METHODS: A Companion to Methods in Enzymology 1.0, 257-269 (1.996)Article No. 0100

Autometallography (AMG) allows the silver amplifi-cation of nanometer-sized catalytic crystals, i.e., crys-tals or crystallattices with the ability to convey elec-trons from reducing molecules, adhering to the surfaceof the particle, to likewise adhering silver ions. Suchcrystals will ignite the AMG process: shells of metallicsilver will grow around them and reveal, with nanome-ter precision, their position in the tissue. AMG is there-fore most valuable not only für tracing colloidal goldparticles used as labels of immunoglobulins, lectins, orenzymes at light microscopy (LM) and electron micros-copy (EM) levels, but in general für tracing AMG ignit-ing heavy metals (1-3).

Immunogold-silver staining (IGSS) utilizes a histochemicalmethod called autometallography (AMG) to amplify ti ny gold par-ticles to sizes easily visible both in light and electron micros-copy. In both applications it is advisable to use the smallestpossible gold diameters (1-6 nm) to obtain the highest sensitiv-ity, thus, allowing minute amounts of the target substance tobe demonstrated. Gold labels smaller than 10 nm in diameterhave been clearly shown to give the highest labeling densitiesof antigen-antibody binding sites. AMG can be used tor thedetection of catalytic crystallattices of metallic gold and silver,and sulfides or selenides of mercury, silver, copper, bismuth,and zinc. The method has its roots in "physical development"technique, transplanted tram photography to histology byLiesegang at the beginning of this century. In 1981, aseries ofpapers were published by one of us with the purpose of introduc-ing a reliable and easy-ta-handle technique tor light microscopi-cal and ultrastructural studies. AMG has a multitude of applica-tions apart tram its use in detecting tissue metals. These includethe highly sensitive and efficient in situ colloidal gold tracing ofpeptides, proteins, and amines by immunocytochemistry usingthe IGSS method, of carbohydrates by lectin IGSS, and of nucleicacids by IGSS in situ hybridization, IGSS in situ polymerase chainreaction, and IGSS in situ self-sustained sequence replication-based amplification (in situ 3SR) techniques, the last two evenperforming with single-copy sensitivity. Applications of pre- andpostembedding AMG tor semithin and ultrathin tissue sectionsare described. @ 1996 Academic Press, Inc.

DETECTION OF CATAL YTIC METALSUSING AMG

Because the field of AMG and its applications is rela-tively young and had received little attention untilAMG was introduced as an amplifier of colloidal goldparticles, knowledge about AMG igniters is stilllim-ited. The postulation that all metal sulfides could beAMG silver amplified does not hold true (2-4). It isnecessary to carefully map which metal can be boundas AMG igniting crystal lattices, under which condi-tions it will take place, and which rules Olle must ob-serve to be sure that the result is correct. It haB nowbeen proven that crystallattices ofmetallic gold, metal-lic silver, silver sulfides, silver selenides, mercury, cop-per, bismuth, and zille are AMG igniters.

Gold

Organisms exposed to gold salts will distribute themetal to different cells where it will eventually end up

257

1 To whom correspondence and reprint requests should be ad-dressed at Institute of Pathological Anatomy, Immunohistochemis-try and Biochemistry Unit, Salzburg General Hospital, MuellnerHauptstrasse 48, A-5020 Salzburg, Austria. Fax: 43-662-4482-882.

1046-2023/96 $18.00Copyright @ 1996 by Academic Press, Inc.All rights ofreproduction in any form reserved.

Gerhard W. Hacker,*'F Wolf gang H. Muss,*Cornelia Hauser-Kronberger,* Gorm Danscher,+Robyn Rufner,§ Jiang Gu,§ Huici Su,* Arne Andreasen,+Meredin Stoltenberg,+ and Otto Dietze**Institute of Pathological Anatomy, Salzburg General Hospital, A-5020 Salzburg, Austria; tMedical ResearchCoordination Center, University of Salzburg, A-5020 Salzburg, Austria; :j:Department of Neurobiology, TheSteno Institute, University of Aarhus, DK-8000 Aarhus, Denmark; and §Deborah Research Institute, BrownsMills, New Jersey 08015-1799

HACKERETAL.258

in the lysosomes. Humans who suff er from rheumaticdiseases are occasionally treated with aurothio com-pounds für curative purposes. Ifbiopsies are taken fromsuch individuals (e.g., during autopsy), or from gold-exposed experimental animals, AMG development ofthe sectioned tissues will show no staining at all. Thisis because gold present in the tissue as gold ions ischemically bound in a way that makes it "invisible"to the AMG developer. If, however, the sections areradiated with ultraviolet light or subjected to a re duc-ing solution, the gold ions will be reduced to gold atomsand create igniting centers (1,2,5). It was this observa-tion in 1981 that led to the introduction of AMG ampli-fication of colloidal gold particles in immunohistochem-istry, leading to immunogold-silver staining (lGSS)techniques (6), and tagging enzymes at cellular andultracellular levels (7). The AMG gold technique hagbeen used für several ultrastructural studies (e.g., 38,40-42), and recently we have worked out a techniquethat makes it possible to differentiate between gold andother defined AMG igniters.

SilverThe fact that metallic silver can initiate the AMG

process hag been known für more than 100 years, hutZieger (8) was the first to demonstrate that silver sul-fide molecules are AMG igniters as weIl. Informationgathered by multielement analysis of biopsies from anargyrotic patient in Norway (9) and autopsies from aJapanese patient poisoned by organic mercury duringthe Minimata accident (10) led to the hypothesis thatBorne metal selenide molecules might have the sameAMG igniting capacity as the corresponding metal sul-fide molecules (2, 11). The idea proved valid when itwas demonstrated that zille selenide crystal latticescreated in vivo by exposing rats intraperitoneally tosodium selenite could be silver amplified (11) andthereby a whole new group of potential AMG igniterswas introduced. The AMG silver method für LM andEM demonstration of silver sulfide crystals in tissuesfrom humans and animals exposed to silver wasworked out in 1981 (12, 13). Several studies have usedthe technique to analyze possible toxicological aspectsof silver, and the method disproved an old undisputedstatement by Liesegang (14) and demonstrated silverin neurons and glial cells from exposed animals für thefirst time.

Bismuth

The autometallographic technique hag received re-newed interest following the introduction of AMG de-velopment für the demonstration of bismuth (22). Onlya few papers are available yet, and the technique hagnot been modified für EM studies to our knowledge.Ross and his group have described the distribution ofBi accumulations in the CNS most carefully. It seemsas if the AMG demonstrable accumulations of Bi arelocated mostly intracellularly and have a regional dis-tribution similar to what is found after mercury expo-sure (16). The reason that the AMG technique works onbismuth-containing tissue sections must be that part orall of the bismuth bound in the tissue is present asbismuth sulfide or bismuth selenide crystallattices (5).

CopperIn 1989, Lonnee et al. (23) introduced AMG silver am-

pIification of copper sulfide crystals created by the cuprol-inic blue technique für visualization of polyanionic glycos-aminoglycans (23, 24). Silver amplification of AMGigniters used in LM and EM histocheInical methods forcarbohydrate tracing haB been recently reviewed (25).Also reviewed are procedures such as periodic acid-thio-carbohydrazide-silver protein-AMG (25). It haB notbeen possible to demonstrate endogeneous or exogenouscopper in tissues, perhaps because copper is cheInicallybound in such a way that a cheInical transfonnation tocopper sulfide/selenide crystallattices is not possible. It

MercurySince Timm's (15) original introduction of AMG for

the demonstration of mercury accumulations in organsfrom exposed animals, a multitude of modificationshave jaded the field and brought into disrepute thisforceful technique. Major wrongdoings included the useof sulfur-controlling fixative or the introduction inother ways of agents that could lead to the creation of

sulfide ions. Sulfide ions, in turn, create AMG ignitingzinc sulfide crystals in tissues that contain chelatablezinc ions. Reapplication of the original Timm approachaugmented with a new AMG developer suitable für ul-trastructural analysis (16) haB increased its use in toxi-cological studies. Through further analysis of Timm'soriginal method, it was demonstrated that the success-ful autoradiographic studies of radioactive mercury de-scribed in the literature always, or in most cases, werea result ofAMG.

When 203Hg-containing sections covered with the au-toradiographic emulsion were developed, the developermoved through the emulsion into the tissue section,carrying with it the reducing silver ions released fromthe silver bromide crystals in the emulsion. Subse-quently, the AMG developer was created in the tissuesections, and crystals ofmercury sulfide/selenide wouldbe silver-amplified. This observation explained the in-comprehensible finding that nonradioactive mercurycould be traced by autoradiography (17) and led to theidiom autometallography für this particular kind of de-velopment of metals in tissues instead of the mis-leading name physical development (18, 19). The tech-nique is easy to use, and mercury sulfide/selenidecrystal lattices can be differentiated from all otherAMG igniters (Fig. 1) (2, 20, 21).

EM IMMUNOGOLD-SILVER STAINING AUTOMETALLOGRAPHY 259

FIG. 1. Electron micrograph ofnanometer-sized mercury sulfide/selenide crystals (arrows) silver amplified with the lactate AMG developer.The tissue came &om a thyroid gland of a 10-year-old sledgedog, killed in 1987 in Thule, Greenland. Bar, 2.5 j},m.

HACKER ET AL.260

sis, selected sections can be reembedded on top of ablank Epon block; or (iv) vibratome sections 150 jJ;mthick can be AMG developed. Blocks, or interestingportions of tissue cut from these sections, are stainedwith osmium tetroxide and uranyl acetate, dehy-drated, and embedded lege artig (3).

is easy to determine whether tissues &om sulfide- or sele-nide-exposed animals contain such copper or zinc sulfide!selenide crystals as a weak acid like 0.1 N HCI will dis-solve a11 the zinc crystals hut leave the copper crystalsunaffected (5).

IMMUNOCYTOCHEMISTRY AND RELATEDTECHNIQUES

ZincPools of zinc present in certain synaptic vesicles in

the CNS and in secretory vesicles of several endocrinecells can be demonstrated cytochemically by two differ-ent AMG methods. Recently developed in vivo tech-niques involve binding ofthe (most likely) free zinc ionsin the vesicles to zinc sulfide/selenide crystallattices(3, 11, 12). These AMG techniques have been inten-sively used in brain research, where a particular groupof neurons, abundantly present in cortical regions andmost likely glutaminergic, is demonstrated. The zinc-enriched neurons (ZEN) have attracted increased sci-entific interest as it haB been found that they are im-portant in memory (26, 27) and that zinc ions have anantagonizing effect on the glutaminergic N-methyl-D-aspartate receptor (28).

A new technique that allows the zinc ions in synapticand secretory vesicles to be transformed to nanometer-sized zinc crystallattices für subsequent AMG develop-ment haB just been described (29). Human brain biop-sieB or other blocks of tissue containing ZEN cells areplaced in liquid nitrogen or frozen by CO2 immediatelyafter removal. The tissue blocks are cut in a cryostat,and the sections are placed on glass slides and trans-ferred to a H2S exposure chamber at -20°C. After peri-ods of 5 min to 24 h, the sections are thawed, fixed,and dehydrated. The sections are then exposed to anAMG developer. AMG causes a silver amplification ofzinc sulfide crystal lattices in the tissue, making thechelatable vesicular zinc pool visible.

The AMGznX.H2S technique can, with substantiallossof quality, be used to locate zinc ions at ultrastructurallevels. It haB been demonstrated that zinc ions in theneocortex of human brain are located in synaptic vesi-eies ofZEN neurons and that the pattern and morphol-ogy of the ZEN terminals is nearly identical to that ofthe neocortex of rat brain.

The extreme sensitivity of AMG hag also made thetechnique useful in immunocytochemistry to detectsubstances recognizable by antibodies (i.e., proteins,peptides, amines) and lectins (i.e., agglutinin-bindingcarbohydrates) (30-55). A number of applications forsingle- or low-copy detection of specific DNA or mRNAsequences using in situ hybridization, in situ polymer-ase chain re action (PCR), and in situ self-sustainedsequence replication-based amplification (in situ 3SR)have been described recently (57-66). In all thesemethods, the detection step is based on colloidal gold-adsorbed macromolecules (immunoglobulins, streptav-idin, and protein A). The immunohistochemical AMGapproach is called immunogold-silver staining (IGSS)(6), and it is highly sensitive and detection-efficientprovided that an adequate silver amplification methodis used (48, 54). Silver lactate AMG (1, 12, 13), andsilver acetate AMG (40,49) are both very efficient (48,52, 55). In addition, the latter hag relatively low sensi-tivity to light (40,52,67). In LM, a dark-gray or blackspecific staining is observed even if very low quantitiesof the labeled substance are present, as is the case inmany applications on semithin resin sections. For post-and preembedding ultrastructural studies, gold parti-eIes of only 1 to 6 nm in diameter can be used to obtainincreased labeling densities and better penetrationproperties (33, 49, 53). Subsequent silver amplificationby AMG makes these initially tiny gold particles visibleeven at low magnifications.

Colloidal gold as a label for immuno-EM was intro-duced by Faulk and Taylor (68). Their immunogold-staining (lGS) technique hag become the method ofchoice for on-grid EM-immunocytochemistry, having anumber of advantages compared to other, nonparticu-late immunostaining techniques (69, 70). Geogheganand collaborators (71) used colloidal gold sols exhib-iting a red color for LM-IGS, hut their method hadlow sensitivity. The real breakthrough came with theintroduction of silver amplification (AMG) of colloidalgold particles by Danscher and Norgaard (7) and Hol-gate et al. (6). Their applications of AMG for enzyme-gold-silver staining and IGSS resulted in substan-tially increased sensitivity and detection efficiencycompared to unamplified IGS and to most other immu-nocvtochemical techniaues (6. 30. 32).~

APPLICATION OF AMG FORUL TRASTRUCTURAL EXAMINATION

Electron microscopical AMG can be implementedin several different ways. (i) Nickel grids with ultra-thin sections can be covered with an autoradio-graphie emulsion and developed with a chemical de-veloper; (ii) the grids can be placed upside-down ona drop of AMG developer; (iii) thick and semithinsections can be AMG developed, and, after LM analy-

261

Immunogold-Silver Staining (IGSS)IGSS, as originally described, is an indirect method

in which specific primary antibodies against the sub-stance to be detected are first applied and a secondlayer of gold-adsorbed antibodies directed against im-munoglobulins ofthe species in which the primary anti-body was raised are added (6). In LM, this indirectIGSS (two-step) setup appears to be preferable to di-reet, protein A-gold-silver staining or streptavidin-gold-silver methods; in our hands, it results in ahigher detection efficiency and yields lower levels ofunwanted background staining. IGSS not only allowsthe highly sensitive immunocytochemistry, hut it hagalso been adapted in a number of related fields, includ-ing in situ molecular biology (56-66). In addition toindirect IGSS, direct methods, bridge methods, strep-tavidin-biotin methods, protein A-gold-silver stain-ing, and various other combinations have been de-scribed (31, 34, 35, 39, 41, 45, 59). Preliminaryexperiments on nucleic acid detection, using a modifiedhut simple protocol involving streptavidin-Nanogold,promise sensitivities similar to those achieved by insitu PCR (Hacker, unpublished results). Suggestedworking procedures für LM and EM application cur-rently used in our laboratories are given under Proto-cols 1 and 2.

In thick paraffin and semithin resin sections, IGSStechniques show a number of advantages compared toother methods. Positive immunostaining may be ob-tained with IGSS where other methods fail; IGSS maytherefore facilitate the demonstration of substancespresent in only negligible quantities (6, 30, 32). Recentantigen retrieval methods employing heat treatmentin a microwave oven, table sterilizer, or pressure cookerallow the reliable detection of fIXation-labile antigenseven after prolonged aldehyde fixation and paraffinembedding, which were formerly not demonstrable byconventional immunocytochemistry (51, 52-75). Mi-crowave irradiation during the incubation of the pri-mary antibody and the subsequent immunogold attach-ment allows IGSS to be completed in only about 30 min(73, 76).

Positive reactions in LM and EM applications can beeasily identified due to the very intense signal, facili-tating the screening of sections even at low magnifica-tions. The technique allows the use of conventionalcounterstains such as hematoxylin and eosin and/ornuclear fast red on paraffin, or azure lI-methylenebille, and basic fuchsin on semithin resin sections. ill-trathin sections can be counterstained with osmiumtetroxide, lead citrate, or uranyl acetate, therebygreatly improving assessment of morphology (30, 32,40, 55, 58, 65). In IGSS, hazardous reagents such asthe potentially carcinogenic diaminobenzidine-tetrahy-drochloride (DAB) commonly used in peroxidase-basedmethods are avoided.

Various tissue fixatives can be used. For ultra-structural studies, it is advisable to use a mild fixa-tion with a buffered mixture of glutaraldehyde andparaformaldehyde, or Stefanini/Zamboni's solution(77). Osmium tetroxide fixation should be avoidedbefore AMG, hut can be performed after silver ampli-fication in conjunction with uranyl en blaG staining(5). The embedding process (we have tested LR-White, Lowicryl, Epon, and Araldite) should in so mecases include low-temperature polymerization (forsame antigens not higher than 40°C).

For the IGSS protocol, LM paraffin and semithinresin sections should be pretreated with Lugol's iodinefollowed by sodium thiosulfate. This treatment ortenincreases the staining efficiency. If iodine treatment isexcluded, the AMG amplification process in most casesneeds to be prolonged. For ultrathin sections, Lugol'siodine may be recommended, hut does not have to beapplied. Most protocols in the literature use Tris-buf-fered saline (TBS) or phosphate-buffered saline (PBS)with a pH of about 7.2 as washing buffers. The use ofbuffers with high galt concentrations and the additionofTriton X-100 or Tween 80 to the buffer system priorto the application of the primary antibody sometimesimproves the staining. Also, an increased NaCI concen-tration sometimes reduces background staining. Thebuffer system to be used before immunogold incubationshould be adjusted to pH 8.2 to stabilize the gold re-agent. However, to make the staining protocol easierto handle, it is also feasible to use a buffer with a pH ofabout 7.6 for all buffered washes. To avoid nonspecificreactions, the addition of 0.1% fish gelatin (e.g., coldwater fish gelatin, Aurion, Wageningen, The Nether-lands) hag proved to be very effective (51, 54). Othertypes of gelatin may suffice hut must be first tested;gelatin quality hag been found to affect staining.

Polyclonal rabbit or guinea pig antisera and mono-clonal mouse or rat antibodies are orten used to detectvarious substances at the LM and EM levels. Primaryantibodies are typically incubated for 60-90 min atroom temperature, however, and this time span can besignificantly reduced by microwave incubation (73). Onthe other hand, longer incubation and/or a second ap-plication of the same antibody may further increasethe detection efficiency and may allow the use of higherantibody dilutions (78). Consequently, the optimumconcentration of antibodies and immunogold reagentsshould be evaluated by testing various dilutions. It hagbeen shown that for IGSS methods, in comparison tomost other immunocytochemical techniques, primaryantibodies can orten be diluted much further, drasti-cally reducing the cast of routine immunocytochemistry(6, 30, 32).

Affinity-purified secondary antibodies adsorbed togold particles (immunogold reagents) or protein-A-gold can be obtained from several companies. Quality

HACKER ET AL.262

forceps should be cleaned für 30 min in a 10% Farmer'ssolution (one part sodium thiosulfate and nine parts10% potassium ferricyanide).

Silver lactate and silver acetate have been widelyused as the ion source in the formation of shells ofmetallic silver around small gold particles (1-5, 33, 48,55). AMG is catalyzed by hydroquinone in a low-pRcitrate buffer. During the process, gold particles in-crease in size and conglomerate if sufficiently near eachother (33). In LM, this is seen as a grayish to blackprecipitate sharply contrasting with the unreactedbackground. The detection efficiency of IGSS isstrongly related to the type of AMG enhancement usedand to the quality of the immunogold reagents (1, 3,12, 13, 48, 51, 52, 55). Experiments comparing silverlactate, silver acetate, and silver nitrate with equiva-lent molarities (0.02 M) under standard conditions ofLM- and EM-IGSS showed that silver lactate (Fig. 2a)and silver acetate AMG (Fig. 2b) gave equally optimalsilver amplification without nonspecific staining (55,58). In contrast, silver nitrate gave an unacceptablebackground staining and strong hut very uneven am-plification (Fig. 2c) (55, 58). We also found that a neu-tral pR of the AMG buffer causes argyrophilic-type"nonspecific" reactions, primarily with collagen. In con-tragt, low pR (pR 3.8) and the addition of gelatin and/or gum arabic as protecting colloids yielded specific,even, uniform, and exactly controllable silver amplifi-cation (unpublished data; see also 48). The addition offish gelatin to the washing buffers before and afterimmunogold-incubation also helps to prevent nonspe-cific silver precipitation, both in LM and EM (51, 54).

Semithin sections are developed vertically in a glasscontainer, e.g., in a slide container according to Schief-ferdecker (coplin jar) containing 80 ml of AMG devel-oper. Ir a silver lactate developer is to be used the con-tainer should be kept in the dark (1-3, 12, 13). BothAMG developers permit a monitoring of the stainingintensity by visual control in LM. When the silver lac-tate developer is used, sections rinsed in double-dis-tilled water can be checked under the microscope andthen further developed if necessary. With silver acetateAMG, rinsing in distilled water before microscopicalobservation is not necessary (40).

For silver amplification ofimmunogold-labeled ultra-thin sections on EM grids, both silver lactate and silveracetate AMG should be covered with a darkbox. In thiscase, amplification is carried out by floating the gridsface-down on drops of freshly prepared developer sup-ported on Parafilm or dental wax (52, 55). The gridsmust not be completely submerged in the AMG solu-tion, otherwise "crunchy preparations" may result. De-velopment ofEM sections should be carried out für 3-15 min at room temperature, depending on the size ofthe gold and the AMG developer used (Fig. 3) (see also2, 3, 19).

and prices differ greatly, and für that reason, it is advis-ahle to test different gold reagents. In Dur experience,high-quality reagents may be purchased from Nano-probes (Stony Brook, NY), Amersham (Amersham,UK), BioCel1 (Cardiff, UK), and Aurion. Dilutions mustbe optimized by titration and are usually between 1/25and 1/200. TBS, pH 7.6 to 8.2, used as gold reagentdiluent, should contain 0.8% bovine serum albumin(BSA) and 0.1% gelatin; this helps to prevent aggrega-tion of gold particles and results in less backgroundstaining. The postfixation in 2% glutaraldehyde afterthe washing steps prevents release of gold reagent fromits binding Bites in the low-pH environment of AMGdeveloper. Glutaraldehyde should be diluted in PBS,pH 7.2, and will remain stahle in this buffer für at least2 weeks at 4°C für reuse.

As clearly demonstrated earlier (33, 49, 53), immuno-labeling is most efficient if gold particles of small size(1-5 nm) are used. Such immunogold reagents pene-trate sections better and achieve particle densitieshigher than with larger gold particles. The most suit-ahle gold particle sizes are 5 nm or even smaller fürLM and EM applications. One of the best immunogoldreagents available today is Nanogold (Nanoprobes).This product does not use colloidal gold hut uses highlyuniform, covalently bound lA-nm gold particles, sur-rounded by an organic sheet made up of proteins (79).Unlike other colloidal golds, Nanogold particles do nothave affinity to proteins; thus reducing backgroundand falBe labeling. We have shown that Nanogold alsogreatly improves the detection ofintranuclear antigensby IGSS, by eliminating the charge interaction betweenconventional gold probes and nucleic acids. AlthoughNanogold is barely visible without silver amplification,a 4- to 6-min incubation with silver acetate or silverlactate AMG will yield electron-dense gold-silver par-ticles of about 50-80 nm in size with the final sizedepending upon the time of development and the com-position of the developer. Silver amplification en blocor on a grid must be completed before any stainingreagents such as osmium tetroxide, lead citrate, or ura-nyl acetate are applied, because these will nucleatesilver deposition in the same manner as gold to producenonspecific staining.

Our laboratories orten use mixtures of severaloptimally diluted IgG subgroup-specific immunogoldbatches of one (EM) or more (LM) gold sizes from differ-ent companies, each mixed in equal amounts (51).These mixtures appear to give better labeling withBorne primary antibodies.

Before AMG gold-silver amplification is carried out,semithin and ultrathin sections must be washed care-fully in ultrahigh-quality glass double-distilled water.For optimal staining results, the purity ofwater is cru-cial. Only thoroughly clean glassware and plastic orTeflonized forceps should be used (51). Ifnecessary, the

FIG.2. (a-c) Representative electron micrographs comparing silver lactate (a), silver acetate (b), and silver nitrate (c) AMG after indirectimmunogold labeling. EMbed-812 sections ofright heart atria ofrat, demonstrating a-atrial natriuretic peptide (a-ANP) ofmyocardiocytesunder standardized conditions (0.02 M of each silver galt, together with 0.05 M hydrochinone in citrate-buffer, pH 3.8). Development for 6min in dark. Note the quite uniform amplification by the silver lactate and acetate developers. Silver nitrate, in addition to specific staining,gave argyrophilic reactions in collagen fibers. Bar, 0.5 p,m.

263

Use of AMG for DNA and RNA DetectionNucleic acid hybridization probes can be applied to

detect viral genomes recognizing infected cells, to in-vestigate biosynthesis of peptides and/or proteins, orto study genetic disease. For nonisotopic labeling, bio-tin or digoxigenin are now used with great success (e.g.,51-66). These labels satisfy the demand ofmost patho-logicallaboratories as probes can be stored and handledwithout the hazard ofradioactivity. Nonradioactive la-bels are also cheaper, easier to handle, and give ahigher resolution than radioactive probes. Nonradioac-tive reporter molecules applied für nucleic acid detec-tion can be easily demonstrated by using direct, indi-reet, or streptavidin-biotin IGSS methods. Optimizedprotocols für in si tu DNA hybridization with biotin-labeled probes and IGSS techniques have been de-scribed (59, 61). Most recently, applications of AMG fürdirect and indirect in si tu PCR, or in situ self -sustainedsequence replication-based amplification (in situ 3SR)are being discussed (62-66). These methods allow, fürthe first time, detection ofsingle copies ofDNA or RNAat the cellular and subcellular level with AMG. DNAor RNA stainings in EM (65, 80) (Fig. 4) (ProtocoI3) canbe accomplished using these techniques in conjunctionwith preembedding methods on formalin- or paraform-aldehyde-fixed cells.

PERSPECTIVES

Protocols für the use of AMG in the detection of cata-lytic tissue metals have been extensively described inthe literature. AMG and its application to immunocyto-chemistry and other IGSS-related methods includingthose für ultrastructural studies, hag manyfold advan-tages over comparable conventional methods. By usingsmall (1- to 5-nm) gold particles, high labeling densitiesand good penetration properties are obtained. Theseproperties are particularly advantageous für preem-bedding techniques applied to tissue sections whereonly scanty immunoreactive structures are present. Inmany cases, higher sensitivities than those yielded byunamplified IGS can be obtained. Olle impedement toa broader use ofIGSS is that many commercial compa-nies appear unable to produce high-quality immuno-

FIG. 3. Human pancreas endocrine cella in the islets of Langer-hans, immunostained for insulin (a, b) or somatostatin (c). (a) Unam-plified 6-nm gold particles sitting on secretory granules. Bar, 100nm. (b) A comparable preparation, hut amplified with silver acetateAMG for 8 min. Note the considerable increase in size and visibilityoflabeled particles. Bar, 100 nm. (c) A higher magnification showingsomatostatin-containing secretory granules labeled by 5-nm goldparticles and silver acetate AMG for 8 min. Note the somewhat un-even amplification of gold particles. Bar, 100 nm.

FIG. 4. TEM examination of a resin (Epon-812)-reembedded paraffin section of fetal lung tissue reacted for cytomegalovirus (CMV)(preembedding in situ hybridization). (a) A low magnification of the nuclear area showing accumulations of positive signal within thenucleus. Bar, 1.0 Jlm. (b) A high-power view of the cytoplasmic area showing direct labeling of CM virus particles (V) with silver-amplifiedgold particles. N, nucleus; NM, nuclear membrane; cy, cytoplasm; V, CMV particles; vaP, virus-associated protein. Bar, 100 Dm.

nucleus, probably because of charge repulsion."Crunchy" looking background staining is orten due topoor optimization of each single step of the protocol.Improperly cleaned glassware, poor quality distilledwater, and low-quality and nonspecific (though com-paratively expensive) commercial "silver enhancement

gold reagents, or even to maintain quality standards.In our experience, several immunogold reagents avail-able on the market did not give a high labeling densityand a very low background, as is required für such asensitive technique. Also, there are frequent problemswith poor penetration of some reagents into the cell

266 HACKER ET AL.

kits" are frequently problematic. Such problems shouldbe solvable so that IGSS with AMG will be the methodof choice in supplementing conventional IGS in ultra-structural studies.

Protocol 1: Indirect IGSS Method on Semithin ResinSections

1.A. lmmunocytochemistry

1. Mount semithin (Epon or Araldite) resin sectionson poly-L-lysine (PLL)- or aminopropyl-triethoxisilane(APES)-coated glass slides (81, 82) and dry für 1 h at60°C. Before staining, treat sections with saturated so-dium ethoxide für about 20 min and thoroughly washin ethanol (3x 2 min).

2. Wash in distilled water (3 min).3. Perform heat antigen retrieval if desired.4. Immerse in Lugol's iodine (1% iodine in 2% potas-

sium iodide; Merck No. 9261, Darmstadt, Germany) (5min).

5. Rinse briefly in tap water, followed by distilledwater.

6. Treat briefly with 2.5% aqueous sodium thiosul-rate until sections become colorless (up to 30 s).

7. Wash in distilled water (2 min).8. Immerse in TBS-gelatin (Tris-buffered saline, pH

7.6, containing 0.1% cold-water-fish gelatin) (10 min).In same cases superior staining is obtained ifthe bufferin this step also contains 0.1% Triton X-I00 or Tween80, and 2.5% NaCI.

9. Apply normal serum of the species providing thesecondary antibody (1/10 in TBS-gelatin) (5 min) anddrain.

10. Incubate with primary antibodies (90 min at RTor overnight at 4°C). The dilution to be used should becarefully determined. The suggested antibody diluentis 0.1 M phosphate- or Tris-buffered saline (pH 7.2-7.6) containing 0.1% bovine serum albumin and 0.1%sodium azide.

11. Wash in TBS-gelatin (3x 3 min).12. Apply normal serum at 1/10 dilution as in step 9.13. Incubate with gold-adsorbed second-layer anti-

bornes (60 min at RT). Optimum dilution is usuallybetween 1/25 and 1/200 and should be determined bytitration.

14. Wash in TBS-gelatin (3x 3 min).15. Postfix in 2% glutaraldehyde in PBS, pH 7.2 (2

min).16. Rinse briefly five times in distilled water (about

30 s each), followed by three washes (3 min each) inthe same.

17. Perform silver acetate autometallography.

(Code 85140, Fluka, Switzerland) in 40 ml of glass dou-ble-distilled water. Dissolve silver acetate crystals bycontinuous stirring für about 15 min.

2. Citrate butTer: Dissolve 23.5 goftrisodium citratedihydrate and 25.5 g citric acid monohydrate in 850 mlof deionized or distilled water. This butTer can be keptat 4°C für at least 2-3 weeks. Before use, adjust to pH3.8 with citric acid solution.

3. Solution B: Dissolve 200 mg hydroquinone in 40ml citrate buffer.

4. Just before use, mix solution A with solution B.5. Silver amplification: Place the slides vertically in

a glass container (preferably with about 80 ml volume,für up to 19 slides) and cover them with the mixture ofsolutions A and B. Staining intensity may be checked inthe light microscope during the amplification process.

6. One may use photographic fixer (e.g., Agefix, AgfaGevaert, FRG, diluted 1/20) to stop the AMG processimmediately. (This solution can be reused für severalstainings.) Leave the slides in this solution für a maxi-mum of 10 s. Altematively, one may use a 2.5% aque-ous solution of sodium thiosulfate.

7. Rinse the slides carefully in tap water für at least3 min. After AMG, one may counterstain sections withazure li-methylene blue and basic fuchsin, dehydratethem, and mount them in DPX (BDH Chemicals, UK).

Protocol 2: Indirect IGSS Method für Ultrathin Sections

1. Mount ultrathin sections on nickel grids (300-400 mesh) and dry für 1 h at room temperature.

2. For Epon sections, etching with H2O2 and/or satu-rated sodium ethotide is recommended. Wash thor-oughly in distilled water (2x 3 min).

3. Immerse in TBS or PBS, pH 7.6, containing 0.1%cold-water-fish gelatin.

4. Apply normal serum of the species providing thesecondary antibody (1/20 in TBS-gelatin plus 1% BSA)(20 min) and drain.

5. Incubate with primary antibody (ovemight at4°C) using a microtiter plate. The dilution should betested carefully. The suggested antibody diluent is 0.1M TBS or PBS containing 1% BSA.

6. Wash in TBS or PBS containing 1% BSA and 0.1%gelatin (3x 3 min).

7. Incubate with gold-adsorbed second layer anti-bodies (ovemight at 4°C). Optimum dilution is usuallybetween 1/20 and 1/100 and should be determined bytitration.

8. Wash in PBS or TBS containing 1% BSA (2x 3min).

9. Wash in PBS (2x 3 min).10. PostfIX in 2% glutaraldehyde in PBS, pH 7.2

(2 min).11. Rinse briefly three times in high-purity glass-

distilled water (about 30 s each), followed by threewashes (3 min each) in the same.

l.B. Silver Acetate Autometallography

1. Prepare fresh mixtures of solutions A and B forevery run. Solution A: Dissolve 80 mg silver acetate

267

15. Treat with 2.5% aqueous sodium thiosulfate un-til sections become colorless and wash in distilled water(3x 2 min).

16. Immerse in TBS-gelatin (Tris-buffered saline,pH 7.6, containing 0.1% cold-water-fish gelatin (2x 3min).

17. Incubate with streptavidin-Nanogold (Nano-probes), gold-adsorbed anti-biotin (Amersham; Nano-probes) or anti-digoxigenin (Boehringer Mannheim, orAurion, antibodies für at least 60 min at RT or over-night at 4°C. Optimum dilution is between 1/25 and 1/50. Antibody diluent is TBS-gelatin containing 0.8%bovine serum albumin.

18. Wash in TBS-gelatin (3x 3 min).19. Postfix in 2% glutaraldehyde in PBS, pH 7.2 (2

min).20. Apply silver acetate autometallography as under

Protocol I, Do not counterstain.

SSC is standard sodium citrate buffer. (Preparationof 20x SSC: 175.32 g NaCI, 88.23 g sodium citrate in1 liter H2O; adjust to pH 7.0 with HCI or citric acid;premixed concentrate Is available from Sigma (No. S-6639)).

Note thai in carrying out the detection ofPCR-ampli-fied nucleic acid sequences by hybridization (indirectin si tu PCR), Lugol's iodine treatment and subsequentreduction by sodium thiosulfate should be avoided.

12. Perform silver acetate autometallography (seeProtocol1) on top of drops of AMG developer placed ondental wax or Parafilm für 3-10 min. Protect from lightwith a darkbox. Use plastic or Teflonized forceps toavoid any impurity.

13. Rinse in double distilled water (2x 3 min).14. Dry ultrathin sections at room temperature (15

min).15. Stain sections as usual with lead citrate and ura-

nyl acetate.16. Examine by EM.

Protocol 3: Preembedding in Situ DNA Hybridization3.A. DNA in Situ Hybridization (45)

This protocol is still under development and is givenhere only as a guideline.

1. Cut thick (20 J.lm) paraffin or cryostat sectionsand mount on APES-coated glass slides (82). Cryostatsections in same cases must be taken into graded as-cending alcohols, followed by xylene für 30 min, andrehydration in graded descending alcohols.

2. PoStfIX cryostat sections in 5% phosphate-buf-fered formaldehyde (pH 7.0) or 2% buffered paraform-aldehyde (pH 7.0) (10 min or longer).

3. Wash sections in 20 mM PBS, pH 7.2 (3x 2 min).4. Soak in 0.3% Triton X-100 (15 min) to permeabil-

ize sections.5. Proteolytic treatment: Lightly digest the tissue

sections with 0.1% proteinase Kin PBS für about 4min (time and concentration depend on the strength ofprefIXation ofthe tissue, the type oftissue, and the sizeof the section) at 37°C.

6. Wash in PBS (3 x 2 min).7. Wash in distilled water (2 min), immerse in 50,

70, and 98% isopropanol (1 min each), and air dry atroom temperature.

8. Prehybridization: Incubate with 50% deionizedformamide and 10% dextran sulfate in 2x SSC, at 50°Cfür 5 min. Drain off excess. Care must be taken thatsections never dry from the hybridization step onward.

9. Place a small drop ofprobe mix (about 20 J.llofreadyto use-probes, or 20 ng ofnick-translated probe) onto thesection, and cover with a 22 X 22-mm coverslip. Biotinyl-ated or digoxigenin-Iabeled cDNA probes have been suc-cessfully used in this procedure.

10. Place the slides on a 92°C heating block and incu-bate für 5-10 min.

11. Move the slides into a 37°C oven and incubatefür two hours or overnight.

12. Remove coverslips by soaking with 4x SSC.13. Wash under stringent conditions in 2x SSC,

O.lx SSC, 0.05x SSC, and distilled water (each >5min) at room temperature or (better) at 37°C.

14. Immerse in Lugol's iodine (1% iodine in 2% po-tassium iodide) (5 min) and briefly rinse in distilledwater.

3.B. Preparation Steps tor Epon EmbeddingThis procedure haB been modified from Kummer et

al. (83).

1. Mter IGSS, in si tu hybridization, in situ PCR, orrelated methods, wash the stained preparations (sec-tions, cytological material, vibratome sections, smalltissue blocks) in 0.1 M PBS, pR 7.2 (3X 5 min).

2. PoStflX in 1% osmium tetroxide dissolved in 0.1 MPBS at room temperature (30 min to 1 h).

3. Wash in 0.05 M maleate butTer (3x 5 min). The0.2 M stock solution of this butTer haB pR 4.6-5.2 andshould be adjusted to pR 5.2 using 0.2 N NaOR.

4. Contrast in 1% uranyl acetate in 0.05 M maleatebutTer (the butTer solution should be adjusted to pH 6.0and will reach pR 5.2 when uranyl acetate is added)(1 h, in the dark).

5. Washin 0.05 M maleate butTer, pR 5.2 (3x 5 min).6. Dehydrate in 30, 50, 70, 90, 96% et~anol (5 min

each). '

7. Immerse in 100% ethanol (3x 5 min).8. Optional: Immerse in ethanol/propylene oxide

mixture (1/1, 5 min).9. Immerse in propylene oxide (2x 5 riiiri).10. Immerse in propylene oxide/Epon (1/1, 30 min).11. Overlay pure Epon and change this several times

(2 h, RT).12. Overlay with a new drop of Epon and place a

268 HACKER ET AL.

weighted Teflonized coverslip on top of the section. Letpolymerize overnight at 60°C.

13. Label areas of positive reactions and, optionally,document photographically with light microscope.

14. Carefully remove coverslip with a razor blade.This is easy to do with very small areas, hut difficultwith large areas of interest. Cut away uninterestingareas with a new razor blade, to leave a rectangularface for the remaining section site(s) to be investigated.Overlay the remaining section with a drop of freshEpon and a previously polymerized blank Epon block.Let polymerize overnight at 60°C.

15. "Pop-off' technique: Dip the warmed slide hori-zontally into liquid nitrogen (LN2) so that the upperedge of the glass slide itself is not covered with LN2.You will probably hear a light "cracking." While it iscold, carefully place the slide onto an insulating surfaceand attempt to break off the Epon block. Prepare andtrim block as for conventional TEM.

16. Very carefully cut one or two semithin (0.5-jJ;m)sections and then cut ultrathin sections. Be sure toobtain good sections from the beginning, using a high-quality diamond knife (e.g., Diatome, Bienne, Suisse).Stain the sections as usual with lead citrate. (Addi-tional uranyl acetate is usually not necessary, as thisis already included as an en bloc stain; compare withstep 4). Take electron micrographs on the first exami-nation-the electron beam may influence the qualityof the gold-silver particles with time.

The maleate buffer is made up from stock solutionsA and B. In preparing solution A, dissolve 23.2 g maleicacid or 19.6 g maleic acid anhydride in distilled waterto yield 1000 ml. Solution B is 0.2 N NaOH in distilledwater. To obtain 0.05 M maleate buffer 5.2, mix 50 mlof solution A and about 7.2 ml of solution B; to obtainbuffer at pH 6.0, mix 50.0 ml of solution A and 26.9 mlof solution B. Bring to a total volume of 200.0 ml withdistilled H2O. A final pH check of the buffer solutionsis recommended.

ACKNOWLEDGMENTS

For the joint collaboration in IGSS related projects, we sincerelythank J. M. Polak and D. Springall (London, UK), P. Lackie (South-ampton, UK), L. Grimelius (Uppsala, S), A.-H. Graf, S. Heil, A.Schiechl, and E. Zipperer (Salzburg, Austria).

REFERENCES

4. Danscher, G. (1988) Acta Neuropathol. 76,107.5. Danscher, G., and Montagnese, C. (1994) J. Histotechnol. 17,

15-22.6. Holgate, C. S., Jackson, P., Cowen, P. N., and Bird, C. C. (1983)

J. Histochem. Cytochem. 31, 938-944.7. Danscher, G., and Norgaard, J. O. (1983) J. Histochem. Cyto-

ehern. 31,394-398.8. Zieger, K. (1938) Wiss. Forschungsber. 48,55-105.9. Aaseth, L., alsen, A., Halse, J., and Hovig, T. (1981) Scand. J.

Clin. Lab. Invest. 41,247-251.10. Shirabe, T. (1979) Neurotoxicology 1, 349-356.11. Danscher, G. (1982) Histochemistry 76, 281-293.12. Danscher, G. (1981) Histochemistry 71, 1-16.13. Danscher, G. (1981) Histochemistry 71, 177-186.14. Liesegang, R. E. (1911) in Kolloidchemische Beihefte (Ergänzungs-

hefte zur Kolloid-Zeitschrift)-Monographien zur reinen undangewandten Kolloidchemie (Oswald, W. Ed.), pp. 1-44,Theodor Steinkopff Verlag, Dresden/Leipzig.

15. Timm, F. (1962) Beitr. Gerichtl. Med. 21, 195-197.16. Danscher, G., and Moller-Madsen, B. (1985) J. Histochem. Cyto-

ehern. 33, 219-228.17. Umeda, M., Saito, K., and Saito, M. (1969) Jpn. Exp. Med. 39,

15-58.18. Danscher, G. (1984) Histochemistry 81, 331-335.19. Danscher, G., and Norgaard, J. O. (1985) J. Histochem. Cyto-

ehern. 33,706-710.20. Norgaard, J. 0., Moller-Madsen, B., Hertel, N., and Danscher,

G. (1989) J. Histochem. Cytochem. 37, 1545-1547.21. Danscher, G., and Rungby, J. (1986) Histochem. J. 18,109-114.22. Ross, J. F., Broadwell, R. D., Poston, M. R., and Lawhorn, G. T.

(1994) Toxicol. Appl. Pharmacol. 124, 191-200.23. Lormee, S., Lecolle, D., Septier, D., Le-Denmant, D., and Gold-

berg, M. (1989) J. Histochem. Cytochem. 37,203-208.24. Ayanoglou, C., Lecolle, S., Septier, D., and Goldberg, M. (1994)

Histochem. J. 26, 213-225.25. Yamada, K (1992) Histochem. J. 25,95-106.26. Frederickson, C. J., and Danscher, G. (1990) Frag. Brain Res.

83, 71-84.27. Slomianka, L., Danscher, G., andFrederickson, C. J. (1990) Neu-

roscience 38, 843-854.28. Westbrook, G. L., and Mayer, M. L. (1987) Nature 328, 640-

643.29. Danscher, G., Juhl, S., Stoltenberg, M., Krunderup, B., and

Schroder, H. D. (1995) submitted for publication.30. Springall, D. R., Hacker, G. W., Grimelius, L., and Polak, J. M.

(1984) Histochemistry 81, 603-608.31. Fujimori, 0., and Nakamura, M. (1985) Arch. Histol. Jpn. 48,

449-452.32. Hacker, G. W., Springali, D. R., Van Noorden, S., Bishop, A. E.,

Grimelius, L., and Polak, J. M. (1985) Virch. Arch. A 406, 449-461.

33. Lackie, P. M., Hennessy, R. J., Hacker, G. W., and Polak, J. M.(1985) Histochemistry 83, 545-550.

34. Scopsi, L., and Larsson, L.-I. (1985) Histochemistry 82, 321-329.35. Van Den Pol, A. N. (1985) Science 228, 332-335.36. Bastholm, L., Scopsi, L., and Nielsen, M. H. (1986) J. Electron

Microsc. Tech. 4, 175-176.37. Schmidt, J., and Peters, W. (1987) Parasitol. Res. 73,80-86.38. Westermark, K., Lundqvist, M., Hacker, G. W., Karisson, A.,

and Westermark, B. (1987) Acta Endocrinol. (Suppl.) 281, 252.

1. Danscher, G. (1981) Histochemistry 71, 81-88.2. Danscher, G. (1991) Pharmacol. Toxicol. 68,414-423.3. Danscher, G., Hacker, G. W., Norgaard, J. 0., and Grimelius, L.

(1993) J. Histotechnol. 16.201-207.

EM IMMUNOGOLD-SILVER STAINING AUTOMETALLOGRAPHY 269

61. Hacker, G. W., Zehbe, I., Hauser-Kronberger, C., Graf, A.-H.,and Dietze, O. (1994) Cell Vision I, 71-73.

62. Zehbe, I., Hacker, G. W., Sällsträm, J., Rylander, E., and Wi-lander, E. (1992) Anticancer Res. 12,2165-2168.

63. Zehbe, I., Sällsträm, J., Hacker, G. W., Rylander, E., Strand,A., Graf, A.-H., and Wilander, E. (1994) in Modern Methods inAnalytical Morphology (Gu, J., and Hacker, G. W., Eds.), pp.297-306, Plenum Press, New York.

64. Zehbe, I., Hacker, G. W., Sällsträm, J. F., Rylander, E., andWilander, W. (1994) Cell Vision 1,20-24.

65. Hacker, G. W., Zehbe, I., Muss, W. H., Hauser-Kronberger, C.,Graf, A.-H., and Dietze, O. (1995) Cell Vision 2, 247-253.

66. Zehbe, I., Hacker, G. W., Sällsträm, J. F., Muss, W. H., Hauser-Kronberger, C., and Wilander, E. (1995) Cell Vision 2, 240-242.

67. Hacker, G. W., Danscher, G., Grimelius, L., Hauser-Kronberger,C., Muss, W. H., Schiechl, A., Gu, J., and Dietze, O. (1995) inImmunogold-Silver Staining: Methods and Applications (Hayat,M. A., Ed.), pp. 20-45, CRC Press, Boca Raton, FL.

68. Faulk, W. P., and Taylor, G. M. (1971) lmmunochemistry 8,1081-1083.

69. Roth, J. (1982) J. Histochem. Cytochem. 30,691-696.70. Roth, J. (1983) in Immunocytochemistry (Bullock, G. R., and

Petrusz, P., Eds.), Val. 2, pp. 217 -284, Academic Press, London.71. Geoghegan, W. D., Scillian, J. J., and Ackerman, G. A. (1978)

lmmunol. Commun. 7, 1-12.72. Shi, S.-R., Key, M. E., and Kalra, K. L. (1991) J. Histochem.

Cytochem. 39,741-748.73. Gu, J. (1994) in Modern Methods in Analytical Morphology (Gu,

J., and Hacker, G. W., Eds.), pp. 67-80, Plenum, New York.74. Shi, S.-R., Gu, J., Kalra, K. L., Chen, T., Cote, R. J., and Taylor,

C. R. (1995) Cell Vision 2, 6-22.75. Shi, S.-R., Cote, R. J., and Taylor, C. R. (1995) Cell Vision 2,

257-259.76. Gu, J., Choi, T.-S., Whittlesey, M. M., Slap, S. E., and Anderson,

V. M. (1995) Cell Vision 2, 257-259.77. Stefanini, M., De Martino, C., and Zamboni, L. (1967) Nature

216,173-179.78. Gu, J., De Mey, J., Moeremans, M., and Polak, J. M. (1981) Reg.

Peptides 1,465-474.79. Hainfeld, J. F., andFuruya, F. R. (1992)J. Histochem. Cytochem.

40,177-184.80. Zehbe, I., Hacker, G. W., Muss, W. H., Sällsträm, J., Rylander,

E., Graf, A.-H., Prämer, H., and Wil~der, E. (1993) J. CancerRes. Clin. Oncol. 119(Suppl.), 22.

81. Huang, W.-M., Gibson, S. J., Facer, P., Gu, J., and Polak, J. M.(1983) Histochemistry 77, 275-279.

82. Maddox, P. H., and Jenkins, D. (1987) J. Clin. Pathol. 40, 1256-1257.

83. Kummer, W., Hauser-Kronberger, C., and Muss, W. H. (1994)in Modern Methods in Analytical Morphology (Gu, J., andHacker, G. W., Eds.), pp. 187-202, Plenum, New York.

39. Skutelsky, E., Gi>yal, V., and Alroy, J. (1987) Histochemistry 86,291-295.

40. Hacker, G. W., Grimelius, L., Danscher, G., Bernatzky, G., Muss,W., Adam, H., and Thurner, J. (1988) J. Histotechnol. 11,213-221.

41. Krenacs, T., Laslik, Z., and Dobo, E. (1989) Acta Histochem. 85,79-85.

42. Krenacs, T., Krenacs, L., Boz6ky, B., and Ivanyi, B. (1990) Histo-ehern. J. 22, 530-536.

43. Krenacs, T., Ivlinyi, B., Boz6ky, B., Laszik, Z., Krenacs, L.,Rlizga, Z., and Ormos, J. (1991) J. Histotechnol. 14, 75-80.

44. Burry, R. W., Vandre, D. D., and Hayes, D. M. (1992) J. Histo-ehern. Cytochem. 40, 1849-1856.

45. Fujimori, O. (1992) Histochem. J. 24, 61-66.46. Roth, J., Saremaslani, P., Warhol, M. J., and Heitz, P. U. (1992)

Lab. Invest. 67,263-269.47. Roth, J., Saremaslani, P., and Zuber, C. (1992) Histochemistry

98, 229-236.48. Stierhof, Y.-D., Humbel, B. M., Hermann, R., Otten, M. T., and

Schwarz, H. (1992) Scanning Microsc. 6, 1009-1022.49. Gu, J., D'Andrea, M., Yu, C.-Z., Forte, M., and McGrath, L. B.

(1993) J. Histotechnol. 16, 19-26.50. Magar, R., Larochelle, R., Robinson, Y., and Dubuc, C. (1993)

Can. J. Veto Res. 57,300-304.51. Hacker, G. W., and Danscher, G. (1994) Gell Vision 1, 102-109.52. Krenacs, T., and Dux, L. (1994) J. Histochem. Cytochem. 42,

967-968.53. Takizawa, T., and Robinson, J. M. (1994) J. Histochem. Cyto-

ehern. 42, 1615-1623.54. Hacker, G. W., Hauser-Kronberger, C., Graf, A.-H., Danscher,

G., Gu, J., Grimelius, L. (1994) in Modern Methods in AnalyticalMorphology (Gu, J., and Hacker, G. W., Eds.), pp. 19-35, Ple-num, NewYork.

55. Rufner, R., Carson, N. E., Forte, M., Danscher, G., Gu, J., andHacker, G. W. (1995) Gell Vision 2, 327-333.

56. Varndell, I. M., Polak, J. M., Sikri, K. L., Minth, C. D., Bloom,S. R., and Dixon, J. E. (1984) Histochemistry 81, 597-601.

57. Liesi, P., Julien, J.-P., Vilja, P., Grosveld, F., and Rechardt, L.(1986) J. Histochem. Cytochem. 34, 923-926.

58. Hacker, G. W. (1989) in Colloidal Gi>ld-Principles, Methods, andApplications (Hayat, M. A., Ed.), Val. 1, pp. 297-321, AcademicPress, San Diego, CA.

59. Hacker, G. W., Graf, A-H., Hauser-Kronberger, C., Wirnsberger,G., Schiech!, A., Bernatzky, G., Wittauer, U., Su, H., Adam, H.,Thurner,J.,Danscher, G., and Grimelius, L. (1993) Chinese Med.J. 106, 83-92.

60. Hacker, G. W., Zehbe, I., Hauser-Kronberger, C., Gu, J., Graf,A.-H., and Dietze, O. (1994) Gell Vision 1, 30-37.~

ILanddskrank8It8n...It,n ,.libO.Path.-Anet. .-lnat./EM-Labef

Eing. 0 1. OKT. 1996

"