Heat fixation

The simplest form of fixation is heat. Boiling or poaching an egg precipitates the proteins and, on cutting, the yolk and egg white can be identified separately. Each component is less soluble in water after heat fixation than the same component of a fresh egg. Picking up a frozen section on a warm microscope slide both attaches the section to the slide and partially fixes it by heat and dehydration. Even though adequate morphology could be obtained by boiling tissue in normal saline, in histopathology heat is primarily used to accelerate other forms of fixation as well as the steps of tissue processing.

Microwave fixation

Microwave heating speeds fixation and can reduce times for fixation of some gross specimens and histological sections from more than 12 hours to less than 20 minutes (Anonymous 2001; Kok & Boon 2003; Leong 2005). Microwaving tissue in formalin results in the production of large amounts of dangerous vapors, so in the absence of a hood for fixation, or a microwave processing system designed to handle these vapors, this may cause safety problems. Recently, commercial glyoxal-based fixatives which do not form vapors when heated at 55°C have been introduced as an efficient method of microwave fixation.

Freeze-drying and freeze substitution

Freeze-drying is a useful technique for studying soluble materials and small molecules; tissues are cut into thin sections, immersed in liquid nitrogen, and the water is removed in a vacuum chamber at −40°C. The tissue can be post-fixed with formaldehyde vapor. In substitution, specimens are immersed in fixatives at −40°C, such as acetone or alcohol, which slowly remove water through dissolution of ice crystals, and the proteins are not denatured; bringing the temperature gradually to 4°C will complete the fixation process (Pearse 1980). These methods of fixation are used primarily in the research environment and are rarely used in the clinical laboratory setting.

Coagulant fixatives

Both organic and non-organic solutions may coagulate proteins, making them insoluble. Cellular architecture is maintained primarily by lipoproteins and by fibrous proteins such as collagen; coagulating such proteins maintains tissue histomorphology at the light microscopic level. Unfortunately, because coagulant fixatives result in cytoplasmic flocculation as well as poor preservation of mitochondria and secretory granules, such fixatives are not useful in ultrastructural analysis.

Dehydrant coagulant fixatives

The most commonly used coagulating fixatives are alcohols (e.g. ethanol, methanol) and acetone. Methanol is closer to the structure of water than ethanol. Ethanol therefore competes more strongly than methanol in the interaction with hydrophobic areas of molecules; thus, coagulant fixation begins at a concentration of 50–60% for ethanol but requires a concentration of 80% or more for methanol (Lillie & Fullmer 1976). Removal and replacement of free water from tissue by any of these agents has several potential effects on proteins within the tissue. Water molecules surround hydrophobic areas of proteins and, by repulsion, force hydrophobic chemical groups into closer contact with each other and hence stabilize hydrophobic bonding. By removing water, the opposite principle weakens hydrophobic bonding. Similarly, molecules of water participate in hydrogen bonding in hydrophilic areas of proteins; so removal of water destabilizes this hydrogen bonding. Together, these changes act to disrupt the tertiary structure of proteins. In addition, with the water removed, the structure of the protein may become partially reversed, with hydrophobic groups moving to the outside surface of the protein. Once the tertiary structure of a soluble protein has been modified, the rate of reversal to a more ordered soluble state is slow and most proteins after coagulation remain insoluble even if returned to an aqueous environment.

Disruption of the tertiary structure of proteins, i.e. denaturation, changes their physical properties, potentially causing insolubility and loss of function. Even though most proteins become less soluble in organic environments, up to 13% of protein may be lost, for example with acetone fixation (Horobin 1982). Factors that influence the solubility of macromolecules include:

Alcohol denatures protein differently, depending on the choice and concentration of alcohol, the presence of organic and non-organic substances, and the pH and temperature of fixation. For example, ethanol denatures proteins > phenols > water and polyhydric alcohols > monocarboxylic acids > dicarboxylic acids (Bhakuni 1998).

Other types of coagulant fixative

Acidic coagulants such as picric acid and trichloroacetic acid change the charges on the ionizable side chains, e.g. (NH₂ → NH₃⁺) and (COO⁻ → COOH), of proteins and disrupt electrostatic and hydrogen bonding. These acids also may insert a lipophilic anion into a hydrophilic region and hence disrupt the tertiary structures of proteins (Horobin 1982). Acetic acid coagulates nucleic acids but does not fix or precipitate proteins; it is therefore added to other fixatives to prevent the loss of nucleic acids. Trichloroacetic acid (Cl₃CCOOH) can penetrate hydrophobic domains of proteins and the anion produced (CCOO⁻) reacts with charged amine groups. This interaction precipitates proteins and extracts nucleic acids. Picric acid or trinitrophenol slightly dissolves in water to form a weak acid solution (pH 2.0). In reactions, it forms salts with basic groups of proteins, causing the proteins to coagulate. If the solution is neutralized, precipitated protein may redissolve. Picric acid fixation produces brighter staining, but the low pH solutions of picric acid may cause hydrolysis and loss of nucleic acids

Non-coagulant cross-linking fixatives

Several chemicals were selected as fixatives secondary to their potential actions of forming cross-links within and between proteins and nucleic acids as well as between nucleic acids and proteins. Cross-linking may not be a major mechanism at current short times of fixation, and therefore ‘covalent additive fixatives’ may be a better name for this group. Examples include formaldehyde, glutaraldehyde, and other aldehydes, e.g. chloral hydrate and glyoxal, metal salts such as mercuric and zinc chloride, and other metallic compounds such as osmium tetroxide. Aldehyde groups are chemically and biologically reactive and are responsible for many histochemical reactions, e.g. free aldehyde groups may be responsible for argentaffin reactions (Papanikolau & Kokkinidis 1997).

Formaldehyde fixation

Formaldehyde in its 10% neutral buffered form (NBF) is the most common fixative used in diagnostic pathology. Pure formaldehyde is a vapor that, when completely dissolved in water, forms a solution containing 37–40% formaldehyde; this aqueous solution is known as ‘formalin’. The usual ‘10% formalin’ used in fixation of tissues is a 10% solution of formalin; i.e., it contains about 4% weight to volume of formaldehyde. The reactions of formaldehyde with macromolecules are numerous and complex. Fraenkel-Conrat and his colleagues, using simple chemistry, meticulously identified most of the reactions of formaldehyde with amino acids and proteins (French & Edsall 1945; Fraenkel-Conrat & Olcott 1948a, 1948b; Fraenkel-Conrat & Mecham 1949). In an aqueous solution formaldehyde forms methylene hydrate, a methylene glycol as the first step in fixation (Singer 1962).

Methylene hydrate reacts with several side chains of proteins to form reactive hydroxymethyl side groups (CH₂OH). If relatively short fixation times are used with 10% neutral buffered formalin (hours to days), the formation of hydroxymethyl side chains is probably the primary and characteristic reaction. The formation of actual cross-links may be relatively rare at the currently used relatively short times of fixation.

Formaldehyde also reacts with nuclear proteins and nucleic acids (Kok & Boon 2003; Leong 2005). It penetrates between nucleic acids and proteins and stabilizes the nucleic acid-protein shell, and it also modifies nucleotides by reacting with free amino groups, as it does with proteins. In naked and free DNA, the cross-linking reactions are believed to start at adenine-thymidine (AT)-rich regions and cross-linking increases with increasing temperature (McGhee & von Hippel 1975a, 1975b, 1977a, 1977b). Formaldehyde reacts with CC and SH bonds in unsaturated lipids, but does not interact with carbohydrates (French & Edsall 1945; Hayat 1981).

The side chains of peptides or proteins that are most reactive with methylene hydrate, and hence have the highest affinity for formaldehyde, include lysine, cysteine, histidine, arginine, tyrosine, and reactive hydroxyl groups of serine and threonine (see Table 4.1) (Means & Feeney 1995).

Gustavson (1956) reported that one of the most important cross-links in ‘over-fixation’, i.e. in tanning, is that between lysine and the amide group of the protein backbone. Due to the shorter fixation times of current diagnostic pathological and biological applications, cross-linking reactions with the protein backbone are unlikely to occur (French & Edsall 1945; Fraenkel-Conrat et al. 1945, 1947; Fraenkel-Conrat & Olcott 1948a, 1948b; Fraenkel-Conrat & Mecham 1949; Gustavson 1956).

Reversibility of formaldehyde-macromolecular reactions

The reactive groups may combine with hydrogen groups or with each other, forming methylene bridges. If the formalin is washed away, reactive groups may rapidly return to their original states, but any bridging that has already occurred may remain.

Washing for 24 hours removes about half of reactive groups, and 4 weeks of washing removes up to 90% (Helander 1994). This suggests that actual cross-linking is a relatively slow process, so, in the rapid fixation used in diagnostic pathology, most ‘fixation’ with formaldehyde prior to tissue processing stops with the formation of reactive hydroxymethyl groups.

For long-term storage in formalin, the reactive groups may be oxidized to the more stable groups (e.g. acids NHCOOH) which are not easily removed by washing in water or alcohol. Thus, following fixation, returning the specimen to water or alcohol further reduces the fixation of the specimen, because the reactive groups produced by the initial reaction with formalin may reverse and be removed. Although it was initially thought that cross-linking was most important in the fixation of tissue for biological uses (based on the limited number of cross-links over short periods of fixation), it is likely that formation of these hydroxymethyl groups actually denatures macromolecules and renders them insoluble. As these washing experiments have not been reproduced, the actual mechanisms and their importance to fixation by formaldehyde are uncertain. As well as simple washing under running water, over-fixation of tissue may be partially corrected by soaking the tissue in concentrated ammonia plus 20% chloral hydrate (Lhotka & Ferreira 1949). Fraenkel-Conrat and his colleagues frequently noted that the addition and condensation reactions of formaldehyde with amino acids and proteins were unstable and could be reversed easily by dilution or dialysis (Fraenkel-Conrat et al. 1945, 1947; Fraenkel-Conrat & Olcott 1948a, 1948b; Fraenkel-Conrat & Mecham 1949).

The principal type of cross-link in short-term fixation is thought to be between the hydroxymethyl group on a lysine side chain and arginine (through secondary amino groups), asparagine, glutamine (through secondary amide groups), or tyrosine (through hydroxyl group) (Tome et al. 1990). For example, a lysine methyl hydroxyl amine group can react with an arginine group to form a lysineCH₂–arginine cross-link; similarly, a tyrosine methyl hydroxyl amine group can bind with a cysteine group to form a tyrosineCH₂–cysteine cross-link. Each of these cross-links between macromolecules has a different degree of stability, which can be modified by the temperature, pH, and type of the environment surrounding and permeating the tissue (Eltoum et al. 2001b). The time to saturation of human and animal tissues with active groups by formalin is about 24 hours, but cross-linking may continue for many weeks (Helander 1994).

When formaldehyde dissolves in an unbuffered aqueous solution, it forms an acid solution (pH 5.0–5.5) because 5–10% of commercially available formaldehyde is formic acid. Acid formalin may react more slowly with proteins than NBF because amine groups become charged (e.g. N⁺H₃). In solution, this requires a much lower pH than 5.5. However, the requirement for a lower pH to produce N⁺H₃ groups may not be equivalent to that required in peptides. Acid formalin also preserves immunorecognition much better than NBF (Arnold et al. 1996), and indeed the success of Taylor in the early days of immunocytochemistry to demonstrate immunoglobulins in paraffin-processed tissue sections, most probably relied on the fixation of the tissues in acid formalin (Taylor et al. 1974). The disadvantage of using acid formalin for fixation is the formation of a brown-black pigment with degraded hemoglobulin. This heme-related pigment, which forms in tissue, is usually not a great problem unless patients have a blood abnormality (e.g. sickle cell disease, malaria).

Formaldehyde primarily preserves peptide-protein bonds and the general structure of cellular organelles. It can interact with nucleic acids but has little effect on carbohydrates and preserves lipids if the solutions contain calcium (Bayliss High & Lake 1996).

Glutaraldehyde fixation

Less is known about glutaraldehyde’s biological reactions and effects compared to formaldehyde, as it has not been used as widely in biological applications. Glutaraldehyde is a bifunctional aldehyde that probably combines with the same reactive groups as does formaldehyde. In aqueous solutions glutaraldehyde polymerizes, forming cyclic and oligomeric compounds (Hopwood 1985), and it is also oxidized to glutaric acid. To aid in stability, it requires storage at 4°C and at a pH of around 5 (Hopwood 1969).

Unlike formaldehyde, glutaraldehyde has an aldehyde group on both ends of the molecule. With each reaction of the first group, an unreacted aldehyde group may be introduced into the protein and these aldehyde groups can act to further cross-link the protein. Alternatively, the aldehyde groups may react with a wide range of other histochemical targets, including antibodies, enzymes, or proteins. The reaction of glutaraldehyde with an isolated protein, such as bovine serum albumin, is fastest at pH 6–7, and is faster (Habeeb 1966), and results in more cross-linking than formaldehyde (Habeeb 1966; Hopwood 1969). Cross-linking is irreversible and withstands acids, urea, semicarbazide, and heat (Hayat 1981). Like formaldehyde, reactions with lysine are the most important for forming cross-links.

Extensive cross-linking by glutaraldehyde results in better preservation of ultrastructure, but this method of fixation negatively affects immunohistochemical methods and slows the penetration by the fixative. Thus, any tissue fixed in glutaraldehyde must be small (0.5 mm maximum) and, unless the aldehyde groups are blocked, increased background staining will result if several histochemical methods are used (Grizzle 1996a). Glutaraldehyde does not react with carbohydrates or lipids unless they contain free amino groups as are found in some phospholipids (Hayat 1981). At room temperature glutaraldehyde does not cross-link nucleic acids in the absence of nucleohistones but it may react with nucleic acids at or above 45°C (Hayat 1981).

Osmium tetroxide fixation

Osmium tetroxide (OsO₄), a toxic solid, is soluble in water as well as non-polar solvents and can react with hydrophilic and hydrophobic sites including the side chains of proteins, potentially causing cross-linking (Hopwood et al. 1990). The reactive sites include sulfydryl, disulfide, phenolic, hydroxyl, carboxyl, amide, and heterocyclic groups. Osmium tetroxide is known to interact with nucleic acids, specifically with the 2,3-glycol moiety in terminal ribose groups and the 5,6 double bonds of thymine residues. Nuclei fixed in OsO₄ and dehydrated with alcohol may show prominent clumping of DNA. This unacceptable artifact can be prevented by pre-fixation with potassium permanganate (KMnO₄), post-fixation with uranyl acetate, or by adding calcium ions and tryptophan during fixation (Hayat 1981). The reaction of OsO₄ with carbohydrates is uncertain (Hayat 1981). Large proportions of proteins and carbohydrates are lost from tissues during osmium fixation; some of this may be due to the superficial limited penetration of OsO₄ (i.e. <1 mm) into tissues or its slow rates of reaction. In electron microscopy, this loss is minimized by initial fixation of tissue in glutaraldehyde.

The best characterized reaction of osmium is its reaction with unsaturated bonds within lipids and phospholipids. In this reaction, osmium in its +8 valence state converts to a +6 valence state, which is colorless. If two unsaturated bonds are close together there may be cross-linking by OsO₄. Although the complex is colorless at this point, the typical black staining of membranes expected from fixation with osmium requires the production of osmium dioxide (OsO₂·2H₂O). Osmium dioxide is black, electron dense, and insoluble in aqueous solution; it precipitates as the above unstable compounds break down and becomes deposited on cellular membranes. The breakdown of osmium +6 valence complexes to osmium dioxide (+4 valence state) is facilitated by a reaction with solutions of ethanol.

In addition to its use as a secondary fixative for electron microscope examinations, OsO₄ can also be used to stain lipids in frozen sections. Osmium tetroxide fixation causes tissue swelling which is reversed during dehydration steps. Swelling can also be minimized by adding calcium or sodium chloride to osmium-containing fixatives (Hayat 1981).

Cross-linking fixatives for electron microscopy

Cell organelles such as cytoplasmic and nuclear membranes, mitochondria, membrane-bound secretory granules, and smooth and rough endoplasmic reticulum need to be preserved carefully for electron microscopy. The lipids in these structures are extracted by many fixatives with dehydrants (e.g. alcohols). Therefore for ultrastructural examination it is important to use a fixative that does not solubilize lipids. The preferred fixatives are a strong cross-linking fixative such as glutaraldehyde, a combination of glutaraldehyde and formaldehyde, or Carson’s modified Millonig’s, followed by post-fixation in an agent that further stabilizes as well as emphasizes membranes such as OsO₄.

Mercuric chloride

Historically, mercuric chloride was greatly favored for its qualities of enhancing the staining properties of tissues, particularly for trichrome stains. However, it is now rarely used in the clinical laboratory due to the health and safety issues involved with the use of a mercury-containing fixative, and also due to the reduced reliance on ‘special stains’. A further major disadvantage of mercuric chloride fixation is the inevitable formation of deposits of intensely black precipitates of mercuric pigment in the tissues. This subsequently gives them inferior value for immunohistochemical and molecular studies. In recently fixed tissues, these precipitates can be readily removed by a Lugol’s iodine step in the staining procedure, followed by bleaching of the section in sodium hypochlorite solution (Hypo). However, this is not effective on mercuric chloride fixed tissues which have been stored for a number of years as paraffin blocks. In these tissues, retrospective analysis by immunohistochemistry and molecular techniques becomes unreliable due to the formation of much larger aggregates of mercuric pigment which cannot be removed subsequently by Lugol’s iodine. The chemistry of fixation using mercuric chloride is not well understood. It is, however, known that mercuric chloride reacts with ammonium salts, amines, amides, amino acids, and sulfydryl groups, and hardens tissues. It is especially reactive with cysteine, forming a dimercaptide (Hopwood 2002) and acidifying the solution:

If only one cysteine is present, a reactive group of RSHgCl is likely.

Mercury-based fixatives are toxic and should be handled with care. They should not be allowed to come into contact with metal, and should be dissolved in distilled water to prevent the precipitation of mercury salts. Mercury-containing chemicals are an environmental disposal problem. These fixatives penetrate slowly, so specimens must be thin, and mercury and acid formaldehyde hematein pigments may deposit in tissue after fixation. Mercury fixatives (Hopwood 1973) are no longer used routinely except by some laboratories for fixing hematopoietic tissues (especially B5). A potential replacement for mercuric chloride is zinc sulfate. Special formulations of zinc sulfate in formaldehyde replacing mercuric chloride in B5 may give better nuclear detail than formaldehyde alone and improve tissue penetration (Carson 1990).