Abstract
The advantages of using immunoelectron microscopy in amyloid research and surgical pathology for the classification of amyloid deposits are well documented. The aim of this study was to improve single-labeling postembedding immunostaining by testing different antigen retrieval (AR) techniques. Etching and AR procedures were applied to sections from aldehyde-fixed and Epon-embedded autopsy specimens of patients who had suffered from generalized AA amyloidosis, systemic senile ATTR amyloidosis, or generalized κ-light chain amyloidosis. The procedures used were no AR, H2O2, saturated aqueous sodium metaperiodate (mPJ), heating in deionized water (dH2O), heating in sodium citrate buffer (SCB), heating in EDTA (each 91C, 30 min), and combinations of etching and heating. Little effect was evident after treatment with H2O2, mPJ, and heating in dH2O, but the signal density markedly increased after heating in 1 mM EDTA. Heating in SCB affected immunolabeling with anti-transthyretin and anti-κ-light chain, whereas no effect was achieved for immunolabeling with anti-AA amyloid. We concluded that AR may significantly improve immunostaining of specimens that have undergone conventional fixation and embedding procedures for electron microscopy. The effect of AR on the detection of amyloid fibril proteins was probably mediated in part through chelation or binding of metal ions by the AR medium.
AMYLOIDOSES ARE characterized by local organ-limited or generalized proteinaceous deposits of autologous origin (Glenner 1980a,b). Depending on the fibril protein, pattern of distribution, and progress of disease, severe and lethal complications may arise (Glenner 1980a,b). The three characteristics of amyloid deposits are (a) a typical green birefringence in polarized light after Congo Red staining, (b) the presence of nonbranching linear fibrils of indefinite length with an approximate diameter of 10–12 nm, and (c) a distinct X-ray diffraction pattern (Glenner 1980a,b; Serpell et al. 1997). The origins of amyloid are diverse, and to date 18 different fibril proteins have been described. The current classification of amyloid and amyloidoses is based on the detection of the amyloid fibril protein, with an individual amyloid syndrome being characterized by the deposition of the protein (Kazatchkine et al. 1993). Most fibril proteins represent a certain underlying disease that is linked to the particular amyloid syndrome (Glenner 1980a,b).
The diagnosis and classification of amyloidosis require histological evidence using either light or electron microscopy, both of which have an identical specificity (Röcken et al. 1996). Histological examination of an amyloidotic tissue specimen has not been replaced by any clinical or laboratory test as yet. Subsequent characterization of the fibril protein is mandatory to link the deposits with an underlying disease and thus facilitate appropriate medical treatment. The fibril protein can be characterized by biochemical procedures such as high-performance liquid chromatography, electrophoresis, immunoblotting, immunodiffusion, or by immunohistochemistry (Linke 1985; Linke et al. 1986; Westermark et al. 1986; Yakar et al. 1995; Röcken et al. 1996; Kaplan et al. 1993,1997).
Immunotyping of amyloid presents several problems. First, the amyloid fibril proteins may differ from their autologous precursors in their tertiary structure resulting from amyloid formation (Serpell et al. 1997). Second, several precursor proteins undergo proteolysis before or after amyloidogenesis, resulting in loss of antigen (Husby et al. 1994; Saido 1998). Third, some individual fibril proteins show variation in their primary structure such as the amyloidoses of λ- and K-light chain origin (Solomon and Weiss 1995). Finally, immunodetection itself is influenced by tissue fixation and embedding procedures (Shi et al. 1997). Some of these problems can be solved either by the use of antibodies specifically directed against the amyloid fibril proteins or by the use of polyclonal antibodies, or simply by trial and error of a broad range of different antibodies. A number of commercially available antibodies have been shown to be useful for immunotyping of amyloid (Linke 1984; Ii et al. 1992; Iversen et al. 1995; Arbustini et al. 1997; Röcken 1998). However, the effect of tissue fixation and embedding is often a limiting variable, particularly in specimens processed for electron microscopy (EM). The aim of this study was to investigate the effect of different etching and AR procedures on postembedding immunolabeling of amyloid in electron microscopic specimens that were fixed and embedded by conventional procedures.
Materials and Methods
Tissues
Specimens were obtained from three autopsy cases. The first was a 64-year-old white man who had suffered from rheumatoid arthritis; specimens from heart, lung, liver, kidney, spleen, and thyroid were fixed overnight in 5% neutralized formalin. The second case, an 82-year-old white man, had suffered from cholecystitis due to cholecystolithiasis; specimens from heart, lung, liver, kidney, spleen, pancreas, duodenum, and prostate were fixed and stored in 5% neutralized formalin, either overnight (paraffin embedding) or for 7 weeks (Epon embedding) before processing. The third case was a 72-year-old white man who had suffered from a plasmacytoma; specimens from heart, lung, liver, and kidney were fixed and stored in 5% formalin for 5 weeks before processing. The autopsies had been performed 1, 4, and 2 days after death, respectively. For light microscopy, specimens were embedded in paraffin. Deparaffinized sections were stained with hematoxylin and eosin. The presence of amyloid was demonstrated by the appearance of green birefringence from alkaline alcoholic Congo Red staining under polarized light (Puchtler et al. 1962). Amyloid was classified immunohistochemically with antibodies directed against AA amyloid, transthyretin, β2-microglobulin, λ-light chain, and κ-light chain, as described elsewhere (Röcken et al. 1999).
Electron Microscopy
For electron microscopy, the autopsy specimens were fixed in different ways. Specimens from the second and third case were postfixed in a mixture of 2% formalin/2.5% glutaraldehyde (pH 7.2, overnight, 4C) and subsequently in 3.125% glutaraldehyde only (7 hr, 4C). Specimens from the first case were fixed in three different ways: Procedure 1, initial fixation in 5% formalin [pH 7.0, overnight, room temperature (RT) followed by 3.125% glutaraldehyde (pH 7.2, 16 hr, 4C); Procedure 2, initial fixation in a mixture of 2% formalin/2.5% glutaraldehyde (pH 7.2, overnight, 4C) and subsequently in 3.125% glutaraldehyde (pH 7.2, overnight, 4C); or Procedure 3, in a mixture of 4% paraformaldehyde/0.2% glutaraldehyde (pH 7.2, 40 hr, 4C) only. After the different fixation procedures, all samples were washed in PBS, postfixed in 2% OsO4-Veronal buffer [2% OsO4 in 2.94 g 5,5-sodium barbital, 1.94 g sodium acetate, 500 ml de-ionized water (dH2O); pH 7.2, 4 hr, 4C], and rinsed thoroughly in Veronal buffer (three times for 5 min, 4C) and dH2O (three times for 5 min, RT). Aqueous uranyl acetate 3% (60 min, RT) was then added and the specimens were dehydrated stepwise in 20, 30, and 50% ethanol (10 min each, RT). After overnight incubation in 70% ethanol (4C), the specimens were treated with 1% uranyl acetate in 70% ethanol (60 min, RT) and further dehydrated in 96% (twice for 30 min, RT) and absolute ethanol (three times for 30 min, RT). Before embedding the samples were incubated in propylene oxide (twice for 15 min, RT) and propylene oxide 1:1 Epon (51 ml Epon 812, 37 ml methylnadic anhydride, 12 ml dodecenyl succinic anhydride; 3 hr, RT). Finally, the specimens were embedded in Epon supplemented with 2% DMP-30 and polymerization took place over 24 hr at 60C. Semithin sections (1 μm) were stained with Toluidine blue. Ultrathin sections (120 nm) were mounted on copper grids and counterstained with 3% aqueous uranyl acetate (30 min, RT), then contrasted with 1% aqueous lead citrate (15 min, RT).
Etching and Antigen Retrieval Procedures
Postembedding immunostaining of ultrathin sections was performed after the following etching and AR procedures were applied: (a) no treatment; (b) etching in H2O2; (c) treatment with saturated aqueous sodium metaperiodate (mPJ); (d) heating in dH2O; (e) heating in sodium citrate buffer (SCB); (f) heating in EDTA; (g) etching in H2O2 followed by heating in dH2O; (h) etching in H2O2 followed by heating in SCB; (i) etching in H2O2 followed by heating in EDTA; (j) treatment with mPJ followed by heating in dH2O; (k) treatment with mPJ followed by heating in SCB; (l) treatment with mPJ followed by heating in EDTA.
For postembedding immunoelectron microscopy, ultrathin sections (120 nm) were mounted on Formvar-coated nickel grids (200 mesh; Plano, Wetzlar, Germany). Etching and treatment were carried out using 10-μl drops of either 3% H2O2 (10 min, RT) or mPJ solution (60 min, RT) on Parafilm sheets in microwell culture plates, using a moist chamber for the latter. Antigen retrieval (AR) was performed using either dH2O (30 min, 91C), 10 mM SCB (10 ml 0.1 M citric acid monohydrate, 90 ml 0.1 M trisodium citrate dihydrate, and 900 ml dH2O; pH 6.0, 30 min, 91C), or 1 mM EDTA (pH 8.0, 30 min, 91C). One mM EDTA was prepared using a stock solution of 500 mM EDTA adjusted to pH 8.0 with NaOH. A working solution was obtained by diluting the stock solution 1:500 with dH2O. For AR, the following heating set-up was applied. Using a scalpel or a razor blade, slim slots were cut along the longitudinal axis of one end of a stiff plastic tube measuring 10–12 cm in length and 6 mm external diameter. The grids were tucked mechanically into the slots. The elastic tension of the tubing was strong enough to maintain the position of the grids during heating, and no glue was required. A cork was attached to the opposite end of the tubing and the whole assembly was put into a test tube filled with the preheated AR solution and placed in a hot water bath. The temperature was controlled by two thermometers. One was placed inside a control test tube without grids and the other in the surrounding water bath. The test tubes were supported by a 4-well plastic rack (see Figures 1 and 2).
Postembedding Immunolabeling
Postembedding immunolabeling was carried out after etching and AR. For immunolabeling with anti-AA amyloid and anti-κ-light chain, the grids were rinsed in 50 mM Tris-buffered saline (Tris-HCl; pH 7.4, three time for 5 min, RT) and incubated in Tris-HCl containing 5% (w/v) fetal calf serum (Tris-FCS; 30 min, RT). Tris-FCS was removed without washing the specimens and the primary antibodies were applied, i.e., anti-AA amyloid (dilution 1:100; 1 hr at RT; monoclonal), or anti-κ-light chain (1:167; 1 hr at RT; polyclonal; both from DAKO, Carpinteria, CA). Immunostaining was visualized by the appropriate 15-nm colloidal gold-labeled secondary antibodies (EM 15-nm rabbit anti-mouse IgG and EM 15-nm goat anti-rabbit IgG; both from Biotrend, Cologne, Germany; 1:20, 60 min, RT). Between the incubations the specimens were washed with Tris-FCS (five times for 5 min). The antibodies were then crosslinked with 1% glutaraldehyde in Tris-HCl (pH 7.4, 15 min, RT). The immunolabeling procedure with anti-transthyretin (1:50; overnight at 4C; polyclonal; DAKO) was identical except that PBS, pH 7.4, was used instead of Tris-HCl. Finally, the sections were counterstained with 3% aqueous uranyl acetate (30 min, RT) and contrasted with 1% aqueous lead citrate (15 min, RT). Fixation, staining, and contrasting were each followed by an extensive washing in dH2O. The specificity of immunostaining was controlled by omitting the primary antibodies. The sections were air-dried and inspected in a Zeiss EM900 electron microscope.
Quantitation of Probe Density
The mean numerical probe density was quantified as follows. In each experiment, 10 pictures were taken (at a magnification of X20,000) at random from amyloidotic areas. This was performed in triplicate and the number of gold particles was counted in the whole frame of 30 pictures. The mean number of gold particles per frame was calculated. At a magnification of X20,000, the individual frame covered 13.68 (μm2. As a control, the number of gold particles in nonamyloidotic areas was quantified in the same manner. Statistics were performed using the V2.04a GraphPad Instat program.
Results
Light Microscopy
In the first and third autopsy cases, a variable amount of vascular and interstitial amyloid deposit was found in every organ investigated (Figures 3 and 7).
In the first case, immunohistochemistry showed intense staining of the amyloid deposits with the antibody directed against AA amyloid only (Figure 4), and the diagnosis was a generalized AA amyloidosis secondary to rheumatoid arthritis. The patient had died from renal insufficiency caused by amyloid deposition.
Equipment and set-up used for heat-mediated AR (for details see Materials and Methods).
In the third case, immunohistochemical classification of the amyloid deposits showed immunostaining with the antibody directed against K-light chain only (Figure 8), and the diagnosis was a generalized κ-lightchain amyloidosis secondary to a plasmacytoma. The patient had died from cardiac failure caused by massive cardiac and pulmonary deposition of amyloid.
In the second case, a variable amount of vascular amyloid deposit was found only in the left and right cardiac ventricles, lung, liver, duodenum, and prostate (Figure 5). No amyloid was evident in the kidneys, spleen, and pancreas. The deposits were immunoreactive for transthyretin only (Figure 6) and were interpreted as systemic senile amyloidosis. The patient had died from cardiac failure after cholecystectomy, which was not attributable to amyloid deposition.
Electron Microscopy
Electron microscopic examination of the specimens from all cases showed irregular arrangements of linear, nonbranching fibrils with an indefinite length and an approximate diameter of 10–12 nm. These features are typical of amyloid.
Immunoelectron Microscopy
Specimen Selection. For immunoelectron microscopic investigation of the first and third cases, specimens were sought with abundant vascular and interstitial deposits of amyloid, interspersed with large areas without amyloid, the latter serving as an internal negative control. In the first case, specimens from the thyroid only proved useful because massive deposits of amyloid encircled follicles homogeneously filled with colloid. For the same reason, specimens from the lung were selected in the third case, in which massive deposits of amyloid encircled optically empty alveolar spaces. In the second case, specimens from the heart only enclosed a sufficient amount of vascular amyloid deposits appropriate for immunoelectron microscopic investigation.
AcknowledgmentsPostembedding Immunolabeling
Controls. Background staining of postembedding immunolabeling was investigated by omitting the primary antibody. Formvar-coated nickel grids yielded a weak background staining with gold particles, whereas no particles were found on blank ultrathin sections of Epon. Irrespective of whether etching or AR was performed, tissue sections from the thyroid, heart, and lung yielded no or a negligible amount, of gold particles when the primary antibody was omitted (Figure 11). Application of primary antibodies against AA amyloid on sections from the heart (ATTR amyloidosis) and lung (Aκ amyloidosis) and against κ-light-chain on sections from the thyroid (AA amyloidosis) and heart (ATTR amyloidosis) showed only weak nonspecific labeling of the specimens and no signal enhancement was found on amyloid deposits. Immunostaining of the lung (Aκ amyloidosis) and thyroid (AA amyloidosis) with anti-transthyretin showed intense nonspecific labeling without signal enhancement on amyloid.
Specimen Preservation After Etching and Antigen Retrieval
Treatment with mPJ significantly reduced the contrast of the sections but the amyloid fibrils were still discernible. Etching with H2O2 occasionally affected the specimens, resulting in reduced contrast and loss of tissue structure. After AR using SCB or EDTA, a variable number of different sized “holes” were found, although the integrity of the ultrathin section and the contrast appeared to be maintained. To further analyze the holes, AR was tested on ultrathin sections mounted on nickel grids without Formvar coating. No holes were evident, and we concluded that they were artifacts attributible to the use of Formvar.
First Case (AA Amyloidosis)
Postembedding immunolabeling of ultrathin sections from the thyroid with the antibody directed against AA amyloid yielded immunostaining immediately related spatially to amyloid fibrils. The mean density of gold particles was 31.1/13.68 μm2 (Table 1). Except for heating in EDTA (82.3/13.68 μm2), no other etching, pretreatment, or AR procedure significantly improved immunolabeling with anti-AA amyloid (Table 1). Background staining was only moderately influenced by AR (8.0/13.68 μm2 after EDTA retrieval vs 1.8/13.68 μm2 without specimen pretreatment) and did not account for increased immunolabeling. Heating in dH2O and etching in H2O2 followed by heating in SCB significantly reduced immunostaining (Table 1).
Effect of different AR procedures on postembedding immunolabeling of amyloid fibrils deposited in a thyroid specimen of an autopsy case with generalized AA amyloidosis (n = 30)
To investigate the influence of fixation on AR, three different aldehyde fixation procedures were tested [i.e., initial fixation in 5% formalin followed by 3.125% glutaraldehyde (Procedure 1), initial fixation in a mixture of 2% formalin/2.5% glutaraldehyde and subsequently in 3.125% glutaraldehyde (Procedure 2), and fixation in a mixture of 4% paraformaldehyde/0.2% glutaraldehyde only (Procedure 3)]. These had little effect on the intensity of immunostaining enhanced by EDTA pretreatment.
Effect of different AR procedures on postembedding immunolabeling of amyloid fibrils in a heart specimen of an autopsy case with senile cardiovascular ATTR amyloidosis (n = 30)
Lung specimens from Case 3 (K-light chain amyloidosis) showing a negative control with omission of the primary antibody (Figure 11), immunostaining of amyloid without AR (Figure 13), and after heating the specimens in EDTA (Figures 12 and 14). Bars = 0.4 μm.
Second case (ATTR Amyloidosis)
Postembedding immunolabeling of ultrathin sections from the heart with an antibody against transthyretin showed immunostaining immediately related spatially to amyloid fibrils (Figures 9 and 10). The mean density of gold particles was 103.3/13.68 μm2 (Table 2). Except for a combined pretreatment with H2O2 and EDTA (306.9/13.68 μm2) and mPJ and EDTA (319.1/13.68 μm2), no other etching, pretreatment, or AR procedure significantly improved immunolabeling with anti-transthyretin (Figures 9 and 10; Table 2). However, background staining in EDTA-treated specimens increased significantly after etching with H2O2 (57.4/13.68 μm2) and mPJ (93.7/13.68 μm2) compared to that found with no etching (24.5/13.68 μm2). Therefore, an increased immunolabeling of amyloid was interpreted as being due to an overall increase in the background staining. H2O2 (10.6/13.68 μm2) and mPJ (19.0/13.68 μm2) alone without subsequent EDTA treatment did not increase the background staining compared to that found with no treatment (20.1/13.68 μm2).
Third Case (Aκ Amyloidosis)
Postembedding immunolabeling of ultrathin sections from the lung with an antibody against κ-light chain yielded immunostaining immediately related spatially to the amyloid fibrils (Figures 12-14). The mean density of gold particles was 67.0/13.68 μm2. Specimen pretreatment with either H2O2, mPJ, or hot dH2O had no significant effect on immunolabeling, as indicated by the mean number of gold particles per frame (see Table 3). A significant increase was achieved after the specimens had been treated either with SCB or EDTA. EDTA showed the most significant effect on immunolabeling, yielding a mean number of 241.1/13.68 μm2 (Figure 13 and 14; Table 3). Furthermore, the effect of EDTA was reduced after etching with H2O2. Neither pretreatment with mPJ nor etching with H2O2 had any significant effect on immunolabeling. The mean numbers of gold particles on nonamyloidotic tissue areas were 3.8/13.68 μm2 (no treatment), 2.5/13.68 μ2 (mPJ + SCB), 34.0/13.68 μ2 (no etching + EDTA), and 20.3/13.68 μ2 (mPJ + EDTA). Therefore, the moderately increased background staining had a negligible effect on the overall enhanced immunolabeling of the amyloid fibrils.
Discussion
Heat-mediated antigen retrieval (AR) is a nonenzymatic pretreatment for immunohistochemical staining that is widely used in research and surgical pathology (for reviews see Boon and Kok 1994; Shi et al. 1997; Leong and Sormunen 1998). AR may facilitate or improve the detection of antigens, particularly those that have been fixed by formalin. It increases the immunohistochemical sensitivity for a wide range of antibodies and retrieves antigens that may otherwise be undetected. AR reduces the incidence of false-negative results and increases the diagnostic accuracy and reproducibility and, therefore, ultimately patient care (Boon and Kok 1994; Shi et al. 1997; Leong and Sormunen 1998).
Several different methods can be used to achieve AR, all of which probably reflect the different but ill-defined and unpredictable biochemical mechanisms of AR (Shi et al. 1993,1997; Boon and Kok 1994; Marani and Horobin 1994; Morgan et al. 1994; Stirling and Graff 1995; Xiao et al. 1996). Microwaves, pressure cookers, autoclaves, water baths, and steam are different modes of heating for AR. These methods use aluminum chloride, saturated lead thiocyanate, citrate, Tris buffer, EDTA, EGTA, glycine-HCl buffer, periodic acid, or dH2O as AR medium. AR can also be performed by nonheating methods, such as immersing the tissue in water, 10% sucrose in PBS, urea, acid, borohydride, or NaOH-methanol for various amounts of time. All these procedures have shown that AR is sensitive to heat, pH, molarity, and metal ions, and that the modification of a protein structure caused by formalin is reversible under certain conditions (Boon and Kok 1994; Shi et al. 1997).
Aldehyde fixatives, such as formalin and glutaraldehyde, are crosslinkers that are preferred for tissue fixation because they are superior for preservation of tissue structure for both light and electron microscopy (Prento and Lyon 1997). They have been commonly used worldwide and guarantee the standardization and comparability of morphology, particularly in surgical pathology. Crosslinking joins two molecules by covalent bonds, and AR may loosen or break cross-linkages caused by aldehyde fixation. In addition, AR may also unmask antigens by extracting diffusible proteins. It may precipitate and stabilize proteins and rehydrate sections to improve the diffusion of the antibody into the tissue sections (Boon and Kok 1994; Shi et al. 1997). These probable mechanisms demonstrate that the effect of AR depends not only on the mode of fixation, but also on the antigen under investigation. There is no evidence thus far for a “one and only” AR method, and it has been recommended that a test battery of different AR procedures should be applied to develop an optimal AR method for a given antigen or antibody (Boon and Kok 1994; Shi et al. 1997).
Effect of different AR procedures on postembedding immunolabeling of amyloid fibrils in a lung specimen of an autopsy case with generalized K-light chain amyloidosis (n = 30)
In this study we applied such a test battery of different etching and AR procedures to improve the postembedding immunohistochemical classification of amyloid in electron microscopic specimens that were fixed and embedded by conventional procedures, i.e., formalin and glutaraldehyde fixation and embedding in Epon. Our results demonstrate that, for these specimens, AR shows antigen/antibody-specific retrieval patterns. Immunolabeling with a monoclonal antibody against AA amyloid was sensitive only to EDTA retrieval, and no other specimen pretreatment had a significant effect. Immunostaining with anti-transthyretin was sensitive to EDTA and to the combinations of H2O2/EDTA and mPJ/EDTA. However, enhanced signal density after the latter two AR procedures was due to a dramatically increased background staining, which was not observed for the other two antigens/antibodies under investigation. The possibility that increased background staining may have been due to increased nonspecific binding of the secondary antibody (Shi et al. 1997) was ruled out, because the negative controls (omission of the primary antibody) showed no increased background staining. Neither etching with H2O2 nor incubation with mPJ influenced background labeling; it was only the combination of etching and AR that had this effect.
Unlike anti-AA and anti-transthyretin, immunolabeling with anti-K-light chain was improved by a combined pretreatment with mPJ/SCB and mPJ/EDTA, and was not merely due to increased background staining. However, it is interesting that none of the antigens/antibodies investigated showed improved staining after heating in dH2O. Therefore, it is likely that the AR was mediated not by heat and rehydration alone but by the presence of a chelating or metal binding agent, such as EDTA or SCB (Morgan et al. 1994).
In immunoelectron microscopy, etching is applied either to increase the surface area of Epon, and thereby the number of antigens available for antibody binding, or to reverse the negative effect of osmium tetroxide fixation on antigen preservation (Polak and Priestley 1992). Previous studies have shown that a combined treatment of etching and AR may be superior to either etching or AR alone (Stirling and Graff 1995). In our study, the use of either H2O2 or mPJ alone had no significant effect on postembedding immunolabeling for any of the three antigens/antibodies investigated. In combination with AR, they showed different effects by improving immunolabeling (κ-light chain amyloidosis: mPJ + SCB), enhancing background staining (ATTR amyloidosis), or counteracting the effect of AR. AR using only EDTA improved immunostaining with anti-AA, but this result was abolished when EDTA was combined with etching in H2O2 or mPJ. Similar observations were made with κ-light chain amyloidosis (H2O2 + EDTA). Therefore, the effect of etching in combination with AR is unpredictable and must be assessed by trial and error using a test battery.
Different methods have been described for the application of heat for antigen unmasking in immunoelectron microscopy. These include microwaving, autoclaving, pressure cooking, and boiling in an open beaker (Stirling and Graff 1995; Xiao et al. 1996). We tested a new method that has not been previously described, but which proved useful for AR and had advantages over boiling in an open beaker. During preliminary experiments (not shown), we observed that heating grids by floating them section side down in an open beaker was difficult to reproduce. The handling of the floating and moving grids was difficult, and only a small number of grids could be handled at one time. In our system, the mode of heating became more consistent because the grids had a fixed position during heating. The plastic tubings could be used many times. Little if any AR medium evaporated during 30 min of heating because we used a closed system with the glass test tube sealed by a cork. Finally, our set-up is cheap, simple to control, and requires only standard laboratory equipment.
In summary, our study shows that AR may significantly improve immunostaining of specimens that have undergone conventional fixation and embedding procedures for electron microscopy. In our series of tests, the effect of AR on the detection of amyloid fibril proteins was probably mediated in part through chelation or binding of metal ions by the AR medium. The application of a test battery proved valuable in assessing the appropriate etching and AR procedure, because the effect of etching and AR is unpredictable for a given antigen/antibody.
Footnotes
Acknowledgements
We thank Ms Becher, Ms Drueg, Ms Koçalkova, and Ms Röder for excellent and skillful assistance. The specimens from the third autopsy case were kindly provided by Professor Dr. Med. Otto (Dortmund, Germany).