Fragmentation of Monoclonal Antibodies

In certain circumstances, it may be desirable to generate antigen-binding fragments of monoclonal antibodies. For example, when cells with receptors for the Fc portion of IgG are present (e.g. macrophages, monocytes), intact IgG molecules may bind nonspecifically via their Fc portions (see Section 5.3.4). Removal of the Fc will prevent this mode of binding. The production of small fragments of IgM (intact Mr 900000) may aid penetration into the tissues for cytochemical studies.

The older immunological literature is replete with accounts of the fragmentation of mouse antibodies, often using conditions that were established for other species such as rabbit or human. In many cases, it was assumed that the procedures worked and no experimental verification was offered. In a few, it is now possible to state with hindsight that they could not possibly have worked. In others, the fragments were identified incorrectly. More recently, there has been an increasing awareness of the problem and the current literature is more solidly based.

It is important to examine the claims in the literature on fragmentation of mouse immunoglobulins with a critical eye. One should not be too surprised if one is unable to reproduce them. In the succeeding paragraphs, I will summarize what I consider to be the most reliable procedures and point out areas where uncertainty still exists.

9.5.1 Preparation of Fab Fragments of Mouse and Rat IgG

Cleavage of IgG at the hinge by the nonspecific protease papain is one other more straightforward way of producing defined fragments of mouse and rat IgG. It is easy to find conditions for complete cleavage. However, it should be pointed out that even in this case, evaluation of the literature is not as straightforward as it would appear.

Part of the problem arises from the fact that papain is a thiol protease (i.e. it has an SH group in the active site, which must be in the reduced form for activity). The presence of reducing agents necessary for the activation of papain (cysteine, mercaptoethanol or dithiothreitol, in order of increasing potency) may also facilitate digestion by their effects on the substrate. Cysteine was a popular reducing agent in the older literature. However, it is very easily oxidized to cystine by air, and results of such experiments may be poorly reproducible unless oxygen is excluded during digestion. All these reducing agents have the potential of cleaving the labile interchain disulfide bonds of immunoglobulins, but the extent to which these effects occur depends on the individual reducing agent and its concentration. In some published work, it is difficult to decide whether apparent heterogeneity in digestion is due to proteolytic cleavage, partial reduction, or a mixture of the two.

Papain is inactivated by heavy metals, which complex with its sulfhydryl group. It is often supplied as 'mercuripapain' to prevent autodigestion. Regardless of whether mercury is present or not, it is good practice to chelate any divalent cations with EDTA to ensure maximal activity of the enzyme. Papain should be 'activated' by a brief incubation in 0.1 M Tris-HCl, pH 8.0, containing 2 mM EDTA and 1 mM dithiothreitol, just prior to use.

Digestion may be terminated by the irreversible alkylation of the thiol group of papain with iodoacetamide. A twofold molar excess of iodo-acetamide over all thiol groups present is recommended. Reduction of iodoacetamide by thiol groups causes the release of HI, so the reaction should be well buffered.

Iodoacetamide absorbs light significantly at 280 nm. In addition, iodoacetamide is light-sensitive, and must be stored and used in the dark. The presence of any brown colour is an indication of decomposition, and such preparations should not be used.

Practical procedure

The extent of digestion of mouse and rat IgG is highly dependent on the concentration of reducing agent. The procedure given below (Oi and Herzenberg, 1979) should be monitored by SDS-PAGE (reducing conditions). Digestion should result in complete disappearance of the y heavy chain (Mr 55000) and appearance of the Fd fragment of y (Mr 27000) and light chains (Mr 22000-25000). An example of a typical digestion is shown in Fig. 9.4.

(1) Activate papain (1-2 mg/ml) in 0.1 M Tris-HCl, pH 8.0, containing 2 mM EDTA and 1 mM dithiothreitol, for 15 min at 37°C.

(2) Digest IgG (1-10 mg/ml) in the same buffer at 37°C, using an enzyme:substrate ratio of 1:100, for 1 h.

Antibody Fragmentation Ficin

Fig. 9.4. SDS-PACE of mouse IgGI and its fragments after digestion with papain. Digestion was terminated with iodoacetamide. The concentration of acrylamide was 10%, and all samples were reduced. A, Molecular weight standards (top to bottom): 95 000; 68 000; 43 000; 30000. B, Undigested IgC. C, IgG after digestion with papain. D: Fab fragment after purification. The lowest band is k chain. The upper bands are Fd fragments, although it is not clear why two bands are present. The mobility of papain (M, 20000) is indicated (P) although there is not enough enzyme present to be visible. I thank Annette Blane for providing this gel.

Fig. 9.4. SDS-PACE of mouse IgGI and its fragments after digestion with papain. Digestion was terminated with iodoacetamide. The concentration of acrylamide was 10%, and all samples were reduced. A, Molecular weight standards (top to bottom): 95 000; 68 000; 43 000; 30000. B, Undigested IgC. C, IgG after digestion with papain. D: Fab fragment after purification. The lowest band is k chain. The upper bands are Fd fragments, although it is not clear why two bands are present. The mobility of papain (M, 20000) is indicated (P) although there is not enough enzyme present to be visible. I thank Annette Blane for providing this gel.

(3) Terminate digestion by addition of iodoacetamide (final concentration 20 mM), and hold on ice for 1 h, protected from light.

(4) Dialyse overnight against phosphate-buffered saline, to remove iodoacetamide.

Comments It is absolutely essential to monitor the completeness of digestion by SDS-PAGE, as the individual IgG subclasses vary in their susceptibility to digestion. The rate of digestion is also influenced by the concentration of reducing agent. Mouse IgG2a and IgG2b are extremely susceptible to cleavage, even in 0.1 mM dithiothreitol, while IgGI is more resistant, and requires at least 1 mM dithiothreitol, and sometimes more. At 1 mM, there will be considerable reduction of interchain disulfide bonds, but antibody activity will probably remain intact. If cleavage is incomplete, the problem will usually be solved by increasing the concentration of reducing agent.

All rat IgG subclasses are easily cleaved by papain in the presence of reducing agent (Rousseaux et al., 1980, 1983; Rousseaux and Rousseaux-Prevost,

1990). It was found that Fab fragments could be produced from all rat IgG subclasses by treatment with 1% (w/w) papain at pH 7.0 in the presence of 10 mM cysteine for 2-4 h at 37°C. Completeness of digestion must be ascertained by SDS-PAGE in every case. Fragments could be separated by gel filtration, ion exchange chromatography or protein A-Sepharose (IgG2c only).

Separation of Fab and Fc

Separation of Fab fragments from Fc and residual intact IgG may usually be accomplished with ease. In the case of IgG subclasses with strong affinity for staphylococcal protein A (mouse IgG2a, IgG2b and IgG3; rat IgG2c), the digest is passed over a column of protein A-Sepharose (Goding, 1976, 1978). (Failure to inactivate the papain with iodoacetamide may result in destruction of the column.) The Fab fragments pass through the column, while Fc and undigested IgG are bound (Fig. 9.5, upper panel). Note that the Fc fragment of mouse IgG3 is very insoluble in water and may precipitate. The column may be regenerated and the bound material recovered by elution with 0.1 M glycine-hydrochloride, pH 2.5.

OD280

Fab lA

Volume

Fig. 9.5. Two methods for separation of Fab fragments of IgG from Fc fragments. Upper panel: separation on protein A-Sepharose. The IgG digest is passed over protein A-Sepharose; the Fab fragment drops through the column. When the pen recorder reaches baseline, the elution buffer (arrow; see text) is applied, and the Fc fragment and any undigested IgG eluted. Lower panel: separation by ion exchange chromatography. The digest is passed over a column of DEAE-cellu-lose. The Fab fragment drops through the column or elutes early in the gradient, followed by undigested IgG, and finally Fc.

Fig. 9.5. Two methods for separation of Fab fragments of IgG from Fc fragments. Upper panel: separation on protein A-Sepharose. The IgG digest is passed over protein A-Sepharose; the Fab fragment drops through the column. When the pen recorder reaches baseline, the elution buffer (arrow; see text) is applied, and the Fc fragment and any undigested IgG eluted. Lower panel: separation by ion exchange chromatography. The digest is passed over a column of DEAE-cellu-lose. The Fab fragment drops through the column or elutes early in the gradient, followed by undigested IgG, and finally Fc.

Alternatively, the fragments may be separated by ion exchange chromatography (see Section 9.2.3), which is the method of choice for IgG subclasses that do not bind to protein A. The digest should be transferred to 0.01M Tris-HC1, pH 8.0, by gel filtration on a small column or by dialysis or dilution. (The Fc fragment of IgG3 will precipitate in this buffer, and should be removed by centrifugation.) The Fab fragments are generally more basic than the Fc or the intact molecule; they will emerge in the 'drop-through' or early in the salt gradient (Fig. 9.5, lower panel). The intact molecules and Fc fragments emerge at higher salt concentrations. (Occasionally, the Fab and Fc fragments may have similar charge, and may coelute from DEAE-cellulose. This is mainly a problem for classes other than IgGl, and is an indication for use of protein A rather than DEAE to separate them.)

A potential disadvantage of Fab fragments of monoclonal antibodies

If both antigen-combining sites of an intact IgG molecule have simultaneous and equal access to identical determinants on a multimeric antigen, the functional affinity (avidity) may approach the product of the individual affinities, because if one site temporarily unbinds, the antibody remains bound via the other site (see Fig. 2.4). The strength of binding of monovalent Fab fragments might therefore be weaker than that of the intact molecule. Thus, if a monoclonal antibody were to have a low affinity for its antigen (as if often the case) the problem would be aggravated by the use of Fab fragments. Normally stable antigen-antibody complexes might dissociate during washing or prolonged incubations.

The effect of valency on the strength of binding is vividly shown in a series of elegant experiments (Ways and Parham, 1983; Parham 1984). A monoclonal antibody to HLA antigens was shown to bind bivalently to cells. It was found that the binding constant for intact or F(ab')2 fragments of this antibody was at least 68-fold greater than for monovalent fragments (Ways and Parham, 1983), and the true difference may be much greater. Bivalency was found to be essential for binding of two other monoclonal antibodies to HLA, as binding was undetectable if monovalent Fab fragments were used (Parham, 1984).

9.5.2 Preparation of F(ab')2 Fragments of Mouse and Rat IgG

Pepsin is a nonspecific protease which is only active at acid pH, and is irreversibly denatured at neutral or alkaline pH as it moves from the stomach into the duodenum. It is therefore essential to make up the pepsin solution from powder in an acidic buffer. If the pepsin is made up in PBS at neutral pH, it will be denatured and the digestion will not work.

Treatment of human or rabbit immunoglobulin with pepsin (enzyme:sub-strate 1:100 at 37°C overnight at pH 4.5 in acetate buffer) usually results in cleavage at the C-terminal side of the inter-heavy chain disulfide bonds (Figs 2.2 and 5.1). The resulting large fragment is named F(ab')2, because it contains two antigen-combining sites (Fig. 2.4). F(ab')2 fragments do not bind to Fc receptors. The production of F(ab')2 fragments of monoclonal antibodies would be very useful, because they would be expected to have considerably higher avidity than Fab (Fig. 2.4).

Unfortunately, the rote application of these procedures to mouse and rat IgG does not always produce satisfactory results. While there have been many claims for production of F(ab')2 fragments of mouse IgG, experience using rigorous techniques to assess the digestion products indicates that the recommended procedures do not always work (see Oi and Herzenberg, 1979).

Mouse IgC

The problem of production of F(ab')2 fragments of mouse IgG has been investigated by Lamoyi and Nisonoff (1983). Immunoglobulins were purified by conventional techniques, and digestion was carried out at pH 4.2 (IgGl and IgG2a) or pH 4.5 (IgG3). The enzyme:substrate ratio was 1:33 (w/w) and the temperature was 37°C. No reducing agents were used. F(ab')2 fragments were obtained in good yield from IgGl, IgG2a and IgG3 proteins, with the relative rates of digestion being IgG3 > IgG2a > IgGl. One IgGl protein out of four tested was rapidly degraded, as were all of five different IgG2b proteins.

Optimal digestion times varied depending on the antibody class, and times which resulted in no undigested IgG generally resulted in somewhat lower yields, presumably owing to further digestion into smaller fragments. Suitable times were found to be 8 h for IgGl (yield c.70%), 4-8 h for IgG2a (yield 25-50%) and 15 min for IgG3 (yield c. 60%). It was not possible to find suitable conditions for production of F(ab')2 fragments from mouse IgG2b. In general, it appears that the best yields are obtained by accepting a certain amount of undigested IgG, which may be removed by gel filtration (Table 9.1).

The production of F(ab')2 fragments of mouse IgG has also been analysed in detail by Parham (1983). It was found that IgGl was rather resistant to cleavage with pepsin, but stable F(ab')2 fragments could be produced by cleavage in 0.1 M citrate, pH 3.5 at 37°C, with an IgG concentration of 1-2 mg/ml and a pepsin concentration of 25 pg/ml. Cleavage was generally complete by 8 h, and yields varied from 25-90% for seven different proteins. The Fc portion was completely degraded. An optimal pH of 3.5 for pepsin digestion of mouse IgGl was also found by Morimoto and Inouye (1992).

Other procedures for production of F(ab')2 fragments from mouse IgGl have been published by Kurkela et al. (1988) who used papain digestion in the absence of thiols, and found that no Fab fragments were obtained, but F(ab')2 fragments could be recovered in 50% yield. Mariani et al. (1991) found that digestion of mouse IgGl with the thiol protease ficin in 50 raM Tris-HCl, 2 mM EDTA, pH 7.0 containing 1 mM cysteine at 37°C at an enzyme:substrate ratio of 1:30 for 2-8 h gave high yields of F(ab')2 fragments. The reaction was stopped by the addition of 1:10 v/v of 100 mM N-ethylmaleimide. The procedure was less effective for mouse IgG2a and IgG2b, as these classes tended to progress rapidly to Fab fragments under the given conditions.

In the case of mouse IgG2a, Parham found that a pH of 4.1 was optimal for pepsin digestion for the three proteins examined, but that it was not possible to find conditions that were general and ideal. Pepsin was capable of cleaving on both sides of the three inter-heavy-chain disulfide bonds. At short times, F(ab')2 predominated, but was contaminated by intact IgG. At longer times, a monomeric Fab' fragment became the major product, and a compromise between yield and purity was unavoidable.

Digestion of mouse IgG2b with pepsin was unsatisfactory and resulted in asymmetrical cleavage. IgG3 was not examined.

Rea and Ultee (1993) made the novel observation that 0.5-0.8 M ammonium sulfate could be used to control the digestion of mouse monoclonal antibodies by pepsin. They found that pepsin sometimes bound to the antibodies and precipitated them, possibly owing to an electrostatic interaction that could be overcome with ammonium sulfate, which accelerated the digestion in some cases and slowed it in others. In some cases, ammonium sulfate could be used successfully to control the digestion of mouse IgG2a by pepsin, but not for IgG2b. The effect of ammonium sulfate appeared to depend to a great extent on the individual antibody.

Yamaguchi et al. (1995) have recently shown that mouse IgG2a and IgG2b can be cleaved into F(ab')2 and Fc using lysyl endopeptidase (see also Kim et al., 1994a, b). The reaction conditions were 50 mM Tris-HCl pH 8.5 at 37°C for 1-6 h, using an enzyme:substrate ratio of 1:50 to 1:1000, and the reaction was terminated by addition of TLCK (tosyl-L-phenylalanine-chloromethyl ketone) to 30 mM final concentration. The reaction products could be separated by affinity chromatography on protein A-Sepharose. This is the first report of reliable production of F(ab')2 fragments of mouse IgG2b, and it will be interesting to see if the results are generally applicable to all monoclonal antibodies of this class.

As there are often very large differences in amino acid sequences between mouse IgG subclasses of different allotypes (see Section 6.1), it cannot be assumed that the results obtained with BALB/c immunoglobulins can be applied to IgG of other allotypes.

In summary, the production of F(ab')2 fragments from mouse IgGl is possible using pepsin or ficin, under carefully controlled conditions. Production of F(ab')2 fragments of mouse IgG3 is also possible using pepsin.

The production of F(ab')2 fragments from mouse IgG2a is sometimes possible, and cleavage into F(ab')2 and Fc by lysyl endopeptidase is very promising. Production of F(ab')2 fragments of mouse IgG2b has been very problematic, but recent work suggests that lysyl endopeptidase may provide a solution, but it will be necessary to determine whether this protocol works for all IgG2b monoclonal antibodies.

Rat IgC

Rousseaux and colleagues have studied the production of F(ab')2 fragments of rat IgG in detail (Rousseaux et al., 1983; Rousseaux and Rousseaux-Prévost, 1990). Their results may be summarized as follows. Rat IgGl and to a lesser extent IgG2a were resistant to pepsin, but complete cleavage to F(ab)2 fragments occurred if the IgG (15 mg/ml) was dialysed against 0.1 M formate buffer, pH 2.8 for 16 h at 4°C, followed by 0.1 M acetate bufTer, pH 4.5 before peptic digestion. Cleavage into F(ab')2 was complete after 4 h at 37°C with 1% (w/w) pepsin, and no other fragments were seen. It would appear that the pH 2.8 treatment caused an irreversible change in the conformation of the Fc, probably the Cy2 domain.

Cleavage of rat IgG2b with pepsin was less satisfactory, and gave results reminiscent of mouse IgG2a (see Parham, 1983, and above). Optimal conditions for pepsin digestion of rat IgG2b consisted of an enzyme:substrate ratio of 5% and an incubation time of 18 h at 37°C at pH 4.5, without the preliminary treatment at pH 2.8.

A more satisfactory way to generate F(ab')2-like fragments of rat IgG2b in good yield was found using staphylococcal V8 protease (3% w/w) in 0.1 M sodium phosphate pH 7.8, for 4 h at 37°C. Digestion was stopped by rapid freezing.

Rousseaux et al. (1983) found that IgG2c was the most susceptible of the rat IgG subclasses to peptic cleavage. Using an enzyme:substrate ratio of 1% (w/w) for 4 h at 37°C in 0.1 M citrate pH 4.5, rat IgG2c was completely cleaved into F(ab')2> and the Fc subfragment pFc' was also obtained. Longer times resulted in the appearance of monomeric Fab' fragments.

The foregoing should serve to emphasize the importance of testing individual IgG proteins for peptic cleavage, and warn against the rote use of 'recipes'. Fragmentation must always be monitored by SDS-PAGE, and failure to do so invites trouble.

Rousseaux and colleagues have recently published a procedure for the fragmentation of rat IgE into F(ab')2-e and Ce4 fragments (Rousseaux and Rousseaux-Prévost, 1990). IgE (10 mg/ml in PBS) was incubated for 10 min at 37°C with 10 mM cysteine, and then digested with papain (1% w/w) at 37°C in PBS pH 7.4 for 10-30 min, and the reaction stopped with iodoacetamide. The fragments were purified by gel filtration.

+1 0

Post a comment