Histology of plastic embedded amphibian embryos and larvae

Abstract

Amphibians including the South African clawed frog Xenopus laevis, its close relative Xenopus tropicalis, and the Mexican axolotl (Ambystoma mexicanum) are important vertebrate models for cell biology, development, and regeneration. For the analysis of embryos and larva with altered gene expression in gain-of-function or loss-of-function studies histology is increasingly important. Here, we discuss plastic or resin embedding of embryos as valuable alternatives to conventional paraffin embedding. For example, microwave-assisted tissue processing, combined with embedding in the glycol methacrylate Technovit 7100, is a fast, simple, and reliable method to obtain state-of-the-art histology with high resolution of cellular details in less than a day. Microwave-processed samples embedded in Epon 812 are also useful for transmission electron microscopy. Finally, Technovit-embedded samples are well suited for serial section analysis of embryos labeled either by whole-mount immunofluorescence, or with tracers such as GFP or fluorescent dextrans. Therefore, plastic embedding offers a versatile alternative to paraffin embedding for routine histology and immunocytochemistry of amphibian embryos. genesis 50:235–250, 2012. © 2011 Wiley Periodicals, Inc.

INTRODUCTION

For more than a century amphibian embryos have been important models for cell and developmental biology. Since the discovery of the first potential molecular regulators of early development (e.g. Vg-1: Weeks and Melton,1987; Activin: Smith et al.,1990; Goosecoid: Cho et al.,1991), the combination of classical embryological assays and molecular biology has been established the early amphibian embryo, especially the South African clawed frog Xenopus laevis and its close relative Xenopus tropicalis as powerful systems to study germ layer formation, axis determination, and organogenesis (Sive et al.,2000). Although X. laevis is not the ideal vertebrate for genetic studies because of its pseudotetraploid genome and long generation times, the recently introduced Xenopus tropicalis which has a diploid genome and shorter generation times has been pushed forward to become a new vertebrate model for genetic studies (Amaya et al.,1998; Hirsch et al.,2002; Carruthers and Stemple,2006; Goda et al.,2006).

Recently, the study of amphibian development raised more specific questions about cell movements, tissue formation, and cell differentiation in a wide variety of developmental processes from gastrulation to organogenesis. Many regulators of these developmental processes have been isolated over the past decades and their specific roles are studied using RT-PCR, biochemistry and proteomics, whole mount in situ hybridization, immunocytochemistry, and, of course, descriptive morphological phenotype characterization of embryos from gain-of-function or loss-of-function experiments. Since many effects on development are subtle and may escape unnoticed in whole-mount preparations, closer inspection by histology becomes a more and more important part of phenotype analysis.

In this article we provide and discuss techniques for histology and immunocytochemistry of plastic or resin embedded embryo samples, some of which are presented here for the first time. We focus on plastic histology at the light microscope level, which provides a simple and versatile alternative to the classic paraffin embedding. It can be used for routine histology as well as immunocytochemistry. The preparation of embryonic tissues for electron microscopy has been recently reviewed elsewhere (Kurth et al.,2010).

Fixation

Fixation is the crucial step in any kind of cell and tissue preparation. The ideal goal is to fix cells in an as-close-to-native-as-possible state within an indefinitely short period of time. Theoretically this can be achieved by ultrarapid freezing of very small samples (up to 200 μm) using high pressure freezing (Dubochet,2007; Moor,1987; Studer et al.,2008; Vanhecke et al.,2008). With chemical fixation, still kept as the main option for embryological studies, this goal is basically never reached (for an excellent review on chemical fixation see chapter 2 in Griffiths,1993). The most important factors influencing the quality of chemical fixation are size and composition of the sample. For tissue samples an ideal block size would be 0.5 × 0.5 × 0.5 mm. Early Xenopus and Ambystoma embryos are larger and filled with semicrystalline yolk platelets. These yolk platelets pose a serious problem for histological preparations because they become brittle after strong fixation and are difficult to section, especially after embedding in paraffin. In later stages, yolk is less abundant, but different cell types of different stiffness and rigidity and of different chemical compositions are lying next to each other in a rather small sample volume. Each of these different cell types would actually need different fixation regimes to achieve optimal preservation. This is, of course, not possible, resulting in the quite common observation that certain well fixed cells lie directly adjacent to other badly fixed cell types (Kelly et al.,1991).

If it is possible to dissect the target tissue before fixation without creating serious damage to the area of interest, it is highly recommended. Alternatively, samples can be prefixed in a weak fixative (e.g. 4% paraformaldehyde in phosphate buffer) for 1 h and dissected within that fixative before transfer to the final, stronger fixative. If dissection is not possible, a compromise fixative is mandatory that is good enough to penetrate the embryo in a reasonable amount of time and is able to fix a wide variety of cell types with reasonable quality.

There are many different fixatives used in histology. Most classic fixatives for routine tissue preparation such as Bouins (picric acid/glacial acetic acid) or Carnoys (ethanol/chloroform/glacial acetic acid) contain strong acids or solvents. These mixtures rapidly penetrate cells and tissues, but produce limited structural preservation (Kelly et al.,1991). Historically, this was improved by adding cross-linking reagents such as formaldehyde. A mixture of saturated mercuric chloride, trichloroacetic acid, and formaldehyde (Romeis fixative) was used for the beautiful whole embryo histology in the Hausen atlas (Hausen and Riebesell,1991). At higher magnification, however, even Romeis-fixed specimens display substantial cytoplasmic extraction, and intracellular details are difficult to see. For bulky samples, such as embryos, the challenge is to combine a fast diffusing fixative with a strong but slow diffusing fixative in a way that the sample gets sufficiently penetrated and properly fixed at the same time. A combination of paraformaldehyde (PFA) and the strongly cross-linking glutaraldehyde (GA), introduced by Karnovsky (1965), is the fixative of choice for high-resolution light microscopy and electron microscopy of larger samples such as embryos. For amphibian embryos, the original recipe has been modified by reducing the strength of the fixative (e.g. Hausen and Riebesell,1991; Kurth et al.,2010; Müller and Hausen,1995) or by adding additional reagents to support tissue penetration (PFA/GA/DMSO: Kelly et al.,1991; PFA/GA/DMSO/acrolein: Kalt and Tandler,1971; PFA/GA/picric acid: Bernardini et al.,1999).

In addition to primary fixation, samples could be postfixed with OsO₄ (1–2% solution), a step routinely performed for EM preparation. Osmiumtetroxide integrates into the double bonds of lipids thereby fixing lipids and membranes and contrasting them. Embryonic cells are rich in lipid vacuoles most of which normally get completely extracted (Fig. 1c,d), even after strong primary fixation (Fig. 2a). Postfixation with OsO₄ stabilizes those lipids and the vacuoles acquire brown color (Figs. 1a,b and 2b). However, embryos fixed that way are difficult to section from paraffin blocks, a problem than can be solved by embedding them into low viscosity resins such as Spurrs resin (Spurr,1969) or the glycol methacrylate Technovit 7100 (see below).

Figure 1. Influence of primary fixation on tissue preservation. (a and b) Primary fixation with glutaraldehyde/paraformaldehyde followed by postfixation with OsO₄. Xenopus tailbud embryo, cross-section through the head (a and a′) and the tail (b and b′). Note the dense staining and the visibility of many subcellular details. No extraction can be seen. (c and d) Fixation in 80% methanol/20% DMSO (Dent's fixative), postfixation with OsO₄. Xenopus tailbud, cross-section through head (c and c′) and tail (d and d′). Staining is much weaker due to massive cytoplasmic extraction, intracellular details are largely obscured. Squares in a–d indicate the regions displayed in a′–d′, respectively. Resin embedded samples, sections stained with TB/Borax.

Figure 2. Effect of postfixation with OsO₄ on tissue preservation. Trunk muscle of axolotl larvae. (a) Fixation in a glutaraldehyde/paraformaldehyde mixture (GA/PFA, modified Karnovsky). (b) Primary fixation in modified Karnovsky followed by postfixation with OsO₄. Technovit sections stained with TB/Borax. Note the better tissue preservation in (b), in particular the preservation of lipid droplets (*).

For immunohistology, both GA and OsO₄ cannot be used because strong GA-crosslinking masks and OsO₄ destroys most antigens. In addition, immunofluorescent signals are obscured by GA-autofluorescence. Therefore, only weak formaldehyde based fixation protocols can be used (2–5% PFA) with no or only minor amounts of GA (0.05–0.2% GA) (Ding et al.,1993; Kurth,2003; Schwarz,1994; Schwarz and Humbel,2007; Slot and Geuze,2007). Before whole-mount immunolabeling, however, even PFA-fixed samples have to be permeabilized by alcohols or detergents such as saponin or triton to improve the diffusion of antibodies. Alternatively, embryos can be fixed in a mixture of methanol and DMSO (Dent et al.,1989), a fixative extractive enough to render specimens already permeable for antibody incubation after primary fixation (Klymkowsky and Hanken,1991). For in situ hybridization, embryos are also fixed in weak formaldehyde solutions (e.g. MEMFA) and later transferred to methanol (Harland,1991). In both cases, embryos can be embedded and sectioned for histological inspection after labeling. However, the histological information that can be drawn out of such samples is rather limited due to cytoplasmic extraction during the methanol or methanol/DMSO steps (see Fig. 1c,d) and the staining procedures that occur before embedding (see also Figs. 7 and 9b for the histology of embryos after in situ hybridization and whole-mount immunolabeling, respectively).

Paraffin Versus Plastic Embedding

Embedding and sectioning

After fixation, samples have to be prepared for sectioning into thin slices. For routine histology and EM, many embedding media are available, such as paraffin (Sive et al.,2000), polyester wax (Steedman,1957), plexiglas (Hausen,1988), polyacrylamid (Hausen and Dreyer,1981), methacrylates (Bennett et al.,1976; Gerrits and Smid,1983; Hausen and Riebesell,1991; Ruddell,1967; Sullivan-Brown et al.,2011) and several resins for transmission electron microscopy (Araldit, Epon 812, Spurr, LR White, LR Gold, Lowicryl; Bozzola and Russell,1999). For Xenopus embryos, two options are regularly used: paraffin or plastic embedding. Paraffin embedding is a classical method and well established in many laboratories. In the clinic, it is routinely used for pathology, and most steps can be performed automatically using embedding stations or staining automats. Paraffin embedding is regarded as relatively simple and cheap and therefore often the first choice for tissue analysis of embryo samples as well (Kelly et al.,1991; Sive et al.,2000). Plastic embedding of Xenopus has also been used for embryos but is mainly considered as more challenging or complex. In combination with the proper fixatives it reveals, however, excellent morphology (Hausen and Riebesell,1991; Kelly et al.,1991; Sive et al.,2000). In contrast to the general view, we would like to stress that embedding into a low viscosity resin, such as the glycol methacrylate Technovit 7100 (Heraeus-Kulzer) is a versatile, simple, and relatively cheap method that reveals excellent structural preservation and many cellular details. The infiltration solution is a transparent liquid of low viscosity, and is easy to handle at room temperature. Orientation of samples during the actual embedding procedure is also convenient, in contrast to paraffin embedding, which is often hindered by solidifying paraffin on top of the embedding mold (even when expensive heated embedding stations are used). Compared with paraffin blocks, cooling of resin blocks is not necessary for sectioning. Technovit 7100 can be sectioned from 1 μm (semithin) to 10 μm (our standard thickness is 2 μm) using commercial single-use metal blades (Heraeus-Kulzer) or glass knifes prepared on a standard knife maker (LKB or Leica). Both options are cheap, easy to use, and therefore highly recommended in a multiuser environment. Alternatively, high quality metal knifes can be used to achieve even better sections. However, all images presented here were from samples sectioned with the single use metal blades or with glass knifes! With a good rotational microtome sectioning itself is relatively easy. In our experience from numerous histology courses, undergraduate students learn to perform sectioning (2–4 μm) within 2 h.

Mounting of sections

Sections are stretched in a warm water bath, mounted to clean normal or to Superfrost® microscope slides, and dried for at least 2 h at 37°C to 50°C (for example on a hot plate). For routine stainings with toluidine blue no additional adhesive is necessary to keep the plastic sections attached, because the staining occurs in aqueous solution. In case of trouble with other staining procedures the slides may be additionally coated with poly-L-lysine (1 mg/ml) or silane (3-aminopropyltriethyloxy-silane, 1–2%).

Staining

Staining of paraffin sections is more versatile than the staining of resin sections. Routinely, we stain Technovit 7100, Epon 812, or Spurr sections with 1% Toluidin blue/0.5% Borax (TB/Borax) (Figs. 1 and 2, Tourte et al.,1984), which is fast and does not require the rehydration and dehydration steps that are needed for paraffin sections (Sive et al.,2000). However, many cellular details not resolvable in routine paraffin preparations are visible in TB/Borax-stained plastic sections (Figs. 1 and 2; see also Figs. 5 and 6). In addition, a few other histological stainings can be performed on resin sections (e.g. Giemsa, hematoxylin-eosin, PAS, Fig. 3; see also Troyer and Babich,1981).

Figure 3. Different histological stainings on frontal Technovit sections through a 1 cm axolotl larva. (a and b) Giemsa. (c and d) hematoxylin-eosin (HE). (e and f) Periodic acid Schiff (PAS) + hemealaun (HA). Abbreviations: epi, epidermis; int, intestine; mu, trunk muscle.

Taken together, Technovit embedding, sectioning, and staining are convenient and reliable for routine histology of amphibian embryos. Recently, a methacrylate with similar properties (JB4) has been reported to give histology of similar quality in zebrafish embryos (Sullivan-Brown et al.,2011).

Tissue Processing and Embedding in Technovit 7100

A useful method for histological preparations should be suitable even for unexperienced students and scientists and it should not depend on any nearly supernatural skills of “the only technician in the world that can do it!” There are two ways to solve that problem: providing a bona fide simple and reliable, fool-proof protocol for tissue processing or provide state-of-the-art equipment that completely takes over. Even for the latter, however, a suitable protocol or program is needed. In the following, we will discuss both options: embryo preparation using either a microwave-based tissue processor or a simple and reliable but time-consuming by-hand protocol.

Microwave-assisted tissue processing

Microwaves (MW) act by dielectric heating which occurs simultaneously throughout the irradiated sample. This process is called “internal heating” in contrast to conventional external heating, which starts from the periphery of the sample (Giberson and Demaree,1995; Kok and Boon,2003; Leonard and Shepardson,1994). This way, MW-irradiation can accelerate cell and tissue preparations and is used during different steps of histological processing, facilitating the penetration of fixatives and other reagents, accelerating dehydration and infiltration, and enhancing histochemical stainings (Leong and Sormunen,1998). It is also used for demasking antigens before immunocytochemistry and for fast EM-preparation (Kok and Boon,2003; Leong,2009; Leong and Sormunen,1998; Schroeder et al.,2006; Webster,2007). Besides accelerating protocols MW-assisted tissue processing has been shown to improve structure preservation in a number of difficult specimens, such as plants (Zechmann and Zellnig,2009), bone (Laboux et al.,2004), Caenorhabditis elegans (Paupard et al.,2001), or cysts of parasitic protists (Kurth et al., in press).

Most MW-ovens including high-end laboratory MW-ovens are equipped with a multimode chamber in which hot and cold spots are formed. To achieve homogenous microwave-induced internal heating, however, it is crucial to prevent the formation of hot or cold spots within the specimen. Therefore, MW-ovens used for tissue processing come with an accurate map of hot and cold spots and defined sites for placing specimens and additional water loads, hence reproducible energy loads are applied to the samples. A recently developed commercial MW-tissue processor (Leica EM-AMW) offers a different solution. It has a small and unique monomode MW-chamber in which a single microwave is irradiating the samples, resulting in highly reproducable energy loads without the need for additional water loads. The AMW was designed to perform automatic MW-assisted tissue processing for transmission electron microscopy (e.g. Knopf et al., 2001; Kurth et al., in press; Zechmann and Zellnig,2009), but it is also useful for fast histological preparations of frog, salamander and fish embryos, and larva (Schnabel et al.,2011; this report; Figs. 4–6). Using the glycol methacrylate Technovit 7100 (Heraeus-Kulzer) as an embedding medium, embryos can be processed from fixation to resin embedding in less than a day (Figs. 4 and 5; Table 1). We use a variation of that protocol for sample embedding into Epon 812 for ultrastructural studies (Fig. 6; Table 1). In both cases the process is fast and reliable.

Figure 4. Sections of anamniote vertebrate models. (a) Xenopus tailbud embryo, cross-section through the head. (b) Ambystoma mexicanum (axolotl), frontal section through head and anterior trunk of a 1 cm larva. (c) Danio rerio (zebrafish), frontal section through 3-day-old larvae. Samples were processed in the Leica EM-AMW tissue processor and embedded in the methacrylate Technovit 7100, sectioned 2 to 5 μm thick, and stained with TB/Borax.

Figure 5. Histology of Xenopus (a–e) and axolotl (f and g) embryos and larvae after microwave-assisted processing and embedding in Technovit 7100. (a and b) Xenopus tailbud (st. 40), cross-section through the trunk, sectioning plane indicated in the insert. (a) Overview. (b) Details at higher magnification; cap, capillary; dm, dermo-myotome; epi, epidermis; ld, lipid droplet; mel, melanocyte; mu, muscle filaments; myo, myotome; not, notochord; nt, neural tube; nuc, nucleus; vac, vacuole; yp, yolk platelet. (c–e) Xenopus tadpole (stage 45), cross-section through the head, sectioning plane indicated in the insert. (c) Different well preserved tissue types are visible such as cartilage (cart), mesenchyme (mes), muscle (mu) or peripheral nerves (ne); cap, capillary; epi, epidermis. (d) Two-layered epidermis with apical mucus containing granules (ag) and a well preserved basement membrane (bm); fib, fibroblast. (e) muscle fiber, peripheral nuclei (nuc), a satellite cell (sc) and details of the contractile apparatus are clearly visible. (f and g) Axolotl larva, frontal section through the head. (f) Pharynx area with epidermis (epi), mesenchyme (mes), and the developing cartilage (cart) of gill arches. (g) Ciliated cell in the epidermis. Cilia (ci) and basal bodies (bb) are visible; ag, apical granules; bm, basement membrane; ld, lipid droplet; nuc, nucleus; pg, pigment granules.

Figure 6. Histology and EM of larval Xenopus retina after microwave-assisted processing and embedding in Epon 812. (a) Semithin epon section (1 μm) stained with TB/Borax; inl, inner nuclear layer; le, lens; on, optic nerve; onl, outer nuclear layer; os, outer segment of photoreceptor cell; rpe, retinal pigment epithelium. (b) EM-micrograpgh of photoreceptor cells (onl, os) and retinal pigment layer (rpe); yp, yolk platelet. (c) Outer segment (os) at higher magnification, regularly packed membrane discs.

Table 1. Leica EM AMW Protocols for Microwave-Assisted Embedding Into Technovit 7100 (T) or Epon 812 (E)

Step	Reagent	Time	Temp	Power (W)	Mode
Modifications: (1) Samples can be fixed in the AMW, or can be collected overtime, fixed conventionally in modified Karnovsky and stored in half strength fixative before processing them all in one AMW-run. (2) For tough samples (e.g yolky early embryos), additional infiltration steps can be included (e.g. epon/ethanol 1:2- and/or 2:1- mixtures), or the infiltration steps can be duplicated (2 × 1:3; 2 × 1:1; 2 × 3:1), or both. (3) Postfixation may be modified to fit to specific sample needs. Osmium phosphate crystals, for example, sometimes pose a problem, which can be circumvented by using OsO4 dissolved in water or cacodylate buffer. Sectioning and staining: Technovit or epon blocks are mounted according to manufacturers instructions. One to 5 μm sections are prepared with metal or glass knifes, stretched on a warm water bath (Technovit) or on a small drop of water (epon), and mounted to microscope slides. Sections are dried and stained with TB/Borax according to Tourte et al. (1984). Stained sections are mounted with Entellan. cont, continuous; conv, conventional; Eth, ethanol; GA, glutaraldehyde; PFA, paraformaldehyde; r.t., room temperature; Temp, temperature; T+H1: Technovit 7100 infiltration solution (including hardener 1). For embedding hardener 2 (H2) is added. Technical notes: In addition to the incubation times, the incubation temperature (Temp); the microwave power that is used to reach the incubation temperature (Power, W) and the mode of the AMW tissue processor to reach the incubation temperature are indicated. Three different modes are available: “Continuous” (cont), temperature is reached as soon as possible and kept stable until the end of the incubation step; “Slope,” temperature increases steadily and reaches the indicated value at the end of the incubation period; “Pulse,” MW switches on and off regularly (10 s on/50 s off).
Fixation (optional or separate step)	2% GA/2% PFA in 50 mM HEPES	15′	37°C	15 W	Cont
		6′	20°C	0 W	Cont
		15′	37°C	15 W	Cont
		6′	20°C	0 W	Cont
Washes	2 × 100 mM HEPES	2 × 3′	35°C	15 W	Slope
	2 × PBS	2 × 3′	35°C	15 W	Slope
Postfixation	1% OsO4/PBS	30′	45°C	25 W	Pulse
Washes	2 × PBS	2 × 4 ′	37°C	15 W	Slope
	2 × Distilled water	2 × 4 ′	37°C	15 W	Slope
Dehydration	30%, 50%, 70%, 95% Ethanol	4 × 5 ′	37°C	15 W	Cont
	Ethanol absolute	5′ (T)	37°C	14 W	Cont
	2 × Ethanol absolute	2 × 8′ (E)	37°C	14 W	Cont
Infiltration	Resin-Mix T + H1/Eth 1:1 or Resin-Mix E/Eth 1:3	15′ or 20′	40°C	11 W	Cont
	Resin-Mix T + H1/Eth 2:1 or Resin-Mix E/Eth 1:1	15′ or 20′	40°C	11 W	Cont
	Resin-Mix T + H1/Eth 3:1 or Resin-Mix E/Eth 3:1	15′ or 30′	40°C	11 W	Cont
	2 × Resin pure T + H1 or E	2 × 20′(T)	40°C	11 W	Cont
		2 × 30′ (E)	50°C	11 W	Cont
Embedding (conv)	Resin pure (E)	65°C
	Resin pure (T+H1+H2)	r.t.

At the light microscopical level, tissue preservation of AMW-processed samples is excellent, revealing the cells in a near-to-native state. Many intracellular details, such as basal bodies, cilia, details of nuclear structures, sarcomere structure of muscle cells, lipid droplets, the mucus producing apical granules of outer epidermal cells or the pigment granules of melanocytes are clearly visible (Figs. 5 and 6a).

At the ultrastructural level, AMW-embedded tailbud embryos reveal good ultrastructural preservation (Fig. 6b). For early embryos, which have more yolk granules and lipid vacuoles, we recommend the use of half-strength modified Karnovsky (1% GA, 2% PFA in 100 mM HEPES) together with a modified infiltration regime, where the infiltration steps are doubled (Table 1), as has been described for the EM-preparation neural tissues of mammals (Moebius,2010). To further improve infiltration, Spurrs resin may be used instead of Epon 812 (Kurth et al.,2010; Spurr,1969).

Standard method for embedding cells and tissues in Technovit 7100

If only a few samples are to be processed or a fancy tissue processor is missing, embryos can be embedded also by hand (Table 2). This protocol is more time consuming (3 days vs. several hours with the AMW) and the workload is higher, but the results are pretty much the same (see Figs. 1 and 2). The process, easy to learn in one run, is a cheap and reliable routine procedure for tissue analysis in a wide variety of experiments.

Table 2. Protocol for Embedding Into Technovit 7100 (Standard and Low Budget)

Step	Reagent	Time	Temperature
Sectioning and staining: see Table 1. GA, glutaraldehyde; H1, H2, Hardener 1 and 2; PFA, paraformaldehyde.
Fixation	2% GA/2% PFA in 50 mM HEPES	o.n.	4°C
Washes	100 mM HEPES	5 × 10 min	r.t.
	PBS	2 × 5 min	r.t.
Postfixation	1% OsO4 in PBS	2–3 h	On ice
Washes	PBS	5 × 5 min	r.t.
	Distilled water	2 × 5 min	r.t.
Dehydration	30%, 50%, 70% Ethanol in water	20 min each	r.t.
	95% Ethanol in water	30 min	r.t.
	100% Ethanol	30 min	r.t.
Infiltration	Technovit + H1/Ethanol mix: 1:1	1 h	r.t.
	Technovit + H1/Ethanol mix: 2:1	1 h	r.t.
	Technovit + H1/Ethanol mix: 3:1	1 h	r.t.
	Technovit + H1 pure	2 h	r.t.
	Technovit + H1 pure	o.n.	r.t.
Embedding	Technovit + H1 + H2	1–2 h before mounting	r.t.

Standard histology of embryos after in situ hybridization

Whole-mount in situ hybridization (ISH) is one of the most frequently used techniques in the field (Harland,1991; Sive et al., 2001). In many cases it may be helpful to analyze the internal morphology of samples after ISH. For that, samples can be dehydrated, infiltrated, and embedded in Technovit 7100 according to Table 2. Postfixation with OsO₄ can be omitted, because there are not much lipids left to be fixed. Stained embryos stored in methanol can be directly infiltrated in methanol-resin-mixtures and embedded. One has to be aware, however, that the histological information of post-ISH embryos will be very limited. Whole-mount ISH is a 3-day procedure that includes weak fixation (MEMFA, methanol), proteinase K treatment, hybridization at 60 to 65°C, constant mechanical stress, and bleaching with H₂O₂ (Harland,1991). After that kind of ordeal only overall cell shape, nuclei, and yolk platelets can be seen, enough to score for gross internal morphology but not sufficient to extract histological details (Fig. 7).

Figure 7. Histology of an early gastrula after in situ hybridization. Embryos were dehydrated, infiltrated in Technovit 7100, embedded, sectioned (2 μm) and stained with TB/Borax. (a) Xbra-exression in the marginal zone, top: animal view (An), bottom vegetal view (Veg). (b) Post-ISH histology of the gastrula displayed in (a) (bottom), the arrow indicates the blastopore. (c) Part of the animal cap indicated by the rectangle in (b) at higher magnification. (d) Animal cap of an early gastrula after MW-assisted tissue processing (see Table 1). Note the much better preservation. (e) Detail of the post-ISH sample; only nuclei and yolk platelets can be seen, cell shapes are obscured. (f) Detail of the MW-processed, well-fixed sample, intracellular details can be discriminated, cell shapes are clearly detectable.

Immunohistology

Protein localization in relation to cell and tissue morphology provides crucial information for the understanding of mutants or morphants compared with normal embryos. In principle, immunohistology could be done in two different ways: (1) postembedding labeling of sections and (2) whole-mount labeling followed by embedding and subsequent sectioning.

Postembedding labeling of sections

On-section staining could be done either on vibratome-, cryo-, paraffin-, or plastic sections. Vibratome sections (50–300 μm) are bulky sections for immunocytochemistry. They could be stained “whole-mount” and immediately analyzed by light microscopy or re-embedded into resin for further sectioning (Blackiston et al.,2010; Ding et al.,1993; Kurth,2003; Kurth et al.,1996,2010). Immunofluorescent staining of cryosections is a simple and reliable method for many antigens (e.g. Fagotto and Gumbiner,1994; Schohl and Fagotto,2002; for review see Fagotto,1999). In cryosections, antigenicity is only slightly affected because antigens suffer only from mild fixation before they are frozen for sectioning. However, the quality and optical resolution of labeled frozen sections is quite limited. This could be improved by using ultrathin (70–200 nm) cryosections, but for that a high end cryoultramicrotome for sectioning at −80°C to −120°C is needed (Slot and Geuze,2007). Even then the cutting of ultrathin cryosections of early embryos remains difficult (Kurth et al.,2010). The overall quality of paraffin sections is superior to those of thick cryosections, but antigens may suffer from dehydration, infiltration, and high temperature during tissue processing. Finally, plastic sections are the most stable option to achieve high-resolution imaging, but the antibodies do not penetrate the plastic, therefore only the surface of the section can be labeled. If antigens are highly abundant or concentrated in a small structure, that kind of labeling leads to precisely resolved fluorescent signals from a indefinitely thin optical plane (for reviews see Schwarz and Humbel,2007,2009). In the embryo, however, the signal is often below the detection level (Kurth,2003; Kurth et al.,2010). A glycol methacrylate resin designed for on-section labeling of thick sections for light microscopy does not give satisfactory results on embryonic tissues (Susanne Weiche, unpublished data).

Due to these limitations, the only on-section labeling methods routinely used for embryological studies in Xenopus or axolotl are the staining of vibratome sections and of “thick” cryosections.

Whole-mount labeling with subsequent embedding

Whole-mount immunostaining is a powerful tool to study protein distribution in embryos and other bulky samples (e.g. Klymkowsky and Hanken,1991; Linask and Tsuda,2000; Müller,2008). In Xenopus embryos, tissue formation could be monitored by use of differentiation markers such 12/101 (somite and skeletal muscle, Kintner and Brockes,1985) or 3G8 and 4A6 (pronephros, Vize et al.,1995), the activity of cell signaling by antibodies against components of different signaling pathways (e.g. Christen and Slack,1999; Schneider et al.,1996), and ectopically expressed proteins by antibodies against protein tags such as GFP, myc-tag, etc. Due to the opacity of early Xenopus embryos, however, the resolution of such whole-mount preparations is rather limited. To gain more detailed insight into tissue organization in relation to specific protein expression, the analysis of histological sections is required. This could be achieved by the embedding of immunolabeled embryos into Technovit 7100 for sectioning. This method combines, for example, whole-mount immunofluorescence with resin embedding and serial sectioning (Table 3). The method works best with samples that are fixed with Dents fixative (Dent et al.,1989). Although the fixation is rather harsh and results in strong cytoplasmic extraction (Kurth,2003; see also Fig. 1c,d), this morphological defect is less evident in fluorescent images (Figs. 8 and 9). Whole-mount immunofluorescence in combination with resin embedding has been mainly used for the visualization of cell adhesion molecules such as cadherins (Angres et al.,1991; Kurth et al.,1999; Müller et al.,1992,1993; Münchberg et al.,1997; Ogata et al.,2007a,b), catenins (Kurth et al.,1996,1999; Schneider et al.,1993), integrins (Gawantka et al.,1992,1994; Joos et al.,1995,1998; Müller et al.,1993), and tight junctional proteins (Fesenko et al.,2000), for cytoskeletal proteins (Kurth et al.,2000) and for different nuclear antigens (e.g. David et al.,1998; Ellinger-Ziegelbauer and Dreyer,1993; Schwab and Dreyer,1997). The procedure is not applicable to all antigens, because proteins that are not firmly attached to cellular structures such as cytoskeleton, membranes, or the nucleus often get extracted during Dent's fixation. Fixation in 4% PFA before transfer into Dent's fixative improves the preservation of such proteins, but in that case antibody incubation, washing, dehydration, and infiltration steps have to be increased, which results in a preparation protocol lasting nearly 2 weeks (Kurth,2003,2005).

Figure 8. Whole-mount Immunofluorescence followed by embedding in Technovit 7100 and analysis of 2 to 5 μm thick sections. (a) Dorsoanterior mesoderm (mes) migrating along the blastocoel roof (bcr); ect, ectoderm; end, endoderm. Antibodies: anti-fibronectin (FN, red), XB/U-cadherin (green). Nuclei counterstained with DAPI. (b) Xenopus animal cap (explanted at stage 9, fixed at stage 22), overview. (c) same cap at higher magnification. Antibodies: anti-β-catenin (P14L, red), anti-phospho-histone H3 (red, due to the overlap with DAPI it appears magenta), anti-tubulin (green). Nuclei counterstained with DAPI. (d) Fluorescence of membrane-anchored GFP (memGFP) in a tailbud embryo after fixation, embedding, and sectioning. memGFP-mRNA was injected into one blastomere at the two-cell stage; epi, epidermis; myo, myotome; not, notochord; nt, neural tube. (e) Section through the dorsal side of stage 11 gastrula. At the 16-cell stage, memGFP-mRNA and rhodamin-dextran were injected into dorsoanterior animal and vegetal cells, respectively, and the embryos cultured until stage 11. After fixation, they were embedded in Technovit 7100, sectioned and scored for GFP (green) and rhodamine (red). (f) Section through the smooth muscle layers of the Xenopus adult intestine. Smooth muscle cells are surrounded by thin layer of ECM and strongly express β₁-integrin (green); bv, blood vessel; mu, smooth muscle; sub, submucosa. (g–i) Axolotl embryos (stage 24) stained with anti-actin (green) and anti-β-Catenin (red). (g) Somite (som), epi, epdermis. (h) Notochord (not), neural tube (nt) and archenteron (arch). (i) Large endodermal cells.

Figure 9. Correlative histology and fluorescence imaging. Consecutive sections from the same block were collected separately on two sets of slides for TB/Borax staining and for fluorescence, respectively. This way, TB/Borax-histology of one section (a, overview; b, details of the tail bud) can be compared with the fluorescence pattern of the next section (c, same region as in b). The two sections in (b and c) are only 2 μm (section thickness) apart and cover the same area.

Table 3. Protocol for Whole-Mount Immunofluorescence and Embedding in Technovit 7100

Step	Reagent	Time	Temperature
Sectioning and staining: Blocks are mounted and stained as described in Table 1. For fluorescence, sections are counterstained with DAPI and mounted with Mowiol-DABCO. For correlative histology and fluorescence 2 series of alternating sections are collected and stained with TB/Borax and DAPI, respectively. Azide; 0.02% sodium azide; H1, H2, Hardener 1,2; NGS, normal goat serum.
Fixation	80% methanol/20% DMSO (Dent's Fix)	o.n.	− 20°C
Rehydration	80%, 70% methanol in water	15′ each	r.t.
	50%, 30% methanol/PBS	15′ each	r.t.
	3 × PBS	3 × 15′	r.t.
Blocking	20% NGS/PBS/Azide	2–3 h	r.t.
1st antibody	Antibody diluted in 20%NGS/PBS	o.n. to 3 days	4°C to r.t.
Washes	PBS	Short, 5′,10′,15′, 30′, 2 × 1 h, 2 h	r.t.
2nd antibody	Antibody diluted in 20%NGS/PBS  (conjugated to a fluorescent dye;  preferably Fab-fragments)	o.n	4°C
Washes	PBS	Short, 5′,10′,15′, 30′, 2 × 1 h, 2 h	r.t.
Postfixation	4% PFA in PBS	1–2 h	r.t.
Washes	PBS	Short, 5′, 4 × 10′	r.t.
Dehydration	30%, 50%, 70%, 90% ethanol/water;  2 × 100% ethanol	15′ each	r.t.
Infiltration	Technovit/ethanol 1:1	30 ′	r.t.
	Technovit/ethanol 2:1	1 h	r.t.
	Technovit pure	1 h	r.t.
	Technovit pure	o.n.	r.t.
Embedding	Technovit + H1 + H2	1–2 h before mounting	r.t.

Whole-mount immunofluorescence in combination with Technovit embedding and semithin sectioning (1–2 μm) reveals a convenient balance of resolution and signal intensity (Figs. 8 and 9). It is also very useful to monitor in vivo markers such as GFP or fluorescent dextrans after embedding and sectioning of fixed samples. Mem-GFP, for example, retains its fluorescence even after fixation, dehydration, and plastic embedding (Fig. 8d,e). This way, different sets of in vivo markers can be combined with antibody stainings. Finally, the immunofluorescence in one section can be correlated to the overall organization of the sample revealed by histochemical staining with TB/Borax of the adjacent section. Since the sections are just 1 to 2 μm apart from each other, this correlative TB/Borax and fluorescence imaging makes it possible to relate tissue organization with protein distribution in the very same area of the embryo (Fig. 9).

SUMMARY AND PERSPECTIVES

Although paraffin embedding is often the method of choice for the histological inspection of many different tissue samples including embryos, resin embedding is a simple and reliable alternative to obtain high quality histology of amphibian embryos, which are traditionally regarded as “tricky” samples. Using low viscosity methacrylates such as Technovit 7100 and microwave-assisted tissue processing, the procedure is also very fast, rendering it a versatile method for routine histology. In combination with epoxy resins such as Epon 812, the microwave-method is also useful for rapid TEM-preparation of embryo samples. Finally, Technovit 7100 is ideal for embedding and sectioning of samples that were whole-mount labeled by immunofluorescence or in situ hybridization.

Pushing the limits of existing technologies and developing new options for the microanalysis of cells and tissues is important for the thorough analysis of morphants and mutants. The successful use of a MW-assisted tissue processor for histology and EM of Xenopus and axolotl embryos, reported here for the first time, is a promising start for further studies on that technology. MW-processing of amphibian samples using other resins such as LR White, LR Gold, or Spurr remains to be evaluated. In addition, one might envision a MW-assisted whole-mount immunostaining procedure, which might accelerate the time consuming procedure presented here (Table 3) to 1 or 2 days. Similarly, it would be important to establish an easy-to-use method for on-section staining of plastic sections, combining excellent section quality with the possibility to label many different proteins on sections of just one sample. At present, such an option is not available. On-section labeling of Lowicryl-sections for light microscopy of embryonic samples is only rarely successful (Fagotto et al.,1999; Kurth,2003; Kurth et al.,1999; Schwab et al.,1998). Finally, state-of-the-art options for fine structure analysis of cells and tissues such as high pressure freezing and freeze substitution (Studer et al.,2008), immunolabeling of ultrathin cryosections (Slot and Geuze,2007), correlative light and electron microscopy (CLEM, e.g. Grabenbauer et al.,2005; Schwarz and Humbel,2009; van Rijnsoever et al.,2008; Verkade,2008; Vicidomini et al.,2008), and the three-dimensional analysis of cellular and subcellular structures by electron tomography (McIntosh et al.,2005) or by focused ion beam milling and serial block face imaging in a scanning electron microscope (FIB-SEM, Hecking et al.,2009) are largely neglected in the field. These technologies will be, however, crucial for the deeper understanding of cellular organization during development. Sooner or later it will be necessary to apply these methods also to embryonic samples.

Acknowledgements

The authors thank Jana Rieckhoff, Jenny Kürth, and Suzanne Manthey for technical assistance, and Christopher Antos, Hans-Henning Epperlein, Elly Tanaka, John Wallingford, and Michael Brand for the kind donation of reagents or test animals.