. 44
( 45 .)


˜Nat1 Acad Scl USA 92,3923-3927.
54 Bessen, R A, Kocrsko, D A, Raymond, G J , Nandan, S , Lansbury, P T , Jr.,
and Caughey, B (1995) Nongenetic propagatron of stram-specific phenotypes of
scrapre prton protem. Nature 375, 698-700.
55 Bolton, D C , Bendhelm, P E., Marmostein, A D , and Potempska, A (1987)
Isolatron and structural studies of the intact scrapte agent protem. Arch Blochem
Blophys 258,579-590.
56 Caughey, B. W., Dong, A., Bhat, K. S., Ernst, D., Hayes, S. F , and Caughey, W S
(199 1) Secondary structure analysis of the scrapre-assocrated protein PrP 27-30
m water by infrared spectroscopy Blochemzstry 30, 7672-7680
lmmunohistochemistry of Resinated Tissues
for Light and Electron Microscopy
Martin Jeffrey and Caroline M. Goodsir

1. Introduction
7.1. Use of Ultrastructural lmmunocytochemistry
in Investigations of the CNS
In spite of the widespread use of tmmunocytochemtcal methods and the com-
mensurate increase m the range of available techniques, a consistent means of
electron mtcroscoptcal antigen locahzatton m the lipid-rich central nervous
system (CNS) has yet to be established. However, some successeshave been
achieved, most notably in neuroanatomy studies Combined morphologtcal and
htstochemtcal techniques for the study of neuronal mtcroctrcutts have been
described (1). Many of the antigens mvesttgated in the CNS, such as neu-
rotransmitters, appear not to withstand conventtonal electron mrcroscoptcal
fixatron and embedding methods and consequently pre-embedding methods
are employed (2). Pre-embedding rmmunocytochemtstry unfortunately suffers
from the disadvantages of unpredictable antibody penetration and subopttmal
morphologtcal structure.
1.2. Application of Ultrastructural lmmunocytochemistry
to the Transmissible Spongiform Encephalopa thies
Relatively few ultrastructural immunocytochemical studies have been per-
formed on the transmtsstble spongiform encephalopathtes. Immunogold stud-
ies of PrP locahzatton performed on gluteraldehyde-fixed, plastic-embedded
tissue confirmed the presence of PrP withm indtvtdual amyloid filaments
located m the subependyma and subpta of scrapte-infected hamsters (3). Simi-
lar studies have been performed on the tissues of human patients affected with

From Methods m Molecular Medmne Prlon Diseases
Edlted by H Baker and A M Rtdley Humana Press Inc , Totowa, NJ

302 Jeffrey and Goodsir

Gerstmann-Straussler-Schemkerdisease(GSS) or with Creutzfeldt-Jakob disease
(CJD). Although these studies were again able to show that fibrtls of amylord
plaques are tmmunoreacttve to PrP, the rather poor tissue quality obtamed from
tmmerston-fixed autopsy material IS mtmtcal to more detailed investigattons.
Ultrastructural pre-embedding immunolocahzatton studtes have been per-
formed on scrapte-infected cells mamtamed m vttro (4.5). PrP accumulates
wtthm cytoplasmlc vesicles of infected cells and 1salso released mto the cul-
ture medium Btochemtcal studies of the localization of the transformation and
accumulation of PrP m scrapte-infected cells mamtamed m vrtro also show
that the transformation occurs at the plasma membrane or along an endocytotic
pathway to lysosomes (6,7). Immunogold electron mtcroscopy has also been
very effectively employed to show accumulation of ubtqumn-protein conJu-
gates m lysosomes of scrapte-infected mouse brains (8)
Our own studies have concentrated on postembeddmg tmmunocytochemi-
cal mvesttgattons of PrP localizatton m scrapre-infected mace. Good localrza-
tton of PrP has been achieved on specimens prepared for routme electron
mtcroscopy and the comments outlmed herem are largely confined to our detec-
tion of anttgen m aldehyde-fixed tissues of murme scrapre.
Immunoelectron mtcroscopy 1slikely to prove helpful m other CNS disease
mvestigattons. The PA4 protein of Alzhetmer™s disease is similar to PrP m that
both are cell membrane proteins that are released from neurons and form extra-
cellular amylotd fibrtls Immunogold studies of PA4 protein localization and
aggregation, parttcularly m the new mouse model of the disease, are likely to
further elucidate the relationship of amyloid fibril formatron and neural deli-
ctts m Alzheimer™s disease. Stmilarly, antibodies raised to vnus protems, or to
particular bacterial antigens, may help to clarify the pathogenesis of infectious
CNS disorders
2. Materials and Methods
2. I. Theory and Practice
of Ultrastructural lmmunocytochemical Staining for PrP
2.7.1. Tissues and Experimental Design.
Nonspecific Binding and Background Deposits
One of the mam prerequisites for light microscopic rmmunocytochemical stud-
ies, that some form of antiserum and tissue control ISessential, also applies to the
tmmunocytochemtstry of resmated sections and for electron mtcroscopy. This 1s
particularly important in studies of PrP, becausenone of the available anttbodtes
IS able to distmgmsh between normal (cell-associated protease or phospholipase-
senstttve) and disease-specific (protease-resistant) PrP, nor have anttbodtes yet
been raised to hypothetical disease-specific conformational changes of PrP.
lmmunohrsfochemrstry of Resinafed Tissues 303

The lmmunogold method is notoriously caprlclous and a low frequency of
nonspeclfic bmding of randomly distributed gold particles as well as occa-
sional clusters of gold partxles are sometimes found throughout the sectlon
(and also on plastic adJacent to the tissue specimen) even in the “cleanest”
preparations. To the unwary these may convey spurious specificity. Such clus-
ters may be.
1 Causedby imperfections in the specimen,
2. ReactIonsbetweenthe etching agentsand the fabric of the grid,
3. Nonspecific bmding to aldehydes,and
4. Other reasonsthat currently are unknown
Therefore, It is particularly important that ultrastructural studies of PrP
locahzatlon are correlated with Individual structures identified m Immunocy-
tochemically stained “thick” (1 pm) sections viewed at the light microscope
level. Accumulations of PrP may then be clearly associated with the dlsease-
specific patterns of PrP recognized m the particular disease under investigation
It is of particular value to deal with infectlons that give rise to characteristic or
mvarlant lesion patterns, such as 1sfound m several of the murme scraple mod-
els (9) and in bovine spongiform encephalopathy @SE).
Although disease-specific PrP accumulations are detected within the cyto-
plasm of infected neurons maintained in vitro, disease-specific PrP accumula-
tlons m vlvo are generally found within the neuropll, although intracytoplasmlc
PrP is reported m several transmissible spongiform encephalopathies. Some
mtracytoplasmic PrP within neurons is unassociated with disease since It may
be detected at specific neuroanatomic sites of both scraple-inoculated mice
and m controls (9). In cattle, mtracytoplasmlc PrP crossreacts with cerold
hpofuscin of both ME-affected animals and m cattle with other types of neu-
rological disease and in controls (10). However, intracytoplasmic PrP may also
be found in human patients with CJD and m cats with feline sponglform
encephalopathy (FSE). It seems probable that at least some of the PrP found
withm neurons m these latter diseasesmay represent disease-specific accumu-
latlon, but the subcellular sites of mtraneuronal disease-specific PrP accumula-
tion in vlvo have not yet been determined.
2.1.2. Technical Comments TISSUE SOURCE
1. The embedding and/or the etching processesused for lmmunocytochemlstry
causesomedeterioration m specimen quality comparedwith methodsdeveloped
for optimal preservation of tissue used in morphological studies Therefore,
immunolocahzatlon studies should, where possible, be performed on tissues
obtained from perfused animals.
304 Jeffrey and Goodsir

2 Ultrastructural mununolabelmg of mature amylold fib& IS usually possible on
Immersion-fixed tissues obtamed at biopsy or at necropsy
1 Many antlgemc epitopes for PrP are preserved in formalm, penodate/lysme/
paraformaldehyde and gluteraldehyde/paraformaldehyde fixatives However,
some antibodies that may recogmze epltopes m cryostat or parrafin wax-embed-
ded tissues may not recogmze the same epltope m tissues treated with harsher
fixation or embeddmg protocols (see the followmg)
2 Optimal tissue structure for electron microscopy 1s obtamed from fixation m
gluteraldehyde/paraformaldehyde fixatives. Although not appropriate for all sub-
cellular studies, most of our mvestlgatlons have been performed on tissues fixed
in mixed aldehydes
Araldlte, Durcopan, and related hard epoxy plastics are the embedding
medium of choice for electron microscopy. However, as previously stated,
these resins are lmpervlous to water and have to be etched prior to Immu-
nolabelmg. Other water-soluble resins, such as LR White and Lowcryl, may
also be employed for Immunolabelmg. However, tissues embedded m water-
soluble resins give inferior morphology and, more important, subsequent treat-
ments with antigen-enhancing reagents causesthe section to lift from the slide
and float off. Where antigens are not exposed or are adulterated with aldehyde
fixatives or with plastic embedding procedures, It is preferable to employ a
pre-embedding regimen rather than to employ one of the water-soluble resins.
Several methods for etching epoxy plastics are avallable:
1 Hydrogen peroxlde (H202, an oxldizmg agent) The effect of this reagent on resm
1sto oxidize the hydrophobic alkane side chains of alcohols, aldehydes, and acids
This has a net effect of Increasing the hydrophillclty of the resin. At LM level,
when used on Its own, this reagent gave much poorer labeling than other etching
reagents. At EM level, for mcubatlon periods of 6 mm, this reagent produced no
slgmficantly enhanced nnmunostammg of PrP Only weak labeling of plaques
was achieved. Myelm became bleached, although other tissue structures were
well preserved
2 Sodmm ethoxlde (removes resm) particularly good for l-pm thick resmated sec-
tions (Fig. I), but much too harsh for electron microscopy studies It is very dlf-
ficult to keep sections on grids followmg sodium ethoxide etchmg
3 Potassmm methoxldelg crown 6/Aq. DMSO (removes resm). produces etchmg
of l-pm thick sections, which 1s almost as good as that of sodium ethoxlde It
also combines good etchmg with maintenance of good tissue structure for elec-
tron microscopy Thts IS the only reagent that may be used m vn-tually identical
protocols for both light and electron mlcroscopy studies of PrP locahzatlon
lmmunohistochemistry of Resinated Tissues

Fig. 1. Araldite-resinated
1-urn thick section, etched with sodium ethoxide, pretreated
with formic acid, and immunostained by the peroxidase antiperoxidase method using
the 1B3 antimurine SAF antibody. The section shows a PrP-releasing neuron from the
lateral hypothalamus of a mouse infected with the 87V murine scrapie strain.

4. Sodium metaperiodate (a strong oxidizing agent): unmasks protein antigenic sites
on gluteraldehyde-fixed, postosmicated tissue. This does not dissolve the araldite
but punctures holes in it. Sodium metaperiodate etching retains the best tissue
structure of all the etching procedures we have employed (Figs. 2 and 3), but was
not as effective as potassium methoxide in revealing antigenic epitopes of PrP.
5. Triton/Tween (detergents): can be used for membrane permeabilization and there-
fore enhance penetration of the immunoreagents. In our experience, incorporat-
ing 0.2% Tween 20 in the wash buffer results in cleaner EM stained grids (fewer
The detection of PrP deposits in tissues prepared for light or electron
microscopy and routinely immunostained without pretreatments is limited to
the detection of plaques. We have employed the following pretreatments for
detection of disease-specific PrP:
1. Formic acid: Good tissue structure was obtained using 98% formic acid incubated
for 30 min for light microscopy and 10 min for electron microscopy. A wide range
of patterns of disease-specific PrP have been detected, including preamyloid forms.
2. Hydrated autoclaving: Good enhancement of immunoreactivity has been
achieved at light microscopy. Because this method is the most simple to perform
Jeffrey and Goodsir

Fig. 2. An 80-nm thick section from the lateral hypothalamus of a mouse infected
with the 87V strain of murine scrapie. This section has been etched with sodium
metaperiodate and pretreated with formic acid, immunostained for lB3 antiserum
using 1 nm immunogold, and enhanced for 5 min with silver. Section shows the elec-
tron microscopical counterpart to Fig. 1 in which part of a morphologically normal
neuron is surrounded by disease-specific PrP accumulation.

it has been used for routine screening of PrP distribution in 1-pm thick sections.
However, at the EM level, tissue structure is considerably compromised. EFFECT OF ANTIBODY
The effect of employing different antibodies for the detection of PrP at LM
level is described in detail elsewhere in this volume. However, in our labora-
tory, several antibodies for the detection of PrP in the brains of BSE, FSE,
ovine scrapie, and murine scrapie-infected tissueshave been investigated. Brief
details are presented in the following:
1. lA8 and lB3 antibodies (11,12) are high-titer polyclonal rabbit sera raised to
murine scrapie-associated fibrils. They recognize several sites of the proteinase-
resistant core of the PrP molecule (13) and recognize murine disease-specific PrP
in all fixative and embedding protocols. These two antibodies also recognize
PrPBSE, PrPFSE, and PrPSC of natural sheep scrapie.
2. 3F4 (14) detects feline PrP in paraffin-wax-impregnated sections but is lost in
gluteraldehyde-fixed tissues.
lmmunohistochemistry of Resinated Tissues 307

Fig. 3. An SO-nm thick section from the lateral hypothalamus of a mouse infected
with the 87V strain of murine scrapie etched with sodium metaperiodate and pretreated
with formic acid, immunostained for 1B3 antiserum using 10-nm colloidal gold. Sec-
tion shows association of 10 nm colloidal gold particles with small bundles of amyloid
fibrils forming in between neurites and glial cell processes.

3. N-terminal peptide antibodies are preserved in gluteraldehyde and paraformalde-
hyde fixatives, but much of the reactivity is lost on embedding in araldite. Pep-
tide-specific antibodies to the protease-resistant core amino acids of PrP (PrPres)
are preserved under the same conditions.
4. A peptide antiserum raised to sheep PrP residues 141-153 (SP30) was effective
in detecting PrP in FSE-affected brains. However, a second sheep peptide antise-
rum raised to PrP residues 219-233 (SP40) was not so effective against disease-
specific PrP in FSE but stained disease-specific PrP in murine scrapie and BSE.
Neither of these two antisera worked well for sheep scrapie.
In summary, prior to investigation of PrP localization in the electron micro-
scope it is necessary to identify the antibody that best recognizes the disease-
specific form of PrP to be detected. The effectiveness of that particular
Jeffrey and Goods/r

antiserum m recognizing specific PrP epitopes under various denaturing, etch-
ing, and fixation regimens then must be determined,
2.1.3. lmmunohistochemical Methods and Visualization of Reaction
Product: Light MIcroscopy and Electron Microscopy
1. To detect antigen in plastic-embedded thick sections it IS possible to employ
either the peroxidase-antiperoxldase (PAP) methods or to use avldin blotm com-
plex (ABC).
2. In our own laboratory, the PAP method IS preferred because it tends to give rise
to fewer problems with background staining.
3. The ABC method IS more sensitive and may have advantages where only low
levels of PrP are present, such as m early preclinical disease, or where antibody IS
In short supply.
4. For light microscopy, reaction product may be visualized by 3-3, dlammoben-
zidme, or alkaline phosphatase.
5 Colloidal gold particles of different diameters may be used to lmmunostam tls-


. 44
( 45 .)