9 How histological stains work
All histological staining methods, from acid dyeing to silver impregnation, are based on the same physicochemical principles, as will be described in this chapter. Examples are provided from several of the application areas discussed in this book. Methods using dyestuffs are emphasized, so some background information concerning dyes is provided at the end of the chapter. A generic troubleshooting guide is also appended.
The answers are often complex, reflecting the multiphase nature of the staining process, in which solid cells and tissues interact with solutions of staining reagents. Thus enzyme histochemistry is not ‘merely’ biochemistry, nor is the periodic acid-Schiff (PAS) procedure merely organic chemistry, nor immunostaining merely immunochemistry. In addition to biochemistry, chemistry and immunochemistry, such staining methods are also influenced by selective uptake of reagents into tissues, and selective losses of products and/or reagents from the tissues. Such uptakes and losses depend on both affinity and rate factors. Nomenclature note: staining always involves the visual labeling of some biological entity by attaching, or depositing in its vicinity, a marker of characteristic color or form. The stain is the marker, or the reagent used to generate the marker.
Stain uptake is often due to dye-tissue or reagent-tissue affinities. In the histological staining literature, to say a tissue component has a high affinity for a dye may merely mean that, under the conditions of use, the component becomes intensely stained. Affinity is however also used to describe those attractive forces thought to bind dye to tissue.
Physical chemists use the term affinity in the former sense, and their usage is adopted here. So in this chapter affinity describes the tendency of a stain to transfer from solution onto a section. The affinity’s magnitude depends on every factor favoring or hindering this movement. Stain-tissue, stain-solvent, and stain-stain interactions must all be considered, as indeed must solvent-solvent interactions. This approach initially assumes staining continues until equilibrium is reached, but in practice this is often not achieved. Moreover uptake of dyes and reagents is often multistep, in both space and time. A reagent may initially enter tissues due to, say, coulombic attractions. Once inside it may form covalent bonds with some tissue grouping. Intensity of staining may also be limited by the solubility of a stain in solvent and tissue environments.
|Interactions||Practical examples where the factor is important|
The hydrophobic effect
|Staining systems using aqueous solutions of dyes or other organic reagents; enzyme substrates for example|
|Reagent-reagent interactions||Metachromatic staining with basic dyes, inorganic pigments in Gomori-type enzyme histochemistry, silver impregnation|
Coulombic attractions, which have also been termed salt links or electrostatic bonds, are widely discussed reagent-tissue interactions. These arise from electrostatic attractions of unlike ions, e.g. the colored cations of basic dyes and tissue structures rich in anions such as phosphated DNA, or sulfated mucosubstances (Lyon 1991; Prentø 2009). In practice, the amount of dye ion binding to a tissue substrate depends not only on the charge signs of dye and tissue but also on their magnitude, on the amount of non-dye electrolyte present in the dyebath, and on the ability of the tissue substrate to swell or shrink (Scott 1973; Bennion & Horobin 1974; Goldstein & Horobin 1974b; Horobin & Goldstein 1974).
Such phenomena are important for all ionic reagents, not just dyestuffs, an example being the periodate anions used as the oxidant in the periodic acid-Schiff procedure (Scott & Harbinson 1968). Even initially uncharged tissue substrates acquire ionic character after binding ionic reagents, e.g., glycogen staining by the PAS procedure and with Best’s carmine.
Van der Waals’ forces include such intermolecular attractions as dipole-dipole, dipole-induced dipole and dispersion forces. These occur between all reagents and tissue substrates, but since molecules with extensively delocalized electronic systems tend to have larger dipoles and are more polarizable, van der Waals’ forces are usually most critical when tissues or stains contain such moieties.
Consequently proteins rich in tyrosine and tryptophan residues, and nucleic acids with their heterocyclic bases, favor van der Waals’ attractions, as do the large aromatic systems of stains such as bisazo dyes and bistetrazolium salts, halogenated dyes (such as rose Bengal and phloxine), and enzyme substrates based on naphthyl and indoxyl with extended conjugation (Horobin & Bennion 1973). For instance, van der Waals’ attractions contribute substantially to stain-tissue affinity when staining elastic fibers – rich in aromatic desmosine and isodesmosine residues – with polyaromatic acid and basic dyes such as Congo red and orcein.
Hydrogen bonding is a dye-tissue attraction arising when a hydrogen atom lies between two electronegative atoms (e.g. oxygen or nitrogen), though it is covalently bonded only to one. Water is hydrogen bonded extensively to itself, forming the clusters important for the hydrophobic effect discussed below, and also to other molecules with hydrogen bonding groups, such as many dyes and tissue components. As there are many more molecules of water present than dye, hydrogen bonding is not usually important for stain-tissue affinity when aqueous solvents are used. An exception arises when hydrogen bonding is particularly favored by the substrate, as is the case with connective tissue fibers (Prentø 2007). In wholly or partially non-aqueous solutions, hydrogen bonding can also be significant, as with Best’s carmine stain for glycogen (Horobin & Murgatroyd 1970).
Covalent bonding between tissue and stain also occurs, which bonds may be regarded merely as another source of stain-tissue affinity. Practical reactive methods, e.g. the Feulgen nuclear and the periodic acid-Schiff procedures, are described elsewhere in this volume. The polar covalent bonds between metal ions and ‘mordant’ dyes are a special case. Dye-tissue binding due to such bonds has been termed mordanting, but is of uncertain status. The characteristic staining properties of mordant dyes may have other, or at least additional, causes. For instance, unlike most cationic dyes used as biological stains, cationic metal-complex dyes are usually markedly hydrophilic (Bettinger & Zimmermann 1991) and consequently resist extraction into alcoholic dehydration fluids (Marshall & Horobin 1973).
A major contribution to stain-tissue affinity when using organic reagents or dyes in aqueous solution is the hydrophobic effect. This is the tendency of hydrophobic groupings (such as leucine and valine side chains of proteins; or biphenyl and naphthyl groupings of enzyme substrates and dyes) to come together, even though initially dispersed in an aqueous environment. The process occurs because water is a highly structured liquid. Many water molecules are held together by hydrogen bonding (see above) in transient clusters, whose formation is favored by the presence of hydrophobic groups. Processes breaking clusters into individual water molecules occur spontaneously, because these events increase the entropy of the system. Consequently, removing cluster-stabilizing hydrophobic groups from contact with water, by placing them in contact with each other, is thermodynamically favored. For background on the hydrophobic effect see textbooks of biochemistry or chemical thermodynamics (Tanford 2004). The effect becomes more important as the substrate and reagent become more hydrophobic, as with staining of fats by Sudan dyes. When these hydrophobic dyes are applied from substantially aqueous solutions, the hydrophobic effect will be a major contribution to affinity. It should be noted that the hydrophobic effect is sometimes termed hydrophobic bonding, even though no special bonds are involved (only water-water hydrogen bonds and, sometimes, stain-tissue van der Waals’ attractions).
Some staining procedures involving Sudan dyes use solvents in which water is absent, or is only a minor constituent. Here the second law of thermodynamics – the tendency of a system to change spontaneously to maximize its disorder (i.e., for entropy to increase as described in texts of chemical thermodynamics) – may again be invoked. Dye dispersed through fat and solvent constitutes a more disordered system than dye restricted to a single phase. Consequently, dye becomes dispersed and staining occurs. Of course, such increases in entropy involving substrate and dye occur in all types of staining system.
Dye-dye interactions can also contribute to affinity. Dye molecules tend to attract each other, forming aggregates. Even in dilute solutions, and especially in aqueous solutions where the hydrophobic effect is important, dimers or larger aggregates of dye ions are often present. van der Waals’ attractions (see above) between dye molecules will be important in both aqueous and non-aqueous solutions. Dye aggregation increases with concentration, e.g. when high dye concentrations build up on tissue sections. With basic (cationic) dyes this occurs on substrates of high negative charge density, e.g. sulfated polysaccharides in mast cell granules, a classic site for metachromatic staining by dyes such as toluidine blue. This phenomenon occurs because dye aggregates have spectral properties different from the monomeric dye. That dye-dye interactions contribute to affinity in tissue sections was demonstrated quantitatively by Goldstein (1962).
Other examples of stain-stain interactions contributing to affinity include metallic nano- and micro-crystals generated by gold or silver impregnation, metal sulfide precipitates formed in Gomori-type enzyme histochemistry, and the purple azure-eosin charge transfer complex produced during Romanowsky-Giemsa staining of cell nuclei (Horobin 2011).
Some stains are not taken up by their tissue targets. In negative staining the shapes of structures are disclosed by outlining or filling them with a stain. Examples include visualizing individual microorganisms using nigrosine, and demonstrating canaliculi of bone matrix using picro-thionine.
Sometimes stains are taken into live creatures, in ways that reflect the biochemical composition and physiological activities of the living cell or organism. Traditionally termed vital staining or supravital staining, this is now usually described as the use of fluorescent probes. Over the past 20 years this methodology has undergone a renaissance; for a recent overview see Celis (2006).
A little-discussed, but nevertheless important, property of stains is their solubility. For instance, when staining fat with Sudan dyes, an upper limit of staining intensity is set by the solubility of the dyes in the target substance, and is also influenced by the solubility in the staining bath solvent. Solubility is also involved in dye retention after staining: see below. The solubility of a staining reagent has complex causes. Put simply, the stronger the reagent-reagent interactions, the lower the solubility. For a general discussion of solubility see physicochemical texts such that of Letcher (2007).
This occurs because the stains either have a very high affinity for tissue elements and/or low affinity for processing fluids and mounting media, or at least dissolve in these latter materials very slowly. To illustrate these points, consider some common stains.
Pigments such as the Prussian blue generated in the Perls’ method for iron, and the lead sulfide produced in Gomori-style enzyme histochemistry, are virtually insoluble in standard solvents. This is also true for microcrystals of silver and gold produced by metal impregnation. Some organic pigments are less satisfactory. Thus, azo dyes, formazans and substituted indigos produced as final reaction products in enzyme histochemistry have low solubility in water, but may dissolve in hydrophobic media such as alcohols, xylene, and polystyrene. In such cases hydrophilic mounting media are used, and staining of lipid-rich tissue elements should be regarded as possibly artifactual.
Solubilities of formazans and azodyes are sometimes reduced by in situ conversion to metal complexes. Other routine metal complex stains are the aluminum, chromium and iron complexes of hematein, and the chromium complex of gallocyanine. These metal complex dyes are not readily removed from tissues by routine processing fluids or mounting media (see above).
This contrasts with routine basic (cationic) dyes such as crystal violet or methylene blue, which freely and rapidly dissolve in the lower alcohols. Routine acid (anionic) dyes, such as eosin Y or orange G, are often less soluble in alcohols, as indeed are hydrophilic basic dyes with large aromatic systems, such as alcian blue. Non-ionic dyes such as Sudan fat stains are soluble in common dehydrating agents and clearing solvents, and in resin mountants. Note: structures of exemplar hydrophilic and lipophilic basic dyes are shown in Figure 9.1.
Sections stained with routine basic dyes must therefore be dehydrated rapidly through the alcohols, or by the use of non-alcoholic solvents, or by air-drying, whereas dehydration is less critical with acid dyes. Sections stained with either acid or basic dyes are usually mounted in non-aqueous media to prevent extraction of dye. Alternatively, dyes may be immobilized, for example, by formation of metal coordination compounds, phosphotungstates or iodine complexes. Non-ionic dyes must be mounted in aqueous media.
This question of selectivity is fundamental to histochemistry, and even routine oversight methods such as hematoxylin and eosin (H&E), Papanicolaou and Romanowsky-Giemsa stains distinguish nuclei from cytoplasm. So we must discover what factors control such selectivities.
Both these factors influence staining. However, in the absence of quantitative investigation, they are not readily distinguishable – so will here be discussed as a single effect. Stain-tissue affinities and numbers of binding sites present in tissues can vary independently.
Sudan dyes can be used as an example. These have a high affinity for fat but low affinity for the surrounding hydrated proteins. Alternatively one may consider staining systems in which covalent bonds are formed. Reagents give colored products only with a limited range of tissue chemical groupings. Thus, the acid hydrolysis-Schiff reagent sequence of the Feulgen nucleal technique gives red derivatives only with DNA.
An understanding of staining systems often requires consideration of patterns of affinities. With the traditional acid dye-basic dye pairs (H&E, Papanicolaou, and Romanowsky) the negatively charged acid dyes have high affinities for tissue structures carrying cationic charges (proteins, under acidic conditions). However they have low affinities for structures carrying negative charges (those rich in sulfated glycosaminoglycans, or in phosphated nucleic acids), with the opposite being the case for basic dyes. This produces two-tone staining patterns in which cytoplasm contrasts with nuclear material.
Practical staining conditions maximize selective affinities. Basic dyes are applied from neutral or acidic solutions, since under alkaline conditions proteins carry an overall negative charge and so may also bind basic dyes. Affinities are also influenced by varying the concentration of inorganic salt present. The various aluminum-hematoxylin, for instance, differ substantially in this regard. The critical electrolyte concentration methodology (Scott 1973) and several other empirical procedures are based on control of electrolyte content. However, staining that distinguishes two structures is still possible even when stain-tissue affinities and the number of stain-binding sites are the same. This is because rate of reagent uptake, or rate of subsequent reaction, or rate of loss of reagent or product, may not be the same in the two structures.
Progressive dyeing methods may be rate controlled, for instance mucin staining using alcian blue or colloidal iron. Selectivity requires short periods of dyeing during which only fast-staining mucins acquire color (Goldstein 1962, Goldstein & Horobin 1974a). If staining is prolonged, additional basophilic materials such as nuclei and RNA-rich cytoplasms can also stain. Dyes used in this way are often large in size, and hence diffuse slowly, maximizing the control possible via differential rate effects.
Selective staining by reactive reagents, yielding colored derivatives, may depend on differential rates of reaction. For instance, periodic acid can oxidize a variety of substrates present in tissues. However, in histochemical applications of the periodic acid-Schiff procedure short oxidation time limits subsequent coloration to fast-reacting 1,2-diol groupings of polysaccharides. Enzyme histochemistry provides further examples of reaction rate controlling selectivity. When incubating at low pH, hydrolysis of an organic phosphate is rapid in tissues containing acid phosphatases whereas in structures containing alkaline phosphatases, with higher pH optima, hydrolysis rates are slow.