acid-hematoxylin (PTAH) for ghial fibers. In addition, areas of both normal and ghiosed brain were repeatedly tested for control purposes. Such areas included the subpial cerebral margins, the epencial

stains

procedure

for

the

demonstration

of ghiosis.

Preparation of antiserum: Antiserum to GFA protein isolated from MS plaques was prepared as described previously (31). Antiserum to GFA protein which was purified by

by comparing the chemical properties

polyacrylamide

gel

electrophoresis

in

SDS

(31)

was prepared in the following manner. An electrophoresis separation consisted of 12 gel cylinders each loaded with 25 tg of purified GFA protein. The GFA protein band in one gel was visualized by immersion in 10% trichloroacetic acid (TCA). The remaining 11 gels were lined up with the TCA-treated gel, and the corresponding GFA protein bands from the untreated gels were sliced and pooled. GFA protein bands from 33 gels had a volume of 7 ml and contained about 800 tg GFA protein. The pooled gel bands were emulsified with 7 ml of incomplete Freund’s adjuvant (Difco Lab. Detroit, Mich.) in a glass homogenizer with a motorized teflon pestle. Ten milliters of this emulsion were further homogenized with 50 mg of H37Ra mycobacterium (Difco Lab.), and 5-nil aliquots were used for the first and second inoculation. The remaining emulsion (4 ml) without the mycobacterium was used in the third and fourth inoculation. All inoculations were done with multiple injections on the back and the sides and spaced at 7-day intervals. Forty five days after the first inoculation, high titer antibody to human GFA protein was present in the rabbit serum. Antiserum to GFA protein from bovine spinal cord was prepared in the following manner. An immunoprecipitate containing bovine GFA protein was prepared by incubating rabbit antiserum to human MS plaque GFA protein with a crude GFA protein fraction prepared from frozen bovine spinal cord by extraction with 0.05 M sodium phosphate buffer, pH 8, containing 4 M urea. The immunoprecipitate was resuspended in 0.9% saline and pelleted by centrifugation twice before use. A 2-nil emulsion composed of immunoprecipitate containing approximately 100 g GFA protein, 1 ml 0.9% saline, 1 ml of incomplete Freund’s adjuvant, and 5 mg H37Ra mycobacterium was prepared by homogenization in a glass homogenizer with a motorized teflon pestle and used for the first and second inoculations. A 2-mI emulsion composed of immunoprecipitate containing approximately 50 tg GFA protein, 1 ml 0.9% saline, and 1 ml of incomplete Freund’s adjuvant was used for the third and fourth inoculations. For the first inoculation, 100-1d aliquots were injected into 2 places in each footpad and the remaining 1.2 ml were injected into multiple sites on the back and sides. Seven days later, the second inoculation was identical to the first. Fourteen days after the initial inoculation, 2 ml of emulsion were inoculated in multiple sites on the back and sides. Twenty-one days after the initial injection, the fourth inoculation was identical to the

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

IMMUNOHISTOCHEMISTRY

AND

third. Forty-five days after the initial injection, high titer antibody specific to bovine GFA protein was present in the rabbit serum. The monospecificity of each antiserum was demonstrated by immunodiffusion, immunoelectrophoresis,

and

electrophoresis

of immunoprecipitates

(31).

We prefer to use antiserum to bovine GFA protein for staining human tissue. Antiserum to human GFA protein must be adsorbed with normal human serum or IgG before use on human tissue. Peroxidase-antiperoxidase (PAP) method: The PAP technique (29) as modified by Taylor and Burns (30) and described in detail by Deck et al. (11) was employed in this study. Sections were cut at 5-7 i, deparaffinized with xylol and treated with 0.3% hydrogen peroxide in anhydrous methanol for 30 mm to block endogenous peroxidase activity. Sections were sequentially incubated at ambient temperature, the sections being washed with phosphate buffered saline (PBS) between each incubation. Incubations were performed with: (a) normal swine serum, diluted 1:20 in PBS, for 10 mm; (b) rabbit antiserum to GFA protein, diluted 1:50 or up to 1:500 in PBS for 45 mm; (c) commercial swine anti-rabbit IgG (Dako, Copenhagen) diluted 1:20 for 30 mm; (d) soluble PAP (Dako) diluted 1:50 in PBS for 30 mm. Sections were then rinsed in Tris buffer, pH 7.6, and stained with freshly prepared 3-3 diamino benzidine tetrahydrochloride (DAB) (Sigma Chemical Co., St. Louis, Mo.) for S minutes. (20 mg DAB/100 ml Tris buffer containing 0.OOS% hydrogen peroxide.) After a thorough rinse in distilled water, the reaction product was enhanced by further treatment with 1% aqueous osmium tetroxide for 1 mm. In a number of cases, the immunoperoxidase staining was followed by light counterstaining

tional

procedure

stain

and

with

did not impair

in a number

of

hematoxylin.

This

instances

even

seemed

enhance it. Considerable advantage was gained additional step in permitting a more accurate fication of normal and neoplastic cells. Control sections were prepared by using preimmunization

rabbit

serum

addi-

the immunoperoxidase

diluted

1:20

anti-GFA protein serum which had been with purified GFA protein. These control showed no specific reaction products for

to

by this identi-

or

adsorbed sections the GFA

protein. RESULTS

Normal

and Reactive neoplastic

Fibrillary Glia Human Material

in Non-

The immunoperoxidase stain for GFA protein strongly and selectively stained the fibrillary glia along the marginal subpial layer of the cerebral cortex and in the subependymal layer. The perikarya of scattered subpial fibrifiary astrocytes and strated

of

the more

subependymal abundant

astrocytes and

more

strongly

demonstain-

CEREBRAL

ing focal staining cells.

cells

deposits,

had

TUMORS with

of the However, acquired

and

negative-to-weak

processes, inflammaor obstructive

reaction for GFA proboth in the cyto-

processes

is in

ependymal ependymal

fibrillated

to neighboring (27, p. 204),

a positive demonstrated

in the

observation

515

of normal in which the

tapering

hydrocephalus, tein was easily plasm

usually

cytoplasm in areas

usually as a reaction tion, tumor invasion

of these

agreement

cells.

with

capacity of normal, reactive dymal cells to develop glial (27, p. 210; 32), and therefore

This

the

known

and neoplastic epenfilaments and fibers confirms the find-

ing previously demonstrated by immunofluorescence (see previous references) that, in appropriate

circumstances,

astrocytes

are

ependymal capable

cells

as well

of synthesizing

the

as pro-

tein. In the peroxidase tological molecular

gbiosed

cerebellar

stain structure layer,

demonstrated the typical hisof fine radial gbiosis in the with the presence of GFA pro-

cortex,

the

immuno-

tein in the Bergmann gbia and in large reactive astrocytes scattered throughout the granular layer. In the white matter of cerebral hemispheres adjacent to a tumor, in areas of inflammation astrocytes

and

of cerebral demonstrated

edema, large an intensely

reactive positive

reaction. Oligodendroglial cells, microglia, neurons, choroid plexus epithelium, vascular endothehum, meningeal cells, fibroblasts, and blood cells showed a negative reaction to GFA protein antibody. A comparison with the Holzer fibrillary demonstrated

either

in PBS

HUMAN

closely white

of

of the immunoperoxidase stain for gbia in areas

gbiosis

is of interest. The Holzer intensely positive staining

packed and

matter

reactive

stain of cortical

intersecting much

astrocytes

more

ghial fibers discrete

scattered

Immunoperoxidase

produced

in

in the staining

the

a much

stain of the

cortex.

more

del-

icate fibrillary network ter and a considerably the perikarya and cell

of ghia in the white matmore intense staining of processes of reactive as-

trocytes in the lated astrocytes

The footplates of fibrilon the capillary blood

vessels

were

cortex. abutting

often

well

an almost complete wall. Astrocytic and Tumor The combined

immunoperoxidase a high

demonstrated,

ring other

material stain

degree

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

forming

around the vascular nuclei were unstained.

for GFA

of specificity

protein with

ex-

ENG

516 treme some

AND

sensitivity which rendered it equal, cases superior to, some of the best

RUBINSTEIN

and in results

obtained with Mallory’s phosphotungstic acidhematoxyhin. A positive reaction to GFA protein antibody was demonstrated in the following tumors: neoplastic cells of astrocytomas, astroblastomas, ependymomas, subependymomas, the astrocytic components of mixed gliomas, ganghioghiomas, mixed gliomas and sarcomas, and the more differentiated astrocytic components of ghioblastomas mor cells

and medullobhastomas. in oligodendrogliomas, and

sarcomas

metastatic

tumors

negative. The stain has so far value in five diagnostic problems: 1. of

The

astrocytic

primary

CNS

The meningiomas,

nature

were

invariably

proven

to

or

tumors

the cells

more primitive unstained. The

are

either

tomas pending

but

the

showed upon

more

peripheral

a putative tion of the

astrocytic origin medubloblastoma

that small for the

of glioblas-

(Fig. either

2). The neoplastic

interpretation or reactive

ever,

a diagnostic

problem

to demonstrate cases focal stainindicating either or focal towards

cells

contrasts

with

further

published

the

completely

negative

fibril formation Ependymomas,

tumors scant

or

gliofibrillogenesis

intensely

dase,

in which absent: may

positive

as reported

be

(Fig. 4). nature is in

demonstrable

by light and electron microsmay in some cases show an staining with immunoperoxiin greater detail elsewhere (11). a rare glioma characterby a papillary arrangement of

the tumor cells around central vessels and in which neuroglial

rarely

by

of intensely elements

of CNS either

demonstrable

with

capillary are

fibrils Mallory’s

blood usually

PTAH

(27,

p. 193), frequently demonstrated an intensely positive staining of the tumor cells around the blood vessels (Fig. 5) and in the pericapillary papillary formations (Fig. 6). Thus, in instances in which an alternative diagnostic interpretation, such as metastatic papillary meningioma, basically

astrocytic

problem gist may

nature

of

the

tumor

of gliomas invading, the leptomeninges

diagnosis

into, sites:

with which be confronted

is situated outside

carcinoma or malignant had been entertained, the was

established.

4. The

ref-

those

component of the glial

metastasizing in extraneural

as how-

CNS:

reaction of the sarcomatous 3. The demonstration

convincingly

differentiaastrocytes

of these constituted, to which

confirming

with demonstration of the gliomatous

The astroblastoma, ized histologically

tiation (Fig. 1). The endothelial cells of vascular endothelial proliferation were invariably negative. Medulloblastomas, which are capable of differentiation towards more mature cells, either glial or neuronal (27, pp. 248-249), but in which difficult in some thereby

fully

in the tumor cells copy (27, p. 210),

various degrees of staining dethe extent of astrocytic differen-

this capacity is often convincingly, showed ing of the tumor cells,

results

Duffy et al. (12), positive reaction

which

or more anapoorly differ-

areas

our

astrocytic

astrocytocould easthe advan-

of the

oligodendrogliomas (mixed gliomas) (Fig. 3), in subependymomas, and in mixed neuronal and glial tumors (gangliogliomas). Several mixed gliomas and sarcomas have also been examined,

or highly

entiated fusiform cells in gliobbastomas form the typical pseudopalisades around foci of early necrosis showed no affinity stain,

of

tumors

The immunoperoxidase stain has provided considerable help in the identification of neoplastic astrocytic elements in mixed astrocytomas and

which

be

composition

which

extremely primitive (embryonal) anaplastic: The more differentiated elements of glioblastomas, malignant mas and cerebellar medulloblastomas ily be identified, the stain possessing tage of heaving plastic tumor

tu-

erence will be made below. 2. The study of mixed

outside the

the its

parenchyma

or and

An important diagnostic neurosurgical patholomay arise when a glioma

normal of the

i.e.,

environment,

brain

or

spinal

Fo;. 1. Glioblastoma multiforme. GFAP-immunoperoxidase (1:100 antiserum dilution) and hematoxylin. a, Area of pseudopalisading (lower left). Occasional immunoperoxidase positive tumor cells (arrows); most of the tumor cells are negative. x300. b, Area from same tumor, away from pseudopahisading. Several tumor cells are immunoperoxidase positive (arrows). x480. Fufl. 2. Cerebellar medulloblastoma. a, Hematoxylin and eosin. The tumor demonstrates high cellularity and poor differentiation with this stain. x300. b, GFAP-immunoperoxidase (1:250 antiserum dilution). Isolated tumor cells in same tumor demonstrating Most tumor cells are unstained. x380. FIG. 3. Mixed oligodendroglioma and

oligodendroglioma.

Note

absence

positive

immunoperoxidase,

astrocytoma.

of neuroglial

(1:50 antiserum dilution) and hematoxylin. astrocytic participation in the tumor. x480.

fibrils Many

indicative

of astrocytic

differentiation.

a, Mallory’s PTAH. Cytological features suggestive of in this preparation. x380. b, GFAP-immunoperoxidase tumor cells are immunoperoxidase positive, indicating

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

.‘

.,

,f”#{149}

#{149} :-‘‘

-1

;“-c

2b’” ,-

/

I.

4j _I

:

517

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

#{149})t 4.’

518

ENG

AND

RUBINSTEIN

FIG. 4. Mixed sarcoma and glioma. a, Iron-hematoxyhin highly cellular areas of poorly differentiated sarcoma, b, GFAP-immunoperoxidase (1:50 antiserum dilution). positive. The sarcomatous areas are unstained. x240.

cord.

The

diagnostic

often compounded the tumor cells,

problem

in

such

cases

by the anaplastic nature as demonstrated in routine

is of his-

tological stains, and by the well-known capacity on the part of gliomas to stimulate an abundant fibrous connective tissue response once they invade the leptomeninges or other extraneural sites (27, pp. diagnosis with meningioma entiated The

or sarcoma, carcinoma

will

immunoperoxidase

us to confirm of primary

336-337). Thus, a mesenchymal

the glial

the differential neoplasm, either

or with

neoplasms

differ-

nature in which

has

mode

of

as was

“meningocerebral recently completed

xanin

by

this

relative

from nature without

gliomas of

the doubt.

John J. report tech-

us to establish the glial nature of which group some members been regarded as representing

meningocerebral fibrous 111). The immunoperoxidase

the

growth was largely leptomeningeal (Fig. 7) and to which the interpretation of malignant meningiomas, sarcomas, atypical fibroxanthomas, and xanthosarcomas had previously been applied. A study of 10 cases of a distinctive form of supratentorial astrocytoma occurring in young subjects, designated thoastrocytomas”,

nique has led these tumors, had previously

gained

permitted

of a number the

our laboratory in collaboration with Dr. Kepes, and will be the subject of a fuller elsewhere (22). The immunoperoxidase

useful in confirming the tastasizing to extraneural

arise. stain

gliomatous

a poorly

and van Gieson stain, demonstrating darkly staining, adjacent to paler fibrillated areas of atypical glia. x200. The gliomatous areas are strongly immunoperoxidase

technique

xanthomas stain

(21, has

identity sites.

of

makes

extraneural

it mandatory

remote deposits An example

27,

also

p.

been

of gliomas meThe information

is important

rarity

of

because metastases

that

the

glial

be established of myxopapillary

ependymoma arising from the region of the cauda equina and metastasizing to the lungs and the lymph nodes has been described and illustrated elsewhere by Deck et al. (11). The exclusion of nonglial tumors that may superficially resemble astrocytomas

5.

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

IMMUNOHISTOCHEMISTRY

AND

HUMAN

CEREBRAL

TUMORS

519

r ‘

p

r-

-. -V

.4

I

*

.

‘

ob

e

S. Astroblastoma, demonstrating immunoperoxidase positive perikarya and cell processes of tumor cells directed towards a central blood vessel. Negative reaction of endothehium and of red blood cells. GFAPimmunoperoxidase (1:50 antiserum dilution). x380. FIG. 6. Astroblastoma. a, Mallory’s PTAH stain, demonstrating perivascular papillary pattern of tumor cells. Note absence of neuroglial fibers in this preparation. x240. b: GFAP immunoperoxidase (1:100 antiserum dilution). Several tumor cells, especially those lining papillary formations, show positive immunoperoxidase reaction in their cytoplasm. x240. FIG.

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

4

I, . *,

h

..

#{149}5

FIG. 8. Island of fibrillated neuroglia included recurrent malignant meningioma. GFAP-immunoperoxidase (1:50 antiserum dilution) and hematoxylin. x240.

in a

FIG. 7. Astrocytoma localized to the leptomeninges. a, Hematoxylin and eosin stain. Note absence of tumor in ghiotic subpial layer. Elongated, rather poorly differentiated cells in leptomeninges. x240. b, Serial section from the same field, stained with Gordon-Sweets’ silver method for reticulin fibers. Abundant reticuhin

fibers intersecting the tumor in the leptomeninges, raising the possibility of a mesenchymal tumor. x240. c, GFAP-immunoperoxidase (1:250 antiserum dilution) field. plasm fusely head).

and hematoxylin. Serial section from the same Positive immunoperoxidase reaction in cytoand processes (arrows) of tumor cells, and difpositive reaction in gliotic subpial layer (arrowx380.

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

IMMUNOHISTOCHEMISTRY in

routine

histological

tastatic nonglial dase

tumors origin reaction,

useful

the

adjunct

these

stains:

and show

in cases

at the

periphery,

more central and invasive

as

cases.

a

of the of diag-

Thus

a

small

with

the

stain

react

frequently

within

the

portions, of metastatic carcinomas mesenchymal neoplasms, such

malignant These cells

meningiomas are interpreted

that are still present While the problem

and sarcomas as reactive

son between the results noperoxidase technique cial neurohistological has

challenge a careful

that may compari-

satisfactorily. CONCLUSION Whereas

the

not,

evidently,

will

problems man brain tive

resolve

concerning tumors,

and

anaplastic

selectivity

and

powerful

tool,

is likely

precise central

immunoperoxidase

to

our

diagnostic

identification neuraxis.

that

its

it a most discrimination,

make

if used almost

of huprimi-

it is apparent

sensitivity

which, prove

of

the interpretation especially those more ones,

its

technique all

with

essential

of difficult

for

the

tumors

more

of the

ACKNOWLEDGMENTS We the

Home

thank

Miss

Cheryl

immunoperoxidase

for

photomicroscopic LITERATURE

Padula

for

procedures

performing and

Mr.

Phil

assistance.

9. Dahl

Bignami

A: Glial

521

fibrillary

acidic

protein

human brain. Purification Res 57:343, 1973

D, Bignami

D,

sulfate fibrilary 1977

on

JH,

of ghioma

A: Glial

Bignami

and

prop-

of the human

fibrillary

ropathol

A: Effect

the immunogenic acidic protein.

Eng

glial

brains.

acidic

method

J

protein

dodecyl

of the glial Meth 17:201,

J: A preliminary

using

the

for the

GFA

study

peroxidase-anti-

protein.

J Neu-

35:362, 1976 11. Deck JHN, Eng LF, Bigbee J, Woodcock SM: The role of glial fibnillary acidic protein in the diagnosis of central nervous system tumours. Acta Neuro-

pathol 12. Duffy

Exp

of sodium

properties J Immunol

LF, Bigbee

morphology

peroxidase

Neurol

(Berl), in press PE, Graf L, Rapport

MM:

Identification

of

glial fibnillary acidic protein by the immunoperoxidase method in human brain tumors. J Neuropathol Exp Neurol 36:645, 1977 13. Eng LF: Chemical characterization of the glial fibnillary acidic protein. Fed Proc 32:485, 1973 14. Eng LF, Bigbee JW: Immunohistochemistry of nervous system-specific antigens. In Advances in Neurochemistry, Vol. 3. Edited by BW Agranoff and MH Aprison. Plenum Press, New York, in press 15. Eng LF, Gerstl B, Vanderhaeghen JJ: A study of proteins in old multiple sclerosis plaques. Trans Am Soc Neurochem 1:42, 1970 16. Eng LF, Kosek JC: Light and electron microscopic localization of the glial fibrillary acidic protein and S-100 protein by immunoenzymatic techniques. Trans Am Soc Neurochem 5:160, 1974 17. Eng LF, Lee YL, Miles LEM: Measurement of ghial fibnillary acidic protein by a two-site immunoradiometric assay. Anal Biochem 71:243, 1976 18. Eng LF, Uyeda CT: Immunologic specificity of the water-soluble and -insoluble glial fibnillary acidic protein. Trans Am Soc Neurochem 4:79, 1973 19. Eng LF, Vanderhaeghen JJ, Bignami A, Gerstl B:

An acidic protein Brain Res 28:351,

CITED

1. Bignami A, Eng LF, Dahi D, Uyeda CT: Localization of the glial fibrillary acidic protein in astrocytes by immunofluorescence. Brain Res 43:429, 1972 2. Chan PH, Huston JS, Dahl D, Bignami A: Purification and initial characterization of astroglial protein from bovine brain. Fed Proc 34:224, 1975 3. Dahl D: Glial fibrillary acidic protein from bovine and rat brain. Degradation in tissues and homogenates. Biochim Biophys Acta 420:142, 1976 4. Dahl D: Isolation and initial characterization of glial fibrillary acidic protein from chicken, turtle, frog, and fish central nervous system. Biochim Biophys Acta 446:41, 1976

D,

normal Brain

TUMORS

from normal and ghiosed human brain. Demonstration of multiple related polypeptides. Biochim Biophys Acta 386:41, 197S 8. Dahl D, Bignami A: Immunogenic properties of the ghial fibrillary acidic protein. Brain Res 116:150, 1976

10. Deck

obtained by the immuand by routine and spestains of adjacent serial permitted us to resolve it

usually

7. Dahi

8).

brain. consti-

CEREBRAL

6. Dahl D, Bignami A: Heterogeneity fibrillary acidic protein in ghiosed Neurol Sci 23:551, 1974

as

(Fig. astrocytes

within the invaded may, in some cases,

tute the major interpretative be offered by this technique,

sections

employed

and of sensitivity raise difficult

will strongly and even

of cells

from erties.

of

the nonglial nature of doubt. However,

high degree of specificity the stain may occasionally nostic problems in such number

be

HUMAN S. Dahl

all me-

CNS tumors immunoperoxi-

can

to establish

neoplasms

Since

all primary a negative stain

AND

isolated 1971

from

fibrous

astrocytes.

20.

Huston JS, Bignami A: Structural properties of the ghial fibnillary acidic evidence for intermolecular disulfide bonds. Biochim Biophys Acta 493:93, 1977 21 Kepes JJ, Kepes M, Slowik F: Fibrous xanthomas and xanthosarcomas of the meninges and the brain. Acta Neuropathologica (Berl) 23:187, 1973 22. Kepes JJ, Rubinstein U, Eng LF: Meningocere-

bral young

xanthoastrocytoma. subjects,

A distinctive

presumably

pial astrocytes, with relatively A study of ten cases. Abstract national ton, D.C.,

Congress September

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014

originating

favorable

to the of Neuropathology, 1978

ghioma from

in sub-

prognosis. VIlIth InterWashing-

ENG

522 23. Lee

YL,

Eng

LF,

Miles

immunologic identity GFA protein. Trans 1976

24. Liem

RKH,

LEM:

of “soluble” Am Soc

Extraction

AND

RUBINSTEIN

and

and “insoluble” Neurochem 7:240,

ML: Identity of the major glial fibnillary acidic protein preparations with tubulin. Brain Res, in press 25. Ludwin SK, Kosek JC, Eng LF: The topographical distribution of S-100 and GFA proteins in the adult rat brain: An immunohistochemical study using horseradish peroxidase-labeled antibodies. J Comp Neur 16S:197, 1976 protein

Shelanski

in “native”

26. Maunoury

R, Delpech

A, Delpech

B, Vidard

MN,

Vedrenne C, Constans JP, Hillereau J: Localisation de la prot#{233}inegliofibnillaire (GFAP) par immunocytochimie dans les tumeurs c#{233}r#{233}brales humaines.

Neuro-Chirurgie

27. Russell of the

nold,

DS, Rubinstein Nervous

London

28. Schachner

System,

23:173,

1977

IA: Pathology 4th edition,

of Tumours Edward Ar-

pp. 1-448, M,

1977 Hedley-Whyte

ET,

Hsu

DW,

Schoonmaker G, Bignami A: Ultrastructural localization of glial fibnillary acidic protein in mouse cerebellum by immunoperoxidase labeling. J Cell

Biol 75:67, 1977 29. Sternberger LA, Hardy PH, Cuculis JJ, Meyer HG: The unlabelled antibody enzyme method of immunohistochemistry. Preparation and properties of soluble antigen-antibody complexes (horseradish peroxidase-antihorseradiah peroxidase) and its use in identification of spirochetes. J Histochem Cytochem 18:315, 1970 30. Taylor CE, Burns J: The demonstration of plasma cells and other immunoglobulin containing cells in formahin fixed paraffm embedded tissues using peroxidase-labeled antibody. J Chin Pathol 27:14, 1974 31. Uyeda CT, Eng LF, Bignami A: Immunological study of the ghial fibrillary acidic protein. Brain Res

37:81,

1972

32. Vraa-Jensen J, Herman MM, Rubinstein IA, Bignami A: In vitro characteristics of a fourth ventricle ependymoma maintained in organ culture systems. Light and electron microscopic observations. Neuropathol Applied Neurobiol 2:349, 1976 33. Yen SH, DahI D, Schachner M, Shelanski ML: Biochemistry of the filaments of brain. Proc Nat Acad Sci (USA) 73:529, 1976

Downloaded from jhc.sagepub.com at Univ Politecnica Madrid on April 24, 2014