Post-embedding Double-labeling of Antigen-retrieved Ultrathin Sections Using a Silver Enhancement-controlled Sequential Immunogold (SECSI) Technique

Abstract

In electron microscopy, the post-embedding immunogold technique provides a high degree of resolution and the possibility of quantitation owing to the intrinsic characteristics of the colloidal gold marker. Application of this technique to the subcellular localization of multiple antigens by differential labeling using gold markers of different sizes, or to double labeling using the same primary antibody isotype with serial silver enhancement, has been reported. We have incorporated this double labeling technique into a modified procedure that produces excellent labeling and ultrastructural preservation, even after exposure of ultrathin sections large enough to cover a 300-μm-diameter single-hole grid to hot antigen retrieval solutions and prolonged labeling protocols.

We have been studying the relative expression of structural proteins that may modulate changes in the shape of epithelial podocyte foot processes and their anchorage to the underlying glomerular basement membrane (GBM) in normal kidney and in nephrosis.

Ultrastructural localization of multiple antigens provides essential information about the spatial and functional relationships between different macromolecules at the subcellular level. Gold conjugates of different particle sizes offer a simple approach to differential antigen localization in double labeling techniques. However, sequential silver enhancement provides an alternative means of particle size segregation, using gold conjugates of a similar size, that has been successfully applied to both pre- and post-embedding labeling of tissue (Bienz et al. 1986; Bienz and Egger 1998; Yi et al. 2001). The use of hydrophilic acrylic resins (e.g., Lowicryl K4M, LR Gold, LR White, methacrylate) enables the application of immunogold techniques to post-embedding labeling of tissue using more conventional-sized gold conjugates (5-15 nm) without the need for complicated specimen preparation to preserve antigenicity and ultrastructure. Although ultrastructural preservation is adequate in tissues embedded in these acrylic resins, some antigens may be suboptimally retained. In contrast to epoxy resin-embedded sections (Stirling and Graff 1995; Brorson 2001), the mechanical fragility of acrylic ultrathin sections and the thermolability of these resins has previously restricted attempts to retrieve antigenicity with conventional heat-mediated retrieval techniques (Xiao et al. 1996).

Here we describe the modifications adapted to antigen retrieval and double labeling of epitopes using a silver enhancement-controlled sequential immunogold-silver technique (SECSI) applied to ultrathin renal sections embedded in LR Gold resin, stabilized, and mounted on large-diameter (300-μm) single-hole grids. In this study we were able to use the same isotype-specific primary antibodies to each epitope with excellent specificity, sensitivity, durability, and preservation of ultrastructure.

Archived renal biopsy and nephrectomy tissue blocks (1-mm³) fixed in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, and embedded in LR Gold resin (London Resin; Basingstoke, Hants, UK) by polymerization under UV light (360 nm) for 24 hr at 4C (Goode et al. 1991) were used under local ethical guidelines and approval (project Nos. 02/068 and 02/072).

The study included tissue showing “normal” histology dissected from the unaffected poles of kidneys that had been surgically removed for treatment of renal carcinoma and renal biopsy tissue from patients with membranous glomerulopathy.

Mouse monoclonal IgG1 anti-human antibodies raised against α-actinin (Insight clone AT6/172), synaptopodin (Progen clone G1D4), and β-1 integrin (CD29) (Neomarkers clone 29C03; Insight Biotechnology, Wembley, Middlesex, UK) and electron microscopy grade goat anti-mouse IgG conjugate (10 nm) (British Biocell International; Cardiff, UK) were used for immunolabeling procedures.

Ultrathin tissue sections, nominally cut at a thickness of 90 nm that showed silver/gold interference colors, were placed on 300-μm single-hole gold grids (Gilder; Grantham, Lincolnshire, UK). Mounting sections on grids to which LR Gold blank sections of similar thickness had been previously applied to the reverse side provided mechanical support for the subsequent demanding handling conditions. All antigens included in this study required unmasking before their detection using the immunogold labeling technique.

Antigen retrieval was performed with a modification of the methods described by Shi et al. (1995) and Stirling and Graff (1995). Briefly, sections were antigen-retrieved by immersion of grids bearing sections in 0.1 M sodium citrate buffer, pH 6.0, at 95C for 10 min. Sections were left to cool to room temperature for 20 min in the citrate buffer, jet-washed (deionized water), immersed in drops of 0.5 M NH4Cl in 0.1 M PBS, pH 7.3, for 20 min and then washed for 5 min in PBS, pH 7.3, containing 1% bovine serum albumin (BSA) and 0.1% Tween-20 (washing buffer). After blocking nonspecific binding in 20% normal goat serum (NGS) diluted in washing buffer for 10 min (blocking buffer), sections were incubated overnight at 4C in each of the primary antibodies, appropriately diluted in PBS, pH 7.3, containing 1% BSA, 0.1% Tween-20, and 5% NGS (antibody buffer). (α-Actinin, integrin β-1, and synaptopodin supernatant antibodies were respectively diluted 1:50; 1:10, and 1:2.) After rinsing in three changes of washing buffer for 5 min each, sections were incubated in blocking buffer for 20 min at 20C, followed by a 2-hr incubation at 20C in goat anti-mouse IgG conjugated to 10-nm gold particles diluted 1:100 in antibody buffer. Sections were jet-washed, silver-enhanced for 12 min (Holgate et al. 1986), and washed again in deionized water for 2 min before the silver stain was fixed for 2 min in 3% sodium thiosulfate. Before application of the second monoclonal antibody, sections were jet-washed and the procedure was repeated as for the first primary antibody. The sections were silver-enhanced for 6 min, jet-washed again, and counterstained in saturated aqueous uranyl acetate for 3 min before being finally rinsed briefly in distilled water and allowed to dry. The specificity of the anti-mouse-IgG gold conjugate for each of the same isotype-specific primary monoclonal antibodies at each stage of the sequential labeling protocol was confirmed by comparison of the double-labeled pattern with that obtained after substitution of either the first or the second primary antibody in the sequence with normal mouse serum, diluted 1:20 in antibody buffer. This dilution provided an excess of mouse serum immunoglobulins compared with the IgG concentration in the primary antibodies. Care was taken to further confirm the validity of the technique by reversal of the sequence of application of each primary antibody.

Sections were examined and photographed at 40 kV using a JEOL JEM 100CX electron microscope. The use of single-hole grids enabled the unimpeded examination of whole glomerular cross-sections from each case.

Scanned images were downloaded to CD-ROM and analyzed using LUCIA L/G imaging software (Nikon UK; Kingston upon Thames, Surrey, UK).

In preliminary studies, it was evident that LR Gold ultrathin sections unsupported on 300-nm-diameter single-hole grids could not survive the rigors of this labeling protocol. Exposure of sections to hot antigen-retrieval solutions accounted for the loss of most sections. Those that survived this initial step did not survive the rigors of extensive grid handling or exposure to the electron beam. However, the use of mechanical support, provided by a blank LR Gold section of similar thickness to the ultrathin section itself, regularly increased section survival to >80%. Our ability to examine sections on these 300-μm-diameter single-hole grids, unimpeded by grid bars, provided obvious advantages of examination and analysis.

Use of the same isotype-specific antibodies in combination with silver enhancement to double label different antigens in ultrathin sections has been previously described (Bienz and Egger 1998; Bienz et al. 1986). Controls that substituted normal mouse serum for the second primary antibody showed no aggregation of smaller silver-enhanced gold particles around the larger, more fully enhanced particles arising from the first phase of the technique. This indicated that the gold conjugate used to visualize the second primary antibody in the sequence was not recognizing any residual reactive sites on the first primary antibody, these sites evidently having been fully encapsulated by the first silver-enhancement step. The relative labeling of each antigen was not influenced by a reversal of the sequence of application of the first and second primary antibodies.

Typical results obtained with our SECSI protocol are shown in Figures 1 and 2A. It can be seen that, in this technique, both small and large silver-enhanced gold particles are of uniform size and shape and can be easily differentiated from each other by both naked-eye examination of digitized images and by using the LUCIA image analysis system (Figure 2B).

Although the use of acrylic resins usually leads to higher efficiency of immunogold labeling than conventional epoxy resin (Brorson and Skjorten 1996), many antigens still require heat retrieval before they can be demonstrated by post-embedding staining of acrylic sections. Exposure of acrylic sections to hot (>90C) antigen retrieval buffers and the subsequent handling procedures would result in a significant loss of sections if these are not adequately supported. We and others have reported successful immunolabeling of antigen-retrieved ultrathin sections mounted on mesh grids (Xiao et al. 1996; Comer et al. 1999). Mesh grids, however, are cumbersome in comparison to single-hole grids owing to the interruption of the field of view by the support bars. Ultrathin acrylic sections mounted on single-hole grids are usually stabilized by application of a formvar support film (Abad 1988). In preliminary studies, we showed that the application of formvar coating was unsuitable for grids exposed to hot solutions because of wrinkling and ultimate failure of the film. The use of blank LR Gold resin section supports in this study provides a simple and robust alternative.

The majority of commercially available mouse monoclonal antibodies are of the IgG1 isotype. Many of these are directed against rare antigens, and therefore the choice of antibody heterogeneity in immunocytochemistry is often limited. The technique of sequential double immunolabeling using antibodies from the same species, with inactivation of the anti-species antibody on the gold grain by silver enhancement, was first reported by Bienz et al. (1986). In contrast to their study, which used a very precise and complex silver enhancement procedure, our silver enhancement technique (Holgate et al. 1986) is simple, cheap, and reproducible, resulting in a clearly differential growth of enhanced gold particles.

This uniformity of shape and the size difference between single-enhanced and doubly-enhanced gold particles enables each antigen to be easily differentiated using the LUCIA imaging system. We preferred to label the most abundant antigen second in the sequence, producing a more esthetically pleasing result and avoiding potential difficulties of detection of fewer, smaller particles among a plethora of large silver-enhanced gold particles.

Figure 1

SECSI-labeled normal glomerular capillary wall segment, showing a distinct distribution of large particles representing β1-integrin chains in the GBM and small particles representing α-actinin, restricted to the foot processes of epithelial podocytes. GBM, glomerular basement membrane; Ep, epithelial podocyte. Bar = 1 μm.

Figure 2

(A) SECSI-labeled glomerular capillary wall segment containing electron-dense deposits (*) from a patient with membranous nephropathy. An electron-dense basal layer of the effaced podocyte above the deposits contains co-localized large (synaptopodin) and small (α-actinin) particles. (B) Binary overlays of A obtained with the LUCIA imaging system demonstrate the ability to differentiate the particle sizes. α-Actinin and synaptopodin are represented by yellow and red, respectively. GBM, glomerular basement membrane; Ep, epithelial podocyte. Bars = 1 μm.

Our protocol enables good-quality, robust immunogold-labeled ultrathin sections to be reliably produced using only modest resources, providing the opportunity to study the relative distribution of multiple antigens by both quantitative and qualitative analyses.