PEG Embedding for Immunocytochemistry

0022-1554193/$3.30 The Journal of Hinochemistry and Cytochemistry Copyright @ 1993 by The Histochemical Society, Inc.

Vol. 41, NO.3, pp. 455-463. 1993 Printed in US.A.

Technical Note

I

PEG Embedding for Immunocytochemistry: Application to the Analysis of Immunoreactivity Loss During Histological Processing' PAUL KLOSEN,' XAVIER MAESSEN, and PHILIPPE VAN DEN BOSCH DE AGUILAR Laboratory of Cellular Biology, Catholic University of Louvain,

Belgium.

Received for publication April 24, 1992 and in revised form October 2, 1992; accepted October 6, 1992 (2T.2661).

I We have developed a protocol for the production and longterm storage of polyethylene glycol (PEG) sections for immunocytochemistry. Sections obtained by this protocol allow immunolabeling for many different antigens, such as intermediate filaments, maaophage markets, or neurotransmitter enzymes. Standard histological staining can also be performed on these sections. This ftation-embedding system may therefore be of interest for histopathology of rare specimens, as well as for experimental research. Multiple labeling can be performed either on the same section or on

Introduction The major problem in immunocytochemistry is the simultaneous conservation of antigenicity and structural organization. In histological processing, two stages are likely to reduce or destroy antigenicity of the molecules of the tissue: fixation and embedding. The problem of fixation has been addressed by many studies. A variety of fixation procedures exist, and usually one of these will provide adequate conservation of antigenicity. For the most part, fixation cannot be avoided if tissue structure is to be preserved at a level suitable for optimal localization of immunoreactivity. Embedding, the second antigenicity reducing factor, can be avoided. Sectioning devices such as the vibratome or cryostat circumvent this stage. However, these techniques provide relatively thick sections or the specimens (e.g., frozen tissue) are difficult to store for long periods of time. The standard paraffin-embedding procedure used in histopathology is believed to cause loss of antigenicity by two unavoidable stages, dehydration and clearing of the tissue, and by embedding in paraffin at high temperature (6).

'

Supported by FNRS and FRFC grants and by the National Lottery of Belgium. PK is the recipient of a research grant by the Ministry of Cultural Affairs of the Grand Duchy of Luxembourg. This article is published within the framework of the Commission of European Communities Concerted Action on Cellular Aging and Diseases (EURAGE). Correspondence to: Dr. P. Klosen, Laboratory of Cellular Biology, Bitiment Claude Bernard, Place Croix du Sud, 5 , B-1348 Louvain-la-Neuvc, Belgium.

consecutive thin sections, thus allowing a more thorough analysis of precious experimentalmaterial. We compare the advantages of PEG vs cryostat or vibratome sections. This protocol has been used to study the inactivation of antigenicity by p a d i n embedding. We have identified the infiltration by paraffin as the antigenicity inactivating step, not dehydration or high temperature as generally thought. ( j Hisrochem Cytochem 41:455463, 1993) KEY WORDS:Immunocytochemistry;Embedding; Polyethylene

gly-

col; Antigenicity; Histology.

Quick-freeze techniques and ultracryotomy can provide tissue for immunocytochemistrywithout fmtion or embedding, but these techniques require expensive hardware usually not available in most laboratories. To circumvent some of the problems of paraffin embedding, polyethylene glycol (PEG) has been used since the 1950s under the name of Carbowax (4,5). In the 1970s, PEG was introduced for histochemistry and immunocytochemistry (13-16,21). PEG shows several advantages over paraffin. It is water and alcohol miscible, thus avoiding the need for clearing agents. Furthermore, it melts, depending on the molecularweight of PEG, at well below the usual 60°C or higher temperatures used in paraffin embedding. Finally, by adjusting the molecular weight of the PEG, variable hardness of the blocks can be obtained, thus allowing optimization of the procedure for different kinds of tissues. However, although sectioning of PEG-embedded tissue is as easy as paraffin sectioning, the mounting of these sections is quite difficult (IJ3). Another major problem presented by these sections is the conservation of immunoreactivity during drying and storage, since the embedding medium (the PEG) is lost during mounting (13,14,21). We have devised a procedure to provide PEG sections for immunocytochemistry, as well as for long-term storage, by combining developments by other authors with our own. Tissue prepared by this technique allows immunolabeling for many different antigens, providing an ideal material for co-localizationstudies by different multiple-labeling techniques. Furthermore, the possibility of long-term storage of tissue blocks and sections may be useful for histopathological studies. 455

456

We also identified infiltration with paraffin as the major antigenicity inactivating step in paraffin embedding, not dehydration or high temperature embedding as formerly thought. Preliminary results of this study have been published in the form of an abstract (11).

Materials and Methods Fixation. Animals were fixed by transcardiac perfusion after terminal ether anesthesia. Immediately before perfusion, 0.2 ml of heparin (5000 UIlml) (Liquemine; Roche, Nutley, NJ) and 0.5 ml of 1% NaNO2 were injected directly into the left ventricle. Perfusion was initiated with 50-100 ml of PBS. After the PBS rinse, the animal was fixed by 100 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.2) followed by 250 ml of 4% paraformaldehyde in 0.1 M borate buffer (pH 11) (3). The brains and other organs were dissected and cut into coronal slices (E. 5 mm thick) for post-fixation in the perfusion fixative (4-6 hr). Some additional tissue specimens were fixed by Methacarn or Clarke's fixative. For immersion-fixationwith formalin, Lillie's formol-calcium acetate provided the best results. Some tissue specimens were either processed directly on a vibratome or cryoprotected, frozen, and cut on a cryostat. Paraffin embedding and sectioningwere performed using standard protocols. Dehydration and Embedding. The embedding procedure was adapted to rat brains from the procedure used by Clayton and Alvarez-Buylla (7) for in situ hybridization. The brains were rinsed overnight with 50% ethanol. Thereafter, dehydration was performed by 70% ethanol (2 hr), 95% ethanol (twice for 2 hr), and 100% ethanol (2 hr). The embedding in PEG was performed at 46°C and was initiated by PEG 1000 overnight, followed by PEG 1000 (3 hr), PEG lOOOlPEG 1500 mix (3 hr), and finally PEG 1000lPEG 1500 mix (overnight). The PEG lOOOlPEG 1500 mix was optimized to each procedure and to the cutting conditions (temperature,humidity). We tested PEG from different sources: Janssen Chimica (Beerse. Belgium), Fluka (Buchs, Switzerland), and Sigma (St Louis, MO). To cast the blocks, fresh PEG lOOOlPEG 1500mix was poured into cardboard molds and the tissue slices were then positioned. The mold with the tissue slices was returned to the oven for 30 min at 46°C. Finally, the mold was retrieved, the pieces reoriented to their final position, and the block hardened at room temperature on a glass plate to ensure a homogeneous heat transfer. The PEG blocks were stored at room temperature in air-tight boxes. Several brains were embedded in PEG without prior ethanol dehydration. After fixation, these brains were rinsed overnight in several changes of 0.1 M phosphate buffer (pH 7.2). They were then progressively infiltrated with PEG 1000-phosphate buffer mixes (30%, SO%, 70%, 100% PEG 1000, 2 hr each). Thereafter, they were processed according to the above protocol. sectioningand Mounting of the Sections. The PEG blocks were mounted on wooden blocks by simply melting some PEG on the wood and the lower surface of the block and then gently pressing the molten surfaces against one another. The trimming of PEG blocks should be done with a warm scalpel blade, because the blocks are harder and more brittle than paraffin blocks and might break. The PEG blocks were sectioned on a standard rotary microtome. Thickness ranged from 3 to 20 Fm. The sections were collected on cardboard and mounted the same day. For long-term storage, the ribbons were put in a cardboard box with some dessicant. Great care had to be taken to avoid humidtfying the ribbons, which resulted in disruption of the sections. Thick sections (8-20 pm) were processed as free-floating sections. These sections were put on TBS buffer to dissolve the PEG. They were then rinsed three times for 10 min with TBS and finally processed for endogenous peroxidase inactivation and immunocytochemistry.After sectioning, a paraffin film was cast over the sectioned tissue by dipping the block into liquid paraffin. This treatment protected the tissue from humidity. The paraffin cast could easily be separated from PEG with tweezers before further sectioning.

KLOSEN, MAESSEN, VAN DEN BOSCH DE AGUILAR

Mounting on slides of thin sections (2-10 pm) was performed in translucent mounting baths put on black cardboard. We tested several spreading solutions: ultrapure water, 0.1% gelatin, 1/10 Finks section glue (9), 5 and 10% sucrose, 1% polyvinylpyrrolidone (PVP-40;Janssen), and 1% methocel. Of these solutions, 5 % sucrose proved to be optimal. The slides were coated with a poly-L-lysine solution (Sigma). The sections were put on the spreading solution with a forceps, where the PEG dissolved instantly and the section spread. The sections should be lowered on the solution by a frank, continuous movement to avoid folding. They are left on this solution for 5-10 min before mounting. The section is brought into contact with a slide held at an angle over 45". The slide was then withdrawn without raising. Afterwards, the slides are dried around the sections. Care must be taken not to leave large drops of the spreading solution on the sections. The solution has to form only a thin film on the sections. They are then put on a drying rack at 37°C. As soon as the sections are dry (10-15 min), the slides are put into TBS for a minimum of 10 min. We usually inactivated endogenous peroxidase by the method of Li et al. (12) before immunocytochemistry or storage. The sections were soaked in the inactivating solution (0.1% NaN3, 0.3% H202 in TBS) for 30 min. After inactivation, the sections were rinsed twice for 10 min with TBS. They were then either used immediately for immunocytochemistry or processed for storage. Storage of Mounted Sections. Three storage procedures allowed conservation of immunoreactivity in PEG sections. For short-term storage, they can be stored in a staining jar in TBS + 0.1% sodium azide. For long-term storage, the slides were coverslipped with PEG 400 as a mounting medium. They were then stored at - 2 0 T , at which the PEG 400 solidifies. Before immunocytochemistry,the sections were thawed, the coverslips removed, and the PEG 400 eluted for 10 min in TBS. Alternatively, they were stored in a staining jar in the cryoprotectant solution described by Watson et al. (21) at -2O'C. This cryoprotectant solution consists of 0.1 M PBS, pH 7.2 (50% vlv), sucrose (30% w / ~ )poly, vinylpyrrolidone (1% w/v) (PVP-40; Sigma), and ethylene glycol (30% vlv). Before immunostaining, the slides were rinsed twice for 10 min in TBS. Experimental Procedures. To determine the stage of antigenicity inactivation, several brains, after alkaline formaldehyde fixation, were cut into coronal slices. Adjacent slices from the same brain were embedded in PEG or paraffin or were processed by procedures intermediate to PEG and paraffin embedding. 1. Alcohol dehydration and embedding in PEG at 6 2 T . 2. Alcohol dehydration, butanol clearing (2 days), and embedding in

PEG at 46°C. 3 . Alcohol dehydration, butanol clearing (2 days), and embedding in PEG at 62°C. This procedure is identical to the paraffin embedding, except that PEG is substituted for paraffin for infiltration and embedding. Furthermore, some mounted PEG sections were taken through xylene and alcohol baths to simulate the dewaxing and rehydrating of paraffin sections. Immunocytochemistry. To check the extent of antigenicity conservation, we have tested a large panel of polyclonal (PAb)and monoclonal antibodies (MAb)against antigens ranging from intermediate filaments to macrophage markers (Table 1). Immunocytochemistry was performed using standard immunoperoxidase techniques. For floating sections, 0.05% Triton X-100 or NP-40 was included in all rinse or incubation buffers.

Tissue Processing and Section Mounting Tissue shrinkage with the described procedure was minimal, uniform, and reproducible. Conversely, when tissue was directly de-

457

PEG EMBEDDING FOR IMMUNOCYTOCHEMISTRY

Table 1. Antibodies tested in this stud’ Antibody

TPb

source

RS-18, 147 (phosphorylated NF-H)

NE 14 (phosphorylated NF-H)

MAb MAb

NE 52 (phosphatase insensitive NF-H) 1G2 (phosphorylated NF-M) NN 18 (phosphatase insensitive NF-M) NR 4 (NF-L)

MAb MAb MAb MAb

a-tubulin, P-tubulin

MAb

Dr. B.H. Anderton Boehringer (Mannheim, Germany) Boehringer Dr. M. Vitadello Boehringer Boehringer, Dako (Santa Barbara, CA) Amersham (Poole, UK)

EDI, ED2, ED3 (macrophage markers)

MAb

Seralab (Sussex, UK)

Synaptophysin SY38

MAb

Choline acetyltransferase

MAb

Tyrosine hydroxylase

PAb

Choline acetyltransferase

PAb

Somatostatin 28

PAb

Gamma (Liege, Belgium) Boehringer, Chemicon (Temecula, CA) Institut J. Boy (Reims, France) Chemicon, Biogenesis (Boumemouth, UK) CRB (Northwich, UK)

GFAP Keratin

PAb PAb

MAP2

PAb

Tau

PAb

Dako

Dako Dr. J.P. Brion Dr. J.P. Brion

PEG sectio&

++ ++

+++ +++

*

++ ++ +++

to

++

paldin secdonsc

+d

+ + +

-

-

+

+++

* *

+++

++

-

-

+++

++

to

+ + +d

-

+++

++

+++

++ +

+ + + to + + + d ++

* *

4% alkaline paraformaldehyde-fmed tissue, embedded as in Materials and Methods. Only the quality of immunolabeling is rated. Morphological quality was always far superior in PEG sections. MAb, monoclonal antibody; PAb, polyclonal antiserum. - no labeling; f , faint labeling with background; + , faint labeling; + + , good labeling: + + + , excellent labeling. The better results were obtained on floating sections.

.

hydrated by PEG-buffer mixtures, tissue shrinkage was very high (&SO%). The tissue blocks obtained by PEG embedding can usually be handed in the same way as paraffin-embedded tissue blocks. Sectioning of PEG blocks on a rotary microtome can be easily performed between 2 and 20 pm. Thicker sections (up to 40 pm and more) can be obtained from pure PEG 1000 blocks on a sliding microtome with a slightly tilted knife. Good ribbons were easily obtained at thicknesses ranging from 3-10 pm. However, they were more brittle than paraffin ribbons and more sensitive to static electricity. They should be handled very carefullywith the help of a forceps and a fine artist’sbrush. Thicker sections did not form ribbons. However, this did not complicate their handling, since they were mostly processed as free-floating sections. The paraffin cast protected PEG-embedded tissue blocks efficiently against uptake of ambient humidity. It also allowed good conservation of both morphology and immunoreactivity. The sections shown in Figures 2b, 2d, and 2f were taken from a block stored for 2 years under such a paraffin cast. The final embedding mixture of PEG 1000/PEG 1500 had to be adapted to ambient conditions during sectioning. In addition, the quality of PEG from several sources varied, and preliminary trials should determine the optimal mixture. A 30:70 mixture was usually adequate for sectioning between 4 and 15 pm. For thinner sections and sectioningon very warm summer days, pure PEG 1500 should be used. Tissue blocks could be easily re-embedded in harder

or softer PEG by simply melting the block at 46°C. putting the tissue into new PEG overnight, and casting a new block. Mounting of sections was performed with sections from ethanoland PEG-dehydrated tissue. Surprisingly, in PEG-dehydrated tissue all immunoreactivities were lost during the drying to ensure section adherence. We have no explanation for this observation. The protective effect of ethanol was tested by partially dehydrating tissue with ethanol before switching to PEG. This effect appears progressively with increased ethanol concentration (results not shown). However, sections from tissue partially dehydrated with ethanol stretched out during rehydration of the sections and were lost during immunocytochemistry. Our initial work was performed with Fink’s glue as the spreading solution (11). However,later batches of this preparation resulted in increased section loss during immunocytochemistry. Therefore, we tested several spreading solutions for antigenicity conservation and section adherence. Ultrapure water resulted in very well-attached sections, but immunolabeling was uneven on these sections. In particular, areas that dried first displayed increased loss of immunoreactivity. Therefore, several agents were tested as protective agents during the drying of the sections. Polyvinylpyrrolidone and methocel interfered very much with section adherence. With these agents the sections did not stick to the slides during mounting, and most of them were lost during immunocytochemistry. Gelatin and 10% sucrose interfered slightly with section adherence but provided good immunolabeling. Five percent sucrose resulted in well-spread sec-

KLOSEN, MAESSEN, VAN DEN BOSCH DE AGUILAR

458

~

'

.*.I

.I-

a.

0 ' I -

,

'I

*

I.

459

PEG EMBEDDING FOR IMMUNOCYTOCHEMISTRY

tions which adhered very well to the slides. Provided that care had been taken to avoid large drops of sucrose on the slides during drying, there was no detectable background on the slides. Morphological conservation afthe tissues was excellent and equal or superior to paraffin sections (Figures la, Ib, 3c, and 3d). Section thickness could easily be reduced to about 50% as compared to paraffin sections. As noted by many authors, these sections can be used for most standard stains provided that they are compatible with the formaldehydefixation. We did note that ferric hematoxylins had a tendency to stain myelin in addition to nuclei. The major problem of the PEG sections was the loss of the embedding medium during mounting. This made the mounting of very small sections (e.g., spinal ganglia) almost impossible. Furthermore, loose tissue (e.g., pancreas) often disrupted on spreading and usable sections were difficult to obtain. Compact tissues (brain, kidney, spleen, liver) usually gave good sections. Sections from stomach or intestines were more difficult to mount but could be obtained with some practice. We tested the procedure proposed by Gao and Godkin (10) for mounting PEG sections. However, the large sections we used did not spread completely on the agarose blocks. Furthermore, section adherence to poly-L-lysine-coated or silanated slides was poor with this procedure.

Immunocytochemistry on Mounted Sections Immunolabeling on PEG sections could be performed by the same standard techniques as those used on paraffin sections. Most antibodies we tested worked fine. Among the exceptions were some anti-neurofilament clones, which seemed to be very sensitive to formaldehydefixation. These antibodiesprovided good results once Clarke’sor Methacarn fixative was used, with either paraffin or PEG embedding. Most antibodies gave results far superior to those obtained on the equivalent paraffin-embedded tissue (Figure 1). Some antisera, such as the anti-choline acetyltransferase (ChAT) antiserum, did not work at all on paraffin-embedded tissues. Other antibodies produced a high background on paraffin sections. Immunolabeling on PEG sections produced “crisper” images. In particular, labeled fibers were smoother and better conserved than in paraffin sections (Figures 1and 3). Furthermore, PEG sections displayed a very good resolution of labeled structures, as typified by immunolabeling for tyrosine hydroxylase (Figure Id). Tissues other than brain could also be used for immunolabeling. As previously stated, some of these tissues presented problems for section mounting, owing to the disappearanceof the PEG matrix on the spreading solution. This problem could be overcome by embedding these tissues, like pancreas, in liver fragments or in segments of intestine, which form a ring around the section frag-

ments and prevent their disruption. Thus, pancreatic islets could be labeled for somatostatin 28 (Figure le).

Immunocytochemistry on Floating Sections PEG sections were also immunolabeled using floating sections. Sections of 10-12 pm were sturdy enough to withstand an immunolabeling procedure without extensive precautions. We have handled floating sections down to 8 pm. Compared with cryostat or vibratome sections of identical thickness, PEG sections proved to be mechanically more resistant. When floating sections were used, the tissue did not have to be dehydrated by ethanol for antigenicity conservation. On the other hand, sections from PEG-dehydrated tissue presented the same problems for the penetration of immunoreagents as vibratome or cryostat sections and had to be permeabilited by buffered ethanol or detergents. Furthermore, they had a greater tendency to curl up than sections from ethanoldehydrated tissue. We did not investigate the conservation of ultrastructure in these sections. Floating sections from ethanol-dehydratedtissue presented many fewer penetration problems and were slightly easier to handle. We noted that the buffer from which these sections are mounted on slides after immunolabeling determined the final site of the sections. Therefore, for quantitative work, it is important to always use the same buffer, to avoid differential shrinkage of the sections. Most antibodies provided similar results on floating sections as on slide-mountedsections. However, some antibodiesworked much better on floating sections (Figure lf). An increase in the signalto-noise ratio was also observed with the antisera to MAP2 and choline acetyltransferase(ChAT). Most other antibodies displayed an increase in labeling quantity on floating sections. To investigate this effect, adjacent 10-pm sections were either mounted on slides before immunolabeling or processed as floating sections. Adjacent 4-pm sections were also slide-mounted and used for comparison. As shown in Figures 2a-2d, an increase in section thickness from 4 to 10 pm resulted, as expected, in an increase in the number of labeled axons and astroglialprocesses in slide-mounted sections. On 10-pm floating sections, a further increase in the number of labeled fibers was apparent (Figures 2e and 2f). With GFAP immunolabeling, no additional astrocytes seemed to be labeled. It appeared rather that, whereas on slide-mountedsections individual astrocytes appeared to be partially labeled, on floating sections the whole extent of the astrocytes contained within the section was labeled.

Inactivation of Antigenicity The solubility of PEG in water, ethanol, and butanol enabled us

Figure 1. (a) Perikaryal labeling of large neurons in the mesencephalic nucleus of the trigeminal nerve with an antibody to phosphorylated NF-H (147) on paraffin sections. Neurites in the vicinity of the cell bodies appear fragmented and can only be distinguishedas dot-like structures. Furthermore, the cell bodies appear shrunken. (b) lmmunolabeling on PEG sections of the mesencephalon with antibody 147 to phosphorylated NF-H. Around the cell bodies, thin labeled neurites can clearly be distinguished. (c) In PEG-embedded cerebellum, antibody 147 labels the basket around the cell body of the Purkinje cells. Note the fine detail obtained in PEG sections. (d) Tyrosine hydroxylase-positive fiber network around a lateral septal neuron on PEG sections. (e) Anti-somatostatin 28 labeling on a pancreatic islet. Note the artifactually dilated extracellular space around the islet. In the islet itself, the extracellular space is not dilated. (f) On 10-pm floating PEG sections, antibody N52 to phosphatase-insensitiveNF-H epitopes labels cortical pyramidal neurons in a network of mons. The labeled mons of these neurons can sometimes be detected (arrows). On slide-mounted sections, labeling with this antibody was barely satisfactory. Bars: a+, e, f = 50 pm; d = 25 pm.

. ' .

-

-J

1

\

..

,

, .

,

:I

',

Figure 2. Effect of section thickness and slide mounting on immunostainingwith (a, c, e) antibody RS18 (phosphorylated NF-H) in the cerebellar cortex and (b, d, 1) an antiserum to GFAP in the hippocampus. (a, b) Slide-mounted 4-Mm sections; (c, d) slide-mounted 10-pm sections; (e, 1) IO-pm floating sections. Bar = 50 wn.

460

Figure3. Comparisonof antigenicity in (a, c) paraffinand (b, d) PEG embedded material. (a) Anti-ChAT labelingin paraffin-embeddedmaterial sometimes displays very faintly labeled perikarya in the striatum, but no fiber labeling. (b) In PEG-embedded material, perikarya are strongly labeled, while a fine fiber network is also demonstrated in the striatum. (c) GFAP-positive astrocytes in the striatum. (d) GFAP labeling on PEG-embedded material reveals astrocytes with smooth processes. Note also the increased number of labeled processes and the labeled astrocyte endfeet around a blood vessel (arrows). Bar = 50 pm.

461

462

to compare this embedding medium directly with paraffin under identical conditions of dehydration,clearing, and embedding temperature. Four antibodies were selected for this study: clone 147 to phosphorylated neurofilaments, clone ED1 to macrophages, an antiserum to choline acetyltransferase(ChAT), and an antiserum to GFAP. Only the results for the ChAT and GFAP antisera are shown in Figure 3. The ChAT antiserum usually did not work on paraffin sections (Figure 3a), whereas on PEG sections it worked fine, even when tissue had been cleared by butanol and embedded at 62°C (Figure 3b). The GFAP antiserum worked properly on both paraffin and PEG sections. However, morphological quality was clearly increased in PEG sections. Similar results were also observed with clone 147 to phosphorylated NFH. Antibody ED1 worked properly only on PEG sections, where it provided results identical to those reported by Dijkstra et al. (8) with cryostat sections. Modifying the PEGembedding procedure to mimic some factors of paraffin embedding (temperature, clearing) did not affect the immunolabeling for either neurofilament, GFAP, or ED1 immunolabeling. Almost all antibodies that worked on formaldehyde-fixed tissue with vibratome or cryostat sectioningperformed well on slidemounted or floating PEG sections. Only two MAb to ChAT did not work at all on PEG sections, whether the tissue was dehydrated by ethanol or PEG. It must be noted that ChAT seemed to be particularly sensitive to ethanol or dehydration, since enhancement of antibody penetration by buffered ethanol notably reduced ChAT immunolabeling on vibratome or cryostat sections with these MAb. Finally, storage of PEG sections is a stage likely to reduce immunoreactivity,as the protective PEG embedding matrix is lost during mounting. Storage in buffer added with azide was mostly used only for a maximum of 1 month. However, we were able to detect MAP2 without appreciable loss of immunoreactivity in sections stored for 9 months at 4°C under these conditions. Optimal results were obtained with storage at -20°C with PEG 400 or Watson’s cryoprotectant. ChAT could still be detected in these sections after 2 years of storage.

Discussion We have devised a PEG embedding and sectioning protocol that allows optimal conservationof immunoreactivity.Among the most significant improvementsof existing procedures are the mounting medium for thin sections and the long-term storage conditions for slide mounted sections and PEG blocks. Furthermore, we have explored the different uses of slide-mounted and free-floating sections for immunocytochemistry.With these techniques, antibodies that allow immunolabeling on paraformaldehyde-fiued vibratome or cryostat floating sections can be used for immunocytochemistry on PEG sections. The advantages of this procedure are multiple. First, the hardness of PEG can easily be adapted to the tissue and section thickness by simply using PEG of different molecular weight. PEG can thus be cut thinner than paraffin, even without cooling of the block or the knife. Second, thin and thick sections can be cut from the same block to perform immunolabeling on either slide-mounted sections or free-floating sections. For studies of neuronal networks using neurotransmitter markets, thick sections may be the method of choice,

KLOSEN, MAESSEN, VAN DEN BOSCH DE AGUILAR

because the analysis and tracing of neuronal fibers is easier on such sections. Thinner sections can be used to study different antigens in small structures or to perform double labeling studies on adjacent sections. It should, however, be noted that a similar methodology has also been proposed for paraffin sections (17) and slidemounted cryostat sections (2). Third, the PEG embedding seems to interfere with immunoreactivity much less than paraffin. This results in an extensive spectrum of antigens that can be detected in PEG-embedded material. Furthermore, PEG apparently works well for in situ hybridization (7). Its potential for histochemistry has been exploited for several years (15,16,20). This versatility may make PEG a method of choice for histopathology and research. Indeed, thinner sections, combined with the number of techniques compatible with this embedding matrix, allow a more thorough analysis. PEG blocks are also easier to store than the frozen tissue fragments usually required for special techniques. Therefore, reinvestigation of important material can be easily performed, should new questions arise. On the other hand, PEG also has disadvantages. The main problem with PEG is its hydrophilicity. Although the storage of PEG blocks is easier than that of frozen tissue, it is more tedious than the storage of paraffin blocks, as PEG blocks have to be stored in air-tight boxes to prevent hydration. However, the casting of a paraffin film over the block reduces this problem considerably. Second, although with some practice PEG sections can be easily obtained, the handling and mounting of these sections remains more difficult than that of paraffin sections. This is mainly due to the loss of the embedding material during mounting. This loss also creates a third problem: the storage of the sections. Whereas paraffin sections can easily be stored for years, dry and at room temperature, PEG sections must be stored in buffers in refrigerators or freezers. However, we believe that these disadvantages are outweighed by the extraordinary conservation of morphology and immunoreactivity in PEG sections, as well as by the versatility of this embedding medium. PEG enabled us to study the loss of antigenicity during histological processing. Conservation of antigenicity is essential for immunocytochemistry. The aim of histological processing is optimal conservation of tissue morphology. These two goals are sometimes incompatible, and a compromise between tissue morphology and antigenicity must be found. To optimally design this compromise, all intervening factors must be known. Fixation is clearly the main reducing factor of antigenicity. Using PEG embedding, we have been able to show that another antigenicity reducing factor is infiluation with paraffin. Dehydration, clearing, or high temperature alone are not responsible for the reduced antigenicity of paraffin-embedded tissue. The exact mechanism of antigenicity inactivation by paraffin is not known. It could be due to some components of the paraffin that may adsorb strongly to the tissue and may not be removed by dewaxing with xylene. They would thus obscure certain antigens. However, other mechanisms are possible. PEG may be interesting as a reference material in the development of special brands of paraffin for immunocytochemisuy. Further comparisons between PEG and paraffin may allow the identification of the paraffin components that inactivate or block antigenicity. Our study also shows that the role of temperature in the loss

463

PEG EMBEDDING FOR IMMUNOCYTOCHEMISTRY

of antigenicity in embedding has been largely overestimated (6). We did not observe a difference between PEG embedding at 46'C and 62'C. Furthermore, recent work by several authors indicates that antigenicity may actually be retrieved by boiling or autoclaving the sections (18,19).However, thermal inactivation may exist for antigens other than those tested in our study. The only difference we observed between vibratome sections and PEG sections was ChAT immunoreactivity. In particular, the MAb to ChAT appeared to be sensitive to ethanol or dehydration of the tissue. Most intriguing was the loss of immunoreactivity during section drying in tissue that was dehydrated with PEG during embedding. We have no explanationfor this effect or for the progressive protection obtained by partial ethanol dehydration. Obviously, this protection by ethanol is not absolute, as mounting with ultrapure water resulted in immunoreactivity loss in those areas that dried fastest. In addition, these sections could not be kept dry for a long period of time without losing most of the immunoreactivity. Therefore, a protective agent had to be included in the mounting solution. This observation clearly must be investigatedfurther, and better protective agents may thus be developed. Finally, our comparison between slide-mounted sections and free-floating sections shows that some immunoreactivity may be lost by mounting the sections on slides. Another explanation for this observation may be that part of the section becomes inaccessible to the antibodies. A first observation in favor of this hypothesis is that GFAP labeling of astrocytes seemed to affect the whole extent of these cells in floating sections, whereas in mounted sections these cells appeared as one would expect of thinner sections. In addition, in section areas that detached during processing, the immunolabeling resembled that of floating sections. A similar effect could also be observed on paraffin sections in areas under which air bubbles had formed during mounting on slides. Again, further investigation of these observations may lead to the development of techniques or materials that prevent this effect.

Acknowledgments The monodondantibodies 147 andRS 18 were Rindi'y donatedby DrB.H. Anderton, antibody 1G2 by Dr M. Vitaa'eh'o, andthe antisera against MAP 2 andMAP tau by DrJl? Brim.

Literature Cited 1. Bard JBL, Ross ASA: Improved method for making high-affinity sec-

tions of soft tissue embedded in polyethylene glycol (PEG): its use in screeningmonoclonal antibodies.J Histochem Cytochem 341237, 1986 2. Barthel LK, Raymond PA Improved methcd for obtaining 3-pm cryosections for immunocytochemistry.J Histochem Cytochem 38:1383, 1990 3. Berod A, Harunan BK. Pujol F Importance of fmtion in immunocytochemistry: use of formaldehyde solutions at variable pH for the localization of tyrosine hydroxylase.J Histochem Cytochem 292344, 1981 4. Blank H: A rapid embedding technic for histologic sections employing a water soluble wax. J Invest Dermatol 12:95, 1949

5 . Blank H, McCarthy PL A general method for preparing histologicsections with a water-soluble wax. J Lab Clin Med 36:776, 1950 6. Brandtzaeg P: Tissue preparation methods for immunocytochemistry.

In Bullock G, Petrusz P, eds. Techniquesin immunocytochemistry.Vol I. London, Academic Press, 1982, 1 7. Clayton DF, Alvarez-Buylla A: In situ hybridization using PEGembedded tissue and riboprobes: increased cellular detail coupled with high sensitivity.J Histochem Cytochem 37:389, 1989 8. Dijkstra CD, Dopp EA, Joling P, Kraal G The heterogeneity of mononuclearphagocytes in lymphoid organs: distinct macrophage subpopulations in the rat recognized by monoclonal antibodies EDI, ED2 and ED3. Immunology 54589, 1985 9. Fink S Some new methods for affiing sectionsto glass slides. I. Aqueous adhesives. Stain Tech 62:27. 1987 10. Gao K, GodkinJD: A new method for transfer of polyethylene glycol-

embedded tissue sections to silanated slides for immunocytochemistry. J Histochem Cytochem 39537, 1991 11. Klosen P, van den Bosch de Aguilar Ph: Polyethylene-glycol sections

for immunocytochemistryof neural antigens: a reappraisal. EurJ Neurosci 3 (suppl):249, 1990 12. Li C-Y, Ziesmer SC, Lazcano-Villareal 0: Use of azide and hydrogen

peroxide as an inhibitor for endogenousperoxidase in the immunoperoxidase method. J Histochem Cytochem 35:1457, 1987 13. MazurkiewiczJE, Nakane PK Light and electron microscopic localiza-

tion of antigens in tissues embedded in polyethylene glycol with a peroxidase-labeled antibody method. J Histochem Cytochem 20:969, 1972 14. Pickel VM, Joh TH. Field PM, Becker CG, Reis DJ: Cellular localiza-

tion of tyrosine hydroxylase by immunohistochemistry.J Histochem Cytochem 23:1, 1975 15. Sandstrom B: Preservation of some histochemically demonstrable en-

zyme activities in polyethyleneglycol-embedded liver tissue during prolonged storage. Acta SOCMed Uppsala 75:274, 1970 16. Sandstrom B, Westman J: Non-freezing light- and electron microscopic

enzyme histochemistry by means of polyethylene glycol embedding. Histochemie 19:181. 1969 17. Schrell U, Sofroniew MY Use of serial 1-2 pm paraffin sectionsin neu-

ropeptide immunocytochemistry for sequential analysis of different substances contained within the same neurons. J Histochem Cytochem 30:512, 1982 18. Shi SR, Key ME, Kaka KL: Antigen retrieval in formalin-fixed,paraffin-

embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J Histochem Cytochem 39:741, 1991 19. Shin RW, Iwaki T, Kitamoto T, Tateishi Hydrated autoclavepretreatment enhances TAU immunoreactivit n formalin-fixed normal and Alzheimer's disease brain tissues. Lab Invest 64:693, 1991 20. Smithson KG, Macvicar BA, Hatton GI: Polyethylene glycol embed-

ding: a technique compatible with immunocytochemisuy, enzyme histochemistry,histofluorescenceand intracellular staining.J Neurosci Methods 7:27, 1983 21. Y%?" E,Wiegand SJ, Clough RW, Hoffman GE: Use of cryoprotec-

tant to maintain long-term peptide immunoreactivityand tissue morphology. Peptides 7:155, 1986