Abstract
The problems of major concern in immunohistochemical practice are discussed in the following order: (a) the mechanism of the Ag-Ab reaction in fixed tissue as opposed to the in vitro reaction; (b) the chemistry of fixation and its influence on the final result of the immunohistochemical reaction; (c) the various procedures used for antigen retrieval in formaldehyde-fixed tissue; and (d) the consideration of the possible mechanism underlying heat-induced antigen retrieval. Suggestions for further work to attempt a clarification of the mechanism involved in the Ag-Ab reaction in immunohistochemistry resorting to existing histochemical methods for the demonstration of protein side groups are presented, together with some examples already published.
Mechanism of the Ag-Ab Reaction in Formaldehyde-fixed Specimens
T
The antibody macromolecule that is diluted in the buffer covering the tissue section is free, during the incubation period, to wander in that buffer, obeying the laws of Brownian kinetics. Of course, the final result must be a consequence of the various factors intervening in these kinetics, e.g., pH, concentration, buffer constituents, salts added, dilution used, temperature. One of the most critical parameters is pH because the isoelectric point of the IgG used in this procedure is in a range of pH between 6.5 and 8.5 (Seppala et al. 1981) and “the optimal immunoreactivities of homogenous monoclonal antibodies could be compromised if the pH… of diluent buffer render the electrostatic charges of the epitope or paratope identical…” (Boenisch 1999). This point will be discussed again below.
Whereas the conformation of the immunoglobulin protein used in the Ag-Ab reaction in tissues is well preserved, usually in its native form, the conformation of the antigen protein, located in the tissue, cannot be considered intact. It has suffered the effect of the formaldehyde fixative and may have been modified in its constitution or its conformation. Even when, according to Mason and O'Leary (1991), “the original secondary structure of proteins, in formaldehyde-fixed tissue, has been effectively locked in by the process of fixation,” routine formaldehyde fixation and the rest of the procedure for paraffin embedding might have altered the protein macromolecules, consequently decreasing the intensity of the final reaction in the IHC procedure.
The kinetics of two macromolecules, such as the antigen and the antibody, as described by Davies et al. (1990), both of which are free to move in the dilution liquid, has no similarity whatsoever to the kinetics of the two macromolecules acting in IHC. In this case, one of them is “anchored” in the tissue section, somewhat altered by the action of the fixative, and is bound to many other macromolecules in a way not clearly understood. The other is in its native state and is free to move in the dilution buffer. Of course, the basic physical and chemical laws are always the same: electrovalent attraction at a certain distance between the macromolecules and van der Waals forces, hydrogen bond formation and hydrophobic adhesion when the distance is smaller. However, the immobile state of the antigen plus the changes that have taken place in its protein side groups are factors that may have modified and conditioned the binding. Its possible association with other proteins or macromolecules in their surroundings may be an obstacle to its binding to the antibody (Bell et al. 1987). The situation just depicted cannot be compared with that of immune precipitation or Western blotting. In these cases, there is neither formaldehyde fixation nor the bizarre variety of many different macromolecules of many kinds, disposed in a section and interconnected with each other.
Summarizing, in IHC the Ag-Ab reaction is chiefly conditioned by the following facts: (a) the state of fixation of the antigen in the tissue section, with the existence of chemical bridges binding reactive side groups of proteins; (b) the relative freedom of the antibody macromolecules free to move in the dilution buffer covering the sections; and (c) the pH of the buffer used for dilution of the primary antibody. The range of best pH for IgG reaction used in the IHC procedure is between 6.5 and 8.5, as will be shown below.
The complex action of each one of the above factors may make it very difficult to interpret the results of the antigen retrieval (AR) procedure that is now so frequently used as a way to increase the final intensity of the color reaction or, even, to make a reaction that was negative without heat positive after its application.
The Chemistry of Fixation
Formaldehyde is the universal fixative in anatomic pathology. Consequently most of the research efforts have concentrated on this method of fixation. As stated by Pearse (1980), it is essential to know as precisely as possible the effect of this fixative on the reactive groups of the various tissue components. Because proteins are the most important of these components, from the point of view of tissue constitution and structure, the action of formaldehyde on these components has attracted many studies.
Formaldehyde is usually used in a 10% solution of the commercial one, usually in neutral phosphate buffer. Formaldehyde exists in this solution in the form of methylene glycol, CH2(OH)2.
According to French and Edsall (1945), the hydroxymethyl compound R-CH2(OH) could condense with a further H atom to form a methylene bridge R-CH2-R′. The formation of methylene bridges had previously been suggested by Meyer (1929). Other possibilities have been suggested, but this one may serve to understand the formation of multiple bonds among proteins in the tissue sections by the establishment of methylene bridges between protein end-groups. In a kinetic study of formaldehyde binding to tissue, Helander (1994) has demonstrated that after 1 day a plateau of formaldehyde concentration in tissue was obtained after 24 hr and that two processes occur during this period: rapid diffusion of the formaldehyde and methylene glycol and binding of formaldehyde to the tissue. If the dimension of the tissue specimen is under the 2–3 mm prescribed, at least for one of the dimensions, 24 hr will consequently be enough for a good fixation.
The groups that may be involved in these bindings are amino, imino, amido, peptide, guanidyl, and carboxyl, SH, and aromatic rings. The investigations of Nitschmann and Hadorn (1944) and the group of Fraenkel-Conrat and colleagues (Fraenkel-Conrat et al. 1947; Fraenkel-Conrat and Olcott 1948a,b; Fraenkel-Conrat and Mecham 1949) have confirmed these assumptions. The chemistry of formaldehyde fixation and its effects on the IHC reaction have been analyzed by Puchtler and Meloan (1985). These authors have underscored the resistance of the methylene crosslinks to the treatment with a high urea concentration, as demonstrated by Fraenkel-Conrat et al. (1947). This point is mentioned here because a buffer containing 5% urea has been proposed as an AR solution in IHC (Shi et al. 1996).
In tissue samples heavily fixed with formaldehyde and stained with hematoxylin-eosin, it is not easy to retain the eosin staining of most histological structures presumed to be bound to the basic components in the tissue (e.g., amino groups of proteins) after ethanol dehydration. Sections treated by heat, as in the AR technique, stain more strongly with eosin and this staining is not easily lost by dehydration. These are only casual observations but are very suggestive of the increased number of protein side groups after heat retrieval procedures.
Of course, the Ag-Ab binding is stronger than the binding between a stain and tissue structures. In the Ag-Ab binding, as has been repeatedly said, there is not only the binding of complementary side groups of the two proteins but also a conformational fitting of two small areas, one each in the antigen and the antibody, the epitope and the paratope.
It is likely that a blockade of the pertinent side groups with existing histochemical methods may help in the confirmation of the above hypothesis. A detailed explanation of all these possibilities is not possible here but it can easily be foreseen.
An attempt to evaluate these binding forces has been attempted by Montero et al. (2000), but much more can be done along this line of thought.
Antigen Retrieval
We cannot help but agree with Cattoretti et al. (1993) when they say that “the antigen is presumed to be masked in the section but not destroyed by fixation.” However, the word “masked” is not, in our opinion, the right one. We can say that the reaction of side groups of proteins with formaldehyde with the formation of methylene bridges and other bindings among the macromolecules in the tissue, have altered the protein macromolecule in such a way that its reaction with the antibody is now impossible or, in the best case, difficult.
Shi and colleagues (1991) are the pioneers in the introduction and development of heat-induced AR, using microwave or oven heating of formaldehyde-fixed, paraffin-embedded tissue sections. In subsequent studies it was found that there is a correlation between temperature of heating and heating time: the lower the temperature, the longer the time needed to obtain a given intensity. Different heating methods were used, but the temperature achieved appeared to be the critical variable. After the pioneer work of Shi et al. (1991), a number of studies have appeared in the literature that support their conclusions (Werner et al. 1996; Frost et al. 2000; Miller et al. 2000; Taylor 2000) as well as a modification of the procedure (Bankfalvi et al. 1994; Biddolph and Jones 1999).
Studying the influence of pH on the antigen-retrieval IHC (Shi et al. 1997), these authors found three different patterns of staining: (a) the stable type; (b) the V-form type; and (c) the ascending type. These can be interpreted as follows: (a) those antigens that result in an acceptable reaction through the whole scale of pHs; (b) those antigens with best results at extremes of pH, both alkaline and acid; and (c) antigens with best results on the alkaline side of the pH. They finally recommended using the AR solution of high pH. In their 1997 report, Shi et al. stated that “three types of antigens were demonstrated based on the pH value of the AR solution…”
Considering that there is a time interval between treatment with the AR solution at a predetermined pH and the incubation with the primary antibody, used at a pH not stated but presumed to be the pH usually used in routine histochemical procedures, the results of IHC appear to depend chiefly on the modifications introduced, not only of heat but also of the pH, in the protein macromolecules of the tissue. Their findings have been summarized recently (Shi et al. 2000).
The pH-dependent modifications of the macromolecules must be different in the various kinds of proteins, something that appears obvious when we consider that those proteins do have different degrees of polarization (i.e., either acidic or basic). This is a point that seems to have received little attention although it is outstanding in the reaction with the antibody.
In dealing with these problems it is always assumed that we are using monoclonal antibodies. Obviously, the reaction of polyclonal antibodies, with so many subsets of antibodies against different specific epitopes in the same solution, is more difficult to block. The mechanism of blockade cannot affect different epitopes in a similar way. On the other hand, monoclonal antibodies with only one possibility of reaction with an epitope are prone to give clear-cut results.
The comment of Boenisch (2001) that “any modification of the diluent buffer can affect the performance of both, the tissue antigen and the diluted antibody” but “that any improvement in quality of staining is consequence of antigen retrieval through a structural modification of the antigen” must be stressed here. The first proposition is true because any modification of the pH or the cations in the incubation medium will affect the Ag-Ab reaction: First, the change in pH because it implies a change in the polarization of both the protein components of the tissue and the antibody proteins (see Bungenberg de Jong 1949), and second, because the cations may affect the performance of monoclonal antibodies. Sodium chloride has been used in immunoabsorption and in chromatography and has been shown to suppress immunoreactivity at alkaline pH (Boenisch 1999).
Heat Retrieval and the Side Groups of Proteins
The statement by Bell et al. (1987) on the formaldehyde sensitivity of GFAP epitope, that “the effect of proteases, improving binding of the antibody to its epitope in fixed tissues, is consistent with the hypothesis that aldehyde fixation reduces the accessibility of antibodies to their epitopes, rather than altering epitope structure,” backs up the general view that alteration of the chemical constitution of the epitope by fixation is not an insurmountable handicap. The demonstration by these authors that the blockade or induced steric change of the structure of the epitope could be due to the crosslinking to GFAP of another protein adds a possibility to bear in mind when a negative result is obtained with a monoclonal antibody.
These and other explanations for negative results occasionally obtained when the IHC procedure is used for the demonstration of tissue proteins further demonstrate the complex problem of elucidating the mechanism of the Ag-Ab reaction in formaldehyde-fixed, paraffin-embedded tissues. Moreover, they tell us that the availability of chemical groups after AR that are responsible for the increased reactivity with the antibody cannot be easily explained in simple and direct terms. The only alternative that appears possible is a detailed study of each one of the many possibilities that may exist in the extensive variety of protein macromolecules existing in organic tissues.
An unpretentious possibility is given, in a report by Montero et al. (1991), which could be modified and extended to alternative antigens. Existing histochemical procedures for the demonstration of protein side or end groups (carboxyl and amino groups; tyrosine, tryptophan or histidine, as well as cystine or cysteine) might be applied to tissue sections before and after the AR technique, together with blockade of the same groups (Pearse 1968).
The use of different monoclonal antibodies for the same macromolecule can also be recommended because not all the epitopes in the same molecules have the same amino acid constitution. At the same time, the use of the corresponding positive and negative controls and of the blocking procedures suitable in each case is advised.
The judicious use of the above-mentioned histochemical procedures for the demonstration of protein side groups, together with the antigen retrieval technique, might be helpful in the interpretation of the results of these procedures.