X-ray crystallography has provided structural data on complete IgG molecules (Fig. 3.5). Mobility around the hinge region of IgG allows for the generation of the Y- and T-shaped structures visualized by electron microscopy.

Fig. 3.5 Model of an IgG molecule

A model of an IgG molecule showing the polypeptide backbones of the four chains as a ribbon. Heavy chains are shown in dark blue and dark green. The antigen-binding sites are at the tips of the arms of the Y-shaped molecule and are formed by domains from both the heavy and light chains. The extended, unfolded hinge region lies at the center of the molecule. Carbohydrate units are shown as ball and stick structures, covalently linked to the Fc region.

For all antibody isotypes there is pairing between VH/VL and CH1/CL domains through extensive non-covalent interactions to form the antigen binding (Fab) region.

Antigen binding is a common feature for IgG-Fab regions of each of the four human IgG subclasses; however, although there is >95% sequence homology between the IgG-Fc regions each IgG subclass exhibits a unique profile of effector activities.

The hinge regions are structurally distinct and determine the relative mobilities of the IgG-Fab and IgG-Fc moieties within the intact molecule. The equivalent of the IgG hinge region is present in all isotypes, except IgM.

In addition to the pairing of the VL/VH and CL/CH1 domains the CH3 domains of the IgG-Fc are also paired through non-covalent interactions.

The CH2 domains are not paired and, potentially, present a hydrophobic surface to solvent. This unfavorable property is avoided by interactions with a hydrophilic N-linked oligosaccharide moiety.

The N-linked oligosaccharide of the CH2 domain, although accounting for only 2–3% of the mass of the IgG molecule, is crucial to the expression of effector functions. The conformation of the CH2 domain protein moiety, and ultimately the IgG-Fc, results from reciprocal interactions between the CH2 protein and the oligosaccharide. The oligosaccharide exhibits structural heterogeneity and effector functions may be modulated depending on the particular oligosaccharide structure (glycoform) attached.

Assembled IgM molecules have a ‘star’ conformation

IgM is present in human serum as a pentamer of the basic four-chain structure (see Fig. 3.w3). Each heavy chain is comprised of a VH and four CH domains. One advantage of this pentameric structure is that it provides 10 identical binding sites, which can dramatically increase the avidity with which IgM binds its cognate antigen. Given that serum IgM commonly functions to eliminate bacteria containing low affinity, polysaccharide antigens, the increased avidity provided by the pentameric structure provides an important functional advantage.

Fig. 3.w3 Structural characteristics of various human immunoglobulins

Carbohydrate side chains are shown in blue. Inter heavy (H) chain disulfide bonds are shown in red, but interchain bonds between H and L chains are omitted. (1) A model of IgG1 indicating the globular domains of H and L chains. Note the apposition of the CH3 domains and the separation of the CH2 domains. The carbohydrate units lie between the CH2 domains. (2) Polypeptide chain structure of human IgG3. Note the elongated hinge region. (3) IgM H chains have five domains with disulfide bonds cross-linking adjacent CH3 and CH4 domains. The possible location of the J chain is shown. IgM does not have extended hinge regions, but flexion can occur about the CH2 domains. (4) The secretory component of sIgA is probably wound around the dimer and attached by two disulfide bonds to the CH2 domain of one IgA monomer. The J chain is required to join the two subunits. (5) This diagram of IgD shows the domain structure and a characteristically large number of oligosaccharide units. Note also the presence of a hinge region and short octapeptide tailpieces. (6) IgE can be cleaved by enzymes to give the fragments F(ab′)₂, Fc, and Fc′. Note the absence of a hinge region.

Covalent disulfide bonds between adjacent CH2 and CH3 domains, the C terminal 18-residue peptide sequence, referred to as the ‘tailpiece’, and J chain link the subunits of the pentamer.

J chain is synthesized within plasma cells, has a mass of ~15 kDa and folds to form an immunoglobulin domain. Each heavy chain bears four N-linked oligosaccharide moieties, however, the oligosaccharides are not integral to the protein structure in the same way as in IgG-Fc. Oligosaccharides present on IgM activate the complement cascade via binding to the mannose binding lectin (see Chapter 4).

In electron micrographs the assembled IgM molecule is seen to have a ‘star’ conformation with a densely packed central region and radiating arms (Fig. 3.6); however, electron micrographs of IgM antibodies binding to poliovirus show molecules adopting a ‘staple’ or ‘crab-like’ configuration (see Fig. 3.6), which suggests that flexion readily occurs between the CH2 and CH3 domains, though this region is not structurally homologous to the IgG hinge. Distortion of this region, referred to as dislocation, results in the ‘staple’ configuration of IgM required to activate complement.

Fig. 3.6 Electron micrographs of IgM molecules

(1) In free solution, deer IgM adopts the characteristic star-shaped configuration. ×195 000. (2) Rabbit IgM antibody (arrow) in ‘crab-like’ configuration with partly visible central ring structure bound to a poliovirus virion. × 190 000.

((1) Courtesy of Drs E Holm Nielson, P Storgaard, and Professor S-E Svehag. (2) Courtesy of Dr B Chesebro and Professor S-E Svehag.)

Secretory IgA is a complex of IgA, J chain and secretory component

IgA present in serum is produced by bone marrow plasma cells and secreted as a monomer with the basic four-chain structure. Each heavy chain is comprised of a VH and three CH domains.

The IgA1 and IgA2 subclasses differ substantially in the structure of their hinge regions:

• the hinge of IgA1 is extended and bears O-linked oligosaccharides;

• the hinge of IgA2 is truncated, relative to IgA1.

A deficit in the addition of O-linked sugars within the hinge region of IgA1 protein has been linked with the disease IgA nephropathy.

IgA is the predominant antibody isotype in external secretions but is present as a complex secretory form. IgA is secreted by gut localized plasma cells as a dimer in which the heavy chain ‘tailpiece’ is covalently bound to a J chain, through a disulphide bond (see Fig. 3.w3).

Electron micrographs of IgA dimers show double Y-shaped structures, suggesting that the monomeric subunits are linked end-to-end through the C terminal Cα3 regions (Fig. 3.7).

Fig. 3.7 Electron micrographs of human dimeric IgA molecules

The double Y-shaped appearance suggests that the monomeric subunits are linked end to end through the C terminal Cα3 domain. × 250 000.

(Courtesy of Professor S-E Svehag.)

The dimeric form of IgA binds a poly-Ig receptor (Fig. 3.8) expressed on the basolateral surface of epithelial cells. The complex is internalized, transported to the apical surface where the poly-Ig receptor is cleaved to yield the secretory component (SC) that is released still bound to the IgA dimer. The released secretory form of IgA is relatively resistant to cleavage by enzymes in the gut and is comprised of:

• two units of IgA;

• J chain; and

• a secretory component (mass 70 kDa) (see Figs. 3.w3 and 3.8).

Fig. 3.8 Transport of IgA across the mucosal epithelium

IgA dimers secreted into the intestinal lamina propria by plasma cells bind to poly-Ig receptors on the internal (basolateral) surface of the epithelial cells. The sIgA–receptor complex is then endocytosed and transported across the cell while still bound to the membrane of transport vesicles. These vesicles fuse with the plasma membrane at the luminal surface, releasing IgA dimers with bound secretory component derived from cleavage of the receptor. The dimeric IgA is protected from proteolytic enzymes in the lumen by the presence of this secretory component.

Serum IgD has antigen specificity but not effector functions

Serum IgD accounts for less than 1% of the total serum immunoglobulin. Although serum IgD has been shown to have specific antigen binding activity no effector functions have been identified. Each heavy chain is comprised of a VH domain and three CH domains with an extended hinge region (see Figs. 3.w3). IgD also functions as an antigen specific receptor on B cells, together with IgM, and as such exhibits the same diversity of antigen specificity.

The heavy chain of IgE is comprised of four constant region domains

It is estimated that ~50% of total body IgE is present in the blood, with the remainder being bound to mast cells and basophils through their high-affinity IgE Fcε receptor (FcεRI). IgE also binds the low affinity receptor (FcεRII) expressed on B cells and cells of the myeloid lineage.

Each heavy chain is comprised of a VH and four CH domains and bears six N-linked oligosaccharides (see Figs. 3.w3). An N-linked oligosaccharide, present in the CH3 domain, and equivalent to CH2 in IgG, influences binding to FcεRII but not FcεRI.

Antibody structural variation

Antibodies show structural variation of three different types – isotypic, allotypic, and idiotypic. The human immunoglobulin isotypes are products of defined immunoglobulin genes encoding the constant regions of heavy and light chains, and the allotypes are polymorphic variants of these genes. The idiotype of an antibody molecule results from antigenic uniqueness reflecting the structural uniqueness of the VH and VL regions (Fig. 3.w4).

Fig. 3.w4 Variability of immunoglobulin structure

All immunoglobulins have the basic four-chain structure. There are three types of immunoglobulin variability. (1) Isotypic variation is present in the germline of all members of a species, producing the heavy (μ, δ, γ, ε, α) and light chains (Igκ, λ) and the V region frameworks (subgroups). (2) Allotypic variation is intraspecies allelic variability. (3) Idiotypic variation refers to the diversity at the antigen-binding site (paratope) and in particular relates to the hypervariable segments.

Antibody isotypes are the products of genes present within the genome of all healthy members of a species

The human immunoglobulin heavy chain isotypes are encoded by μ, δ, γ3, γ1, α1, γ2, γ4, α2, and ε genes on chromosome 14. The single κ and multiple λ isotypes are encoded by Cκ and Cλ genes on chromosomes 2 and 22, respectively.

Allotypes result from genetic variation at a locus within the species

Immunoglobulin allotypes refer to proteins that are the product of polymorphic immunoglobulin genes (i.e. the result of genetic variation at a given locus within the species). The WHO-adopted nomenclature – for example G3m(g) – requires:

• definition of the heavy chain class, G;

• followed by the subclass, G3;

• an m (indicating ‘marker’), G3m; and

• the designated number or letter in brackets G3m(g).

The sequence correlates have been determined. For example G1m(f) and G1m(z) allotypes are characterized by the presence of, respectively, an arginine or lysine residue at position 214 in the CH1 or Cγ1 domain of the heavy chains.

In humans allotypes are encountered within:

• IgG;

• IgA; and

• κ light chains.

Because all heavy chain genes are encoded within the same locus, they are inherited as a complex or haplotype. Studies of immunoglobulin haplotypes help in the identification of different population groups of humans.

Allotypes of human IgG were first defined serologically, with antisera obtained from multi-parous women and individuals receiving multiple transfusions. In each case an allotype mis-match resulted in anti-allotype responses.

Idiotypes result from antigenic uniqueness

The structural uniqueness of antibody variable regions can be reflected in antigenic uniqueness recognized by antisera. However, in addition to antigenic uniqueness, cross-reactivity may be observed for two V regions that are highly homologous. The terms private and public (or cross-reactive and recurrent) idiotypes are used to describe this property. Idiotypes:

• may be specific for individual B cells or plasma cell products (antibodies) (private idiotypes);

• are sometimes shared between different B cell clones and or monoclonal antibodies (public, cross-reacting, or recurrent idiotypes).

Idiotypy (from the Greek ‘idios’, meaning ‘private’) originally referred to the antigenic uniqueness of an individual antibody molecule as recognized by antisera raised in rabbits and mice, by immunization of an individual with a single antibody molecule raised in another member of the same species and allotype. For human IgG proteins heterologous antisera are raised, in rabbits or mice, and absorbed with polyclonal IgG, to absorb cross-reactive antibody. In the modern era monoclonal anti-idiotype antibodies would be generated. They are important reagents when generating assays for quantitative and qualitative analysis of an antibody therapeutic (see Method box 3.1).

Method box 3.1 Recombinant antibodies for human therapy

Recombinant antibody therapeutics (rMAbs) are likely to become the largest family of disease-modifying drugs available to clinicians. Their efficacy results from specificity for the target antigen and biological activities (effector functions) activated by the immune complexes formed. Currently, 26 antibody products are licensed and hundreds are in clinical trials or under development. Initial trials administered mouse monoclonal antibodies specific for human targets and provided ‘proof of principle’.

Q. What problems could you envisage with the use of mouse antibodies to treat diseases in humans?

A. The antibodies might not interact appropriately with human effector molecules, e.g. Fc receptors, complement, etc. In the longer term an individual might mount an immune response against the non-self mouse antibodies.

In practice, patients mounted immune responses against the non-self mouse antibodies and the development of these human anti-mouse antibody responses (HAMA) meant that repeated dosing was not possible.

In response, scientists then produced:

• chimeric antibodies, in which the mouse V domains were linked to human antibody C domains;

• humanized antibodies, for which the CDRs of the mouse V region were transferred to a human V region framework sequence;

• fully human antibodies, from phage display libraries of expressed human antibody genes or transgenic mice expressing human antibody genes.

Although such therapeutic antibodies are termed ‘human’, they are derived in a manner that generates specificities that would ordinarily be eliminated by an individual’s immune system because they are directed against self. These developments have resulted in reduced immunogenicity but have not eliminated anti-drug antibody (ADA) responses; referred to as human anti-human antibody (HAHA) responses. In practice a variable proportion of patients produce ADA; parameters that influence their production includes, the type of antibody therapeutic (chimeric, humanized, fully human), the disease treated, the dosing regimen, the genetic background of the individual, etc.

The success of antibodies such as infliximab (anti-TNFα) and rituximab (anti-CD20) has resulted in demands for their production in metric tonnes. The biopharmaceutical industry has met the challenge with the construction of mammalian cell culture facilities (10 000–20 000-liter capacities) to produce rMAbs. Productivity has been greatly enhanced in recent years such that downstream processing is now the bottleneck. The cost of treatment with these drugs remains very high.

All full length antibody therapeutics currently licensed have been produced by mammalian cell culture using Chinese hamster ovary (CHO) cells or the mouse NSO or Sp2/0 plasma cell lines; a Fab therapeutic is produced in E. coli. Other systems under development and evaluation include transgenic animals, yeast, plants, etc.

The efficacy of an antibody therapeutic is critically dependent on appropriate post-translational modifications (PTMs) and each production system offers a different challenge because PTMs show species, tissue, and site specificity. Essential human PTMs are relevant to potency; non-human PTMs may increases the potential immunogenicity of the product.

At present we extrapolate from effector functions activated in vitro to presumed activity in vivo, for example complement-dependent cytotoxicity (CDC), antibody-dependent cellular cytotoxicity (ADCC), and the induction of apoptosis.

Further improvement of the potency of rMAbs depends on optimizing the biological effector functions activated in vivo by the immune complexes formed. This offers a considerable challenge due to the difficulty of monitoring events in vivo.

A number of genetically modified animal models are under development that might mimic human biology, e.g. mice transgenic for human Fc recptors.

In recent years in-vitro studies have unequivocally established that essential effector functions of IgG antibodies are dependent on appropriate glycosylation of the antibody molecule.

Glycosylation has therefore been a focus of attention for the biopharmaceutical industry for the past several years because regulatory authorities require that a consistent human-type glycosylation is achieved for rMAbs, irrespective of the system in which they are produced.

Chinese hamster ovary (CHO) cells and mouse myeloma cells (NS0, Sp2/0) can produce human IgG heavy chains bearing human type glycoforms but not the full range. It is now evident that effector function efficacy differs between glycoforms; thus non-fucosylated antibodies have enhanced ADCC, compared to the fucosylated form. These production cell lines are being engineered to produce pre-selected glycoforms, depending on the effector activities considered to be optimal for a given application. These cell lines can also generate non-human glycoforms that may be immunogenic; therefore, targeted glycosyltransferases have been ‘knocked-out’.

Fig. MB3.1.1 Production of Fv antibodies by phage display.

Artificial antibodies such as Fv fragments, consisting of a hybrid of antibody VH and VL regions, can be generated and selected by phage display technology. Antibody VH and VL genes are first amplified from B cell mRNA by the polymerase chain reaction. The genes are joined together with a spacer to give a gene for an Fv fragment. Bacteria are then transfected with the gene in a phagemid vector containing a leader sequence, a fragment of the gene expressing phage coat protein 3, and an M13 origin of replication, and then infected with M13 phage. The phages replicate and express the Fv on their tips. Phages displaying the right specificity are isolated by panning on antigen-coated plates, and amplified. The antigen-specific phage can be used to infect strains of bacteria that allow the secretion of the Fv protein into the culture medium. Following selection a Fv can be used to generate a full length antibody.

Antigen–antibody interactions

The conformations of the epitope and the paratope are complementary

Protein molecules are not rigid structures, but exist in a dynamic equilibrium between structures that differ in their ability to form a primary interaction with specific ligands.

Following a primary interaction, each ‘partner’ may influence the final conformation within the complex. This concept approximates to the ‘induced fit’ model of protein–protein interactions.

An examination of the interaction between the Fab fragment of the mouse D1.3 monoclonal antibody and hen egg white lysozyme (HEL) reveals the complementary surfaces of the epitope and the antibody’s combining site (paratope); comprised of 17 amino acid residues of the antibody and 16 residues of the lysozyme molecule (Fig. 3.9). All hypervariable regions of the heavy and light chains contribute, though the third hypervariable region in the heavy chain appears to be most important.

Fig. 3.9 The Fab–lysozyme complex

(1) Lysozyme (green) binds to the hypervariable regions of the heavy (blue) and light (yellow) chains of the Fab fragment of antibody D1.3. (2) The separated complex with Glu121 visible (red). This residue fits into the center of the cleft between the heavy and light chains. (3) The same molecules rotated 90° to show the contact residues that contribute to the antigen–antibody bond.

The paratope of the D1.3 monoclonal antibody may be regarded as ‘classical’. The structures of other lysozyme– antibody complexes have been solved and show differing involvement of hypervariable and framework residues.

Such structural studies are essential when engineering antibody molecules (e.g. when ‘humanizing’ a mouse antibody to generate an antibody therapeutic, see Method box 3.1).

Antibody affinity is a measure of the strength of interaction between a paratope and its epitope

The affinity of a protein–protein interaction is a thermodynamically defined measure of the strength of interaction between reciprocal binding sites, i.e as for the paratope of a Fab fragment and the epitope of an antigen. Since an antibody has two Fab moieties it can form multiple or three dimensional complexes; thus the apparent affinity is enhanced and is referred to as the avidity.

Antibodies form multiple non-covalent bonds with antigen

The antigen–antibody interaction results from the formation of multiple non-covalent bonds. These attractive forces consist of:

• hydrogen bonds;

• electrostatic bonds;

• van der Waals forces; and

• hydrophobic forces.

Each bond is relatively weak in comparison with covalent bonds, but together they can generate a high-affinity interaction.

The strength of a non-covalent bond is critically dependent on the distance (d) between the interacting groups, being proportional to 1/d² for electrostatic forces, and to 1/d⁷ for van der Waals forces.

Thus interacting groups must be in intimate contact before these attractive forces come into play.

For a paratope to combine with its epitope (see Fig. 3.9) the interacting sites must be complementary in shape, charge distribution, and hydrophobicity, and in terms of donor and acceptor groups, capable of forming hydrogen bonds.

Close proximity of two protein surfaces can also generate repulsive forces (proportional to 1/d¹²) if electron clouds overlap.

In combination, the attractive and repulsive forces have a vital role in determining the specificity of the antibody molecule and its ability to discriminate between structurally similar molecules.

The great specificity of the antigen–antibody interaction is exploited in a number of widely used assays (see Method box 3.2).

Method box 3.2 Assays for antibodies and antigens

Immunoassays use labeled reagents for detecting either antibodies or antigens. They are very sensitive and economical in the use of reagents.

Solid-phase assays for antibodies using detection reagents labeled with radioisotopes (radioimmunoassay, RIA) or with enzymes (enzyme-linked immunosorbent assay, ELISA) are probably the most widely used of all immunological assays, because large numbers can be performed in a relatively short time.

Fig. MB3.2.1 Immunoassay for antibody.

(1) Antigen in saline is incubated on a plastic plate or tube, and small quantities become absorbed onto the plastic surface. (2) Free antigen is washed away. (The plate may then be blocked with excess of an irrelevant protein to prevent any subsequent non-specific binding of proteins.) (3) Test antibody is added, which binds to the antigen. (4) Unbound proteins are washed away. (5) The antibody is detected by a labeled ligand. The ligand may be a molecule such as staphylococcal protein A that binds to the Fc region of IgG – more often it is another antibody specific for the test antibody. By using a ligand that binds to particular classes or subclasses of test antibody it is possible to distinguish isotypes. (6) Unbound ligand is washed away. (7) The label bound to the plate is measured. A typical titration curve is shown in the graph. With increasing amounts of test antibody the signal rises from a background level through a linear range to a plateau. Antibody titers can only be detected correctly within the linear range. Typically the plateau binding is 20–100 times the background. The sensitivity of the technique is usually about 1–50 ng/mL of specific antibody. Specificity of the assay may be checked by adding increasing concentrations of free test antigen to the test antibody at step 3; this binds to the antibody and blocks it from binding to the antigen on the plate. Addition of increasing amounts of free antigen reduces the signal.

Fig. MB3.2.2 Enzyme-linked immunosorbent assay (ELISA).

The ELISA plate is prepared in the same way as the immunoassay up to step 4. In this system, the ligand is a molecule that can detect the antibody and is covalently coupled to an enzyme such as peroxidase. This binds the test antibody, and after free ligand is washed away (6) the bound ligand is visualized by the addition of chromogen (7) – a colorless substrate that is acted on by the enzyme portion of the ligand to produce a colored end-product. A developed plate (8) is shown in the lower panel. The amount of test antibody is measured by assessing the amount of colored end-product by optical density scanning of the plate.

Fig. MB3.2.3 Assay of antigen.

(1) Competitive assay. The test antigen is placed together with labeled antigen onto a plate coated with specific antibody. The more test antigen present, the less labeled standard antigen binds. This type of assay is often used to measure antigens at relatively high concentrations, or hormones that have only a single site available for combination with antibody. (2) Two-site capture assay. The assay plate is coated with specific antibody, the test solution then applied, and any antigen present captured by the bound antibody. After washing away unbound material, the captured antigen is detected using a labeled antibody against another epitope on the antigen. Because the antigen is detected by two different antibodies, the second in excess, such assays are both highly specific and sensitive.

Antigen–antibody interactions are reversible

The affinity of an antibody is the sum of the attractive and repulsive forces resulting from binding between the paratope of a monovalent Fab fragment and its epitope. This interaction will be reversible, so at equilibrium the Law of Mass Action can be applied and an equilibrium constant, K (the association constant), can be determined (Fig. 3.10). In practice, due to the divalency of antibody and multiple epitopes expressed on an antigen, large three dimensional complexes may be formed that do not readily dissociate; when a paratope–epitope interaction is broken both the antibody and antigen remain in proximity due to other paratope–epitope interactions, thus reassociation is favored.

Fig. 3.10 Calculation of antibody affinity

All antigen–antibody reactions are reversible. The Law of Mass Action can therefore be applied, and the antibody affinity (given by the equilibrium constant, K) can be calculated. (Square brackets refer to the concentrations of the reactants)

Avidity is likely to be more relevant than affinity

Because each antibody unit of four polypeptide chains has two antigen-binding sites, antibodies are potentially multivalent in their reaction with antigen.

In addition, antigen can be:

• monovalent (e.g. small chemical groups, haptens); or

• multivalent (e.g. microorganisms).

The strength with which a multivalent antibody binds a multivalent antigen is termed avidity to differentiate it from the affinity, which is determined for a univalent antibody fragment binding to a single antigenic determinant.

The avidity of an antibody for its antigen is dependent on the affinities of the individual antigen-combining sites for the epitopes on the antigen. Avidity will be greater than the sum of these affinities if both antibody-binding sites bind to the antigen because all antigen–antibody bonds would have to be broken simultaneously for the complex to dissociate (Fig. 3.11).

Fig. 3.11 Affinity and avidity

Multivalent binding between antibody and antigen (avidity or functional affinity) results in a considerable increase in stability as measured by the equilibrium constant, compared with simple monovalent binding (affinity or intrinsic affinity, here arbitrarily assigned a value of 10⁴ L/mol). This is sometimes referred to as the ‘bonus effect’ of multivalency. Thus there may be a 10³-fold increase in the binding energy of IgG when both valencies (combining sites) are used and a 10⁷-fold increase when IgM binds antigen in a multivalent manner.

In physiological situations, avidity is likely to be more relevant than affinity because antibodies are at least divalent and most naturally occurring antigens are multivalent.

In practice we determine the association constant at equilibrium when the rate of formation of complex (ka) is equal to the spontaneous rate of dissociation (kd). The association or equilibrium constant is defined as K = ka/kd.

It has been suggested that B cell selection and stimulation during a maturing antibody response depend upon selection for the ability of antibodies to bind to antigens both:

• rapidly (kinetic selection); and

• tightly (thermodynamic selection).

Cross-reactive antibodies recognize more than one antigen

Antigen–antibody reactions can show a high level of specificity, but can also be cross-reactive, binding to a structurally related but different antigen (Fig. 3.12). Thus, monoclonal antibodies to hen egg lysozyme (HEL) may also bind structurally homologous duck egg lysozyme (DEL). A polyclonal antiserum to HEL will contain populations of antibodies specific for HEL and others that cross-react with DEL.

Fig. 3.12 Specificity, cross-reactivity, and non-reactivity

Antiserum specificity results from the action of a population of individual antibody molecules (anti-X, anti-Y, anti-Z) directed against different epitopes (X, Y, Z) on the same or different antigen molecules. Antigen A (AgA) and antigen B (AgB) have epitope Y in common. Antiserum raised against AgA (anti-XYZ) not only reacts specifically with AgA, but cross-reacts with AgB (through recognition of epitopes Y and X′). The antiserum gives no reaction with AgC because there are no shared epitopes.

Antibodies recognize the conformation of antigenic determinants

Analysis of antibodies to protein antigens reveals that specificity may be for epitopes:

• consisting of a single contiguous stretch of amino acids (a continuous epitope);

• dependent on the native conformation of the antigen and formed from two or more stretches of sequence separated in the primary structure (discontinuous or conformational epitopes).

Continuous epitopes are unique three dimensional structures whilst discontinuous epitopes may be formed of a flexible peptide that assumes a unique conformation when bound to a paratope. i.e., the paratope may influence the conformation of the epitope by an induced fit mechanism.

Q. How might these differing characteristics of antigen be relevant when producing antibodies for immunological assays?

A. Antibodies specific for discontinuous epitopes may not bind denatured antigen, for example on immunoblots (see Method box 3.3), whereas antibodies to continuous epitopes may bind denatured antigen.

Method box 3.3 Assays to characterize antigens

The techniques of immunoprecipitation and immunoblotting (also known as western blotting) are both used to characterize antigens, giving information on their molecular mass, purity, heterogeneity, and limited information on relative abundance.

These techniques may be useful for comparing the abundance of antigens, but are less well suited to the quantitation of antigens than the immunoassays described in Method box 3.2.

Immunoblotting is a simple technique, but only continuous epitopes or continuous epitopes that can reform under the conditions of the experiment can be detected.

Immunoprecipitation is more often used where antibodies that recognize only undenatured antigens are available.

Fig. MB3.3.1 Immunoblotting.

In immunoblotting, denatured antigen samples are separated in an analytical gel (e.g. an SDS polyacrylamide gel or an isoelectric focusing gel). The resolved molecules are transferred electrophoretically, in a blotting tank, to a nitrocellulose membrane. The blot is then exposed to antibody, washed and further exposed to a radiolabeled conjugate (second antibody) to detect antibody bound to antigen. The principle is similar to that of a radioimmunoassay (RIA, see Method box 3.2) or enzyme-linked immunosorbent assay (ELISA, see Method box 3.2). After washing again, the blot is placed in contact with X-ray film in a cassette. The autoradiograph is developed and the antigen bands that have bound the antibody are visible. This immunoblotting technique can be modified for use with a chemiluminescent label or an enzyme-coupled conjugate (as in ELISA), where the bound material can be detected by treatment with a chromogen, which deposits an insoluble reagent directly onto the blot.

Fig. MB3.3.2 Immunoprecipitation.

In immunoprecipitation the antigens being tested are labeled with ¹²⁵I, and antibody is added that binds only to its specific antigen. The complexes are precipitated by the addition of co-precipitating agents, such as anti-immunoglobulin antibodies or staphylococcal protein A. The insoluble complexes are spun down and washed to remove any unbound labeled antigens. The precipitate is then resolubilized (e.g. in SDS), and the components are separated on analytical gels. After running, the fixed gels are autoradiographed, to show the position of the specific labeled antigen. Frequently the antigens are derived from the surface of radiolabeled cells, which are solubilized with detergents before the immunoprecipitation. It is also possible to label the antigens with biotin, and detect them at the end chromatographically using streptavidin (binds biotin) coupled to an enzyme such as peroxidase (see ELISA technique in Method box 3.2).

Antibodies are capable of expressing remarkable specificity and are able to distinguish small differences in the shape and chemical properties (e.g. charge, hydrophobicity) of epitopes. Small changes in the epitope, such as the position of a single chemical group, can therefore abolish binding (Figs. 3.13 and 3.14).

Fig. 3.13 Specificity and cross-reacitivity

Antiserum raised to the meta isomer of aminobenzene sulfonate (the immunizing hapten) is mixed with ortho and para isomers of aminobenzene sulfonate, and also with the three isomers (ortho, meta, para) of two different, but related, antigens: aminobenzene arsonate and aminobenzene carboxylate. The antiserum reacts specifically with the sulfonate group in the meta position, but will cross-react (although more weakly) with sulfonate in the ortho position. Further, weaker cross-reactions are possible when the antiserum is reacted with either the arsonate group or the carboxylate group in the meta, but not in the ortho or para, position. Arsonate is larger than sulfonate and has an extra hydrogen atom, while carboxylate is the smallest of the three groups. These data suggest that an antigen’s configuration is as important as the individual chemical groupings that it contains.

Fig. 3.14 Configurational specificity

The lysozyme molecule possesses an intrachain bond (red), which produces a loop in the peptide chain. Antisera raised against whole lysozyme (anti-lysozyme) and the isolated loop (anti-‘loop’ peptide) are able to distinguish between the two. Neither antiserum reacts with the isolated loop in its linear reduced form. This demonstrates the importance of tertiary structure in determining antibody specificity.

Q. Consider an antibody that recognizes an antigen on the outer envelope of a virus. If the antigen mutated, what would be the effect on its ability to bind antibody?

A. It depends on the nature of the mutation. Only mutations that affect the structure of the epitope could affect binding. However, if the mutation caused a radical change in the epitope structure (e.g. charge change or the insertion of a bulky amino acid residue), then the ability to bind might be lost or the affinity of binding drastically reduced.

Q. What effect would it have on the virus if a mutation of the epitope prevented it from being recognized by the antibody?

A. Viruses that mutate to avoid recognition by antibody will be at a selective advantage because they will not be destroyed by the immune system or prevented from attaching to host cells. Many viruses, including influenza and HIV, mutate their surface molecules and so evade recognition by antibodies.

Antibody effector functions

Antibodies are bifunctional because they both:

• specifically bind antigen to form large complexes thus limiting the spread of pathogens in vivo; and

• the immune complexes formed elicit host responses to facilitate their removal and destruction.

The nature of the constant region determines the effector function of the antibody and the host responses elicited, e.g. complement activation, phagocytosis.

In antibody–antigen complexes the antibody molecules are essentially aggregated such that the multiple Fc regions are able to engage, cross link and activate ligands or receptors (e.g. FcγR and C1q) (Fig. 3.15). Antibody is said to be an opsonin; it opsonizes the antigen (bacterium, virus); to opsonise means to make the antigen more ‘tasty’ or ‘attractive’ targets for phagocytic cells, i.e. they ingest or eat the complexes.

Fig. 3.15 Biological properties of human immunoglobulins

The biological functions of different antibody classes and subclasses depend on which receptors they bind to and the cellular distribution of those receptors mediated by acting as adapters that bind via their Fc region to Fc receptors on different cell types.

IgM predominates in the primary immune response

IgM is the first antibody produced in the primary immune response and is largely confined to the intravascular pool. It is frequently associated with the immune response to antigenically complex, blood-borne infectious organisms.

Once bound to its target, IgM is a potent activator of the classical pathway of complement.

Although the pentameric IgM antibody molecule consists of five Fcμ regions, it does not activate the classical pathway of complement in its uncomplexed form. However, when bound to an antigen with repeating identical epitopes, it forms a ‘staple’ structure (see Fig. 3.6), undergoing a conformational change referred to as dislocation, and the multiple Fcμ presented in this form are able to initiate the classical complement cascade.

IgG is the predominant isotype of secondary immune responses

The four IgG subclasses are highly homologous in structure, but each exhibits a unique profile of effector functions; thus, in activating the classical pathway of complement, complexes formed with:

• IgG1 and IgG3 antibodies are efficient;

• IgG2 antibodies are less effective;

• IgG4 antibodies are inactive.

The IgG subclasses also interact with a complex array of cellular Fc receptors (FcλR) expressed on various cell types (see Fig. 3.15 and below). IgG equilibrates between the intravascular and extravascular pools, so providing comprehensive systemic protection.

In humans, the newborn infant is not immunologically competent and the fetus is protected by passive IgG antibody selectively transported across the placenta (Fig. 3.16). Transport is mediated by the neonatal Fc receptor (FcRn) – all IgG subclasses are transported but the cord/maternal blood ratios differ, being approximately 1.2 for IgG1 and approximately 0.8 for IgG2.

Fig. 3.16 Immunoglobulins in the serum of the fetus and newborn child

IgG in the fetus and newborn infant is derived solely from the mother. This maternal IgG has disappeared by the age of 9 months, by which time the infant is synthesizing its own IgG. The neonate produces its own IgM and IgA; these classes cannot cross the placenta. By the age of 12 months, the infant produces 80% of its adult level of IgG, 75% of its adult IgM level, and 20% of its adult IgA level.

Q. How do you interpret these ratios with respect to IgG transport across the placenta?

A. IgG1 is transported more efficiently than IgG2. Also, because the ratio for IgG1 is >1, the antibody is moving up a concentration gradient; therefore it must be active transport.

In some species (e.g. the rat), maternal immunoglobulin, present in colostrum or milk, is transferred to the offspring in the postnatal period through selective transport of IgG across the gastrointestinal tract via a homologous FcRn receptor.

Serum IgA is produced during a secondary immune response

Serum IgA is a product of a secondary immune response. Immune complexes of IgA opsonize antigens and activate phagocytosis through cellular Fc receptors (FcαR).

A predominant role for IgA antibody is in its secretory form, affording protection of the respiratory, gastrointestinal, and reproductive tracts.

The IgA1 subclass predominates in:

• human serum (approximately 90% of total IgA); and

• secretions such as nasal mucus, tears, saliva, and milk (70–95% of total IgA).

In the colon, IgA2 predominates (approximately 60% of the total IgA). Many microorganisms that can infect the upper respiratory and gastrointestinal tracts have adapted to their environment by releasing proteases that cleave IgA1 within the extended hinge region, whereas the short hinge region of IgA2 is not vulnerable to these enzymes.

IgD is a transmembrane antigen receptor on B cells

IgD has no known effector functions as a serum protein. It functions as a transmembrane antigen receptor on mature B lymphocytes.

IgE may have evolved to protect against helminth parasites infecting the gut

Despite its low serum concentration, the IgE class:

• is characterized by its ability to bind avidly to circulating basophils and tissue mast cells through the high-affinity FcεRI receptor (see below);

• sensitizes cells on mucosal surfaces such as the conjunctival, nasal, and bronchial mucosae.

IgE may have evolved to provide immunity against helminth parasites, but in developed countries it is now more commonly associated with allergic diseases such as asthma and sensitivity to peanuts, eggs, fish, etc.

Fc receptors

Antibodies sometimes protect just by binding to a pathogen, so preventing the pathogen from attaching to cells of the body and infecting them. More often, the biological and protective functions of antibodies are mediated by acting as adapters that bind antigen, via the paratope, and via their Fc region to Fc receptors (FcR) expressed on the above cell types. Three classes of FcR are defined that recognize the Fc regions of IgG (FcγR), IgA (FcαR) and IgE (FcεR); within each class of FcR there are several different types, e.g. FcγRI, FcγRII and FcγRIII.

The three types of Fc receptor for IgG are FcγRI, FcγRII, and FcγRIII

Three types of cell surface receptor for IgG (FcγR) are defined in humans:

• FcγRI (CD64);

• FcγRII (CD32); and

• FcγRIII (CD16).

Each receptor is characterized by a glycoprotein α chain that has an extracellular domain, homologous with immunoglobulin domains, that binds to antibody (Fig. 3.17) – that is, they belong to the immunoglobulin superfamily, as do receptors for IgA (FcαR) and IgE (FcεRI).

Fig. 3.17 Fcγ receptors

Receptors for Fcγ in humans belong to the immunoglobulin superfamily, and have either two or three extracellular immunoglobulin domains. Motifs (ITAM, ITIM) on the intracellular segments or on associated polypeptides are targets for tyrosine kinases involved in initiating intracellular signaling pathways.

FcγRs are expressed constitutively but differentially on a variety of cell types but may be upregulated or induced by environmental factors (e.g. cytokines).

Biological activation results from aggregation (cross-linking) of the FcγR on the cell surface with consequent signal transduction and subsequent activation of immunoreceptor tyrosine-based activation (ITAM) or immunoreceptor tyrosine-based inhibitory (ITIM) motifs in the cytoplasmic sequences.

Phosphorylation of the ITAM motif triggers activities such as:

• phagocytosis;

• antibody-dependent cell-mediated cytotoxicity (ADCC);

• apoptosis;

• mediator release;

• enhancement of antigen presentation.

In contrast, phosphorylation of ITIM blocks cellular activation.

FcγRI is involved in phagocytosis of immune complexes and mediator release

It is common practice to refer to FcR as being of high, medium or low affinity. However, these descriptions are imprecise and relative; the affinity values vary considerably depending on the methods of determination.

FcγRI (CD64) binds:

• monomeric IgG1 and IgG3 with high affinity; relative to

• IgG4, low affinity;

• IgG2 does not have detectable binding affinity.

FcγRI has a restricted cellular distribution but is expressed on all cells of the mononuclear phagocyte lineage, and is involved in the phagocytosis of immune complexes, mediator release, etc.

The extra-cellular region of the α chain is comprised of three immunoglobulin domains; the cytoplasmic domain is associated with a γ chain that bears an ITAM motif.

FcγRII is expressed in two forms

FcγRII (CD32a; CD32b) is expressed as structurally and functionally distinct FcγRIIa and FcγRIIb forms with wide but differing cellular distribution.

The α chain of FcγRIIa:

• has moderate affinity for monomeric IgG1 and IgG3;

• binds complexed (multivalent, aggregated) IgG with high avidity;

• expresses an ITAM motif within its cytoplasmic tail; and

• may be produced as a polymorphic form which can bind monomeric IgG2.

Polymorphism in the FcγRIIA gene results in the presence of histidine or arginine at position 131 in the extracellular domains – the His¹³¹ allotype binds and is activated by immune complexes of IgG2.

The FcγRIIb molecule:

• expresses an ITIM motif within its cytoplasmic tail; and

• when cross-linked inhibits cellular activation, particularly on B cells (see Fig. 11.13).

FcγRIII is expressed as FcγRIIIa and FcγRIIIb

FcγRIIIa (CD16a) is structurally and functionally distinct from FcγRIII (CD16b) and they have different cellular distributions. Both are extensively glycosylated.

FcγRIIIa is a transmembrane protein (like FcγRI, FcγRIIa, and FcγRIIb), whereas FcγRIIIb is GPI (glycosyl phosphatidyl inositol) anchored (see Fig. 3.17).

The α chains of FcγRIIIa:

• have a moderate affinity for monomeric IgG; and

• may be associated with γ/ξ and/or β chains bearing ITAM motifs.

FcγRIIIa is expressed on monocytes, macrophages, NK cells, and a fraction of T cells.

The FcγRIIIb form:

• is selectively expressed on neutrophils and basophils; and

• has a low affinity for monomeric IgG.

Engagement and cross-linking of FcγRIIIb can result cellular activation due to lipid raft formation by association with other membrane proteins bearing signaling motifs.

Polymorphism in FcγRIIIa and FcγRIIIb may affect disease susceptibility

Polymorphism in the FcγRIIIA gene results in the presence of phenylalanine (Phe) or valine (Val) at position 158 in the extracellular domains; the Val¹⁵⁸ allotype is associated with higher binding affinity for IgG1/IgG3 and consequently more efficient NK cell activation.

Polymorphism in the FcγRIIIB gene results in multiple amino acid sequence differences, including the generation of an N-linked glycosylation motif in the FcγRIIIb-NA2 and consequently the extent of glycosylation.

The relative expression of the FcγRIIIb-NA1 and FcγRIIIb-NA2 forms are reported to be associated with differing susceptibility to infection.

Q. What factors determine whether a particular IgG subclass will have a particular biological function (e.g. the ability to opsonize a bacterium for phagocytosis by a macrophage)?

A., The ability of each subclass of antibody to bind to Fc receptors, the affinity of binding and the distribution of Fc receptors between different effector cells.

IgG Fc interaction sites for several ligands have been identified

Application of site-directed mutagenesis, X-ray crystallography and nuclear magnetic resonance spectroscopy has allowed elucidation of the molecular topography of IgG Fc interaction sites for several ligands that bind overlapping non-identical sites at the CH2/CH3 interface, for example:

• staphylococcal protein A;

• streptococcal protein G;

• FcRn; and

• FcγRIIIB.

The interactions between maternal IgG and the MHC class I molecule-like FcRn expressed on the intestinal epithelium of the neonatal rat have now been studied at high resolution (Fig. 3.18) and are believed to mimic closely the binding of the human placental counterpart, hFcRn, with maternal IgG. Titration of IgG histidine residues in the binding site for FcRn may explain its:

• pH sensitivity;

• ability to bind at pH 6.5 (the pH within vacuoles); and

• dissociation at pH 7.4 (the pH of blood).

Fig. 3.18 Neonatal rat intestinal Fc receptor FcRn

Principal interactions between neonatal rat intestinal FcRn and the Fc of maternal IgG (derived from milk) are illustrated by ribbon diagrams of FcRn (domains α1, α2, α3, and β₂m are shown in red, light green, purple, and gray, respectively) and of Fc (CH2 and CH3 domains are shown in blue and yellow). The main contact residues of the FcRn (α1 domain, 90; α2, 113–119 and 131–135; β₂m, 1–4 and 86) are depicted as space-filling structures.

(Reproduced from Ravetch JV, Margulies DH. New tricks for old molecules. Nature 1994;372:323–324.)

Given the symmetry of the Fc region, the fragments used in the experiments above are functionally divalent and may form multimeric complexes.

If monomeric IgG were divalent for FcγR and C1q, however, it would not function properly because circulating monomeric IgG could form activating multimers. The interaction sites for these ligands (e.g. FcγR and C1q) have been ‘mapped’ to the CH2 domain next to the hinge. The crystal structure of an IgG Fc/FcγRIIIb complex reveals an asymmetric interaction site embracing the CH2 domains of both heavy chains, thereby ensuring monovalency.

IgM–antigen complexes are very efficient activators of the classical complement system, but the mechanism by which IgM binds C1q appears different from that of IgG. The conformational change from a ‘star’ to a ‘staple’ conformation upon binding to multivalent antigen is thought to unveil a ring of occult C1q-binding sites that are not accessible in the star-shaped configuration (see Fig. 3.6).

Glycosylation is important for receptor binding to IgG

N-linked glycosylation of the IgG Cγ2 domain is essential for the binding and activation of FcγRI, FcγRII, FcγRIII, and C1q. The oligosaccharide is of the complex type and is heterogeneous, generating multiple glycoforms. In addition to the covalent link to asparagine 297 it forms multiple non-covalent interactions with the CH2 domain such that it appears to be sequestered within the protein structure and inaccessible for direct interactions with the FcγRI, FcγRII, FcγRIII, and C1q ligands.

Fidelity of glycosylation has become an important issue in the production of monoclonal antibodies for therapy because glycosylation is a species- and tissue-specific post-translational modification (see Method box 3.1).

The FcR for IgA is FcαRI

The receptor for IgA is FcαRI (CD89):

• it is comprised of two Ig superfamily domains;

• it is associated with a γ chain;

• it is expressed on myeloid cells;

• can trigger phagocytosis, cell lysis and the release of inflammatory mediators; and

• can bind both IgA1 and IgA2.

The two types of Fc receptor for IgE are FcεRI and FcεRII

Two types of Fc receptor for IgE (FcεR) are defined in humans (Fig. 3.19):

• the high-affinity FcεRI which is expressed on mast cells and basophils and is the ‘classical’ IgE receptor; and

• the low-affinity FcεRII (CD23), which is expressed on leukocytes and lymphocytes.

Fig. 3.19 Models for FcεRI and FcεRII

Models for the high-affinity IgE receptor (FcεRI), which binds to IgE via its α chain, and the low-affinity receptor (FcεRII), which binds using its lectin domains. Both receptors are shown with IgE bound in the activation and release of histamine and other vasoactive and inflammatory mediators.

The α chain of FcεRI is a glycoprotein and has two extracellular domains homologous to immunoglobulin domains and is a member of the immunoglobulin superfamily.

The low-affinity FcεRII is not a member of the immunoglobulin superfamily, but has substantial structural homology with several animal C-type lectins (e.g. mannose-binding lectin [MBL]).

Cross-linking of IgE bound to FcεRI results in histamine release

The high-affinity receptor (FcεRI) is present on the surface of mast cells and basophils as a complex with a β (33 kDa) and two γ (99 kDa) chains to form the αβγ₂ receptor unit (see Fig. 3.19).

FcεRI binds IgE with an affinity of approximately 10¹⁰ L/mol such that, although the serum concentration of IgE is very low, the receptors are permanently saturated. Cross-linking of the IgE bound to these receptors results in the activation and release of histamine and other vasoactive mediators.

FcεRII is a type 2 transmembrane molecule

FcεRII is the low-affinity (CD23) receptor and is a type 2 transmembrane molecule (i.e. one in which the C termini of the polypeptides are extracellular, see Fig. 3.19). The two forms of human CD23 are:

• CD23a, which is expressed in antigen-activated B cells and influences IgE production;

• CD23b, expression of which is induced in a wide range of cells by IL-4.

CD23a and CD23b differ by six or seven amino acids in their cytoplasmic N termini and contain different signaling motifs that modify their functions.

IgE receptors bind to IgE by different mechanisms

Decades of research and controversy surround the identification of the interaction site on IgE for the high-affinity FcεRI receptor.

Recent crystal structures of a Cε2Cε3Cε4 fragment and Cε3Cε4–FcεRI complex seem to have resolved the issue, and there is a striking structural homology between this site and the site on IgG Fc that binds FcγRIII. One distinctive feature is that, unlike IgG binding to FcγR, binding of IgE to its FcεRI receptor does not appear to depend on glycosylation of the immunoglobulin; however, binding to FcεRII does because it is a lectin. These studies should be interpreted with caution because the antibody fragments were not produced in mammalian cells.

Possible models for the interaction of IgE with the FcεRI and FcεRII receptors are illustrated in Figure 3.19. The low-affinity IgE Fc receptor, FcεRII (CD23), is a C-type lectin and is therefore sensitive to glycosylation status.

Development of the antibody repertoire by gene recombination

How can a finite genome provide the information required for the vast repertoire of antibody molecules that an individual can synthesize? This question posed an intellectual and practical challenge for decades.

Ehrlich’s side-chain hypothesis, put forward at the beginning of the 20th century (Fig. 3.20), proposed antigen-induced selection. His model is close to our present view of clonal selection, except that he placed receptors of several different specificities on the same cell and did not address the question of how the diverse receptors were generated.

Fig. 3.20 Ehrlich’s side-chain theory

Ehrlich proposed that the combination of antigen with a preformed B cell receptor (now known to be antibody) triggered the cell to produce and secrete more of those receptors. Although the diagram indicates that he thought a single cell could produce antibodies to bind more than one type of antigen, it is evident that he anticipated both the clonal selection theory and the idea that the immune system could generate receptors before contact with antigen.

Later it was shown that light chains were comprised of a variable N-terminal half and a constant C-terminal region. This led Dreyer and Bennett to propose the ‘two genes one polypeptide chain’ hypothesis, i.e. the light chain sequence is encoded by at least two different genes, contrary to a central dogma of molecular biology at the time (‘one gene, one protein’).

Shortly thereafter, Tonegawa proved the Dreyer and Bennett hypothesis by demonstrating that the DNA encoding the variable and constant region were located at distant sites, on the same chromosome. Subsequently it was shown that light and heavy variable regions were comprised of the product of two (VL, JL) or three (VH, DH, and JH) separate gene libraries of DNA sequences, respectively, which are rearranged into a complete antigen receptor gene via recombination events (termed ‘V(D)J recombination’) that occur at specific times during B cell differentiation. Combinatorial assembly of antigen receptor genes from these libraries of gene segments, described in more detail below, allows a limited amount of genetic material to generate the vast antigen receptor repertoire required for humoral immune protection. A similar process is used to assemble the genes encoding the antigen binding domains of T cell receptors.

Q. Given that one human can produce more heavy chains and more light chains than there are genes in the entire human genome, what mechanism(s) could lead to the presence of the great diversity of immunoglobulin genes seen in different B cells, each encoding a specific antibody?

A. The antibody genes could be produced by splicing together different gene segments, with a different combination of segments used in different B cells. Another mechanism could involve mutation of the original gene(s) with different mutations occurring in different B cells. In fact, both of these mechanisms take place in human B cells.

The two mechanisms occurring within individual B cells – gene segment recombination and gene mutation – described above, are termed somatic recombination and somatic mutation to distinguish them from the related processes that occur in germ cells.

In addition to the diversity generated by somatic recombination and somatic mutation, the pairing of a unique VL domain with a unique VH domain generates a unique antigen-binding site. This is referred to as combinatorial pairing.

An exception to this rule is present in the Camelidae family in which a high proportion of circulating antibody is comprised of a functional heavy chain only. This suggests possibilities for using just VH domains as therapeutics that may have advantages for penetrating solid tumors.

Heavy chain gene recombination precedes light chain recombination

The revealed organization of mammalian genomes invalidated the earlier dogma of ‘one gene, one polypeptide chain’ and replaced it with ‘genes in pieces’ as coding (exons) and non-coding (introns) DNA sequences were identified. Generally:

• an open reading frame is transcribed to nuclear RNA;

• the introns are ‘spliced’ out of nuclear RNA to yield a continuous messenger RNA (mRNA);

• protein is translated from continuous mRNA.

Germline DNA encoding immunoglobulin polypeptide chains shows a further level of complexity. Information for the variable domains is present in two or three libraries containing multiple alternative versions of V, D, and J gene segments. These families of gene segments are widely separated from exons encoding the constant regions. Recombination events during B cell differentiation choose a particular V, D, and J gene segment from each library and assemble them into a continuous DNA sequence encoding the V domain. However, in the initial nuclear RNA transcript the information for the V and C domains is still widely separated.

The DNA encoding the leader peptide to the end of the C gene, including introns, is transcribed into heterogeneous nuclear RNA (hnRNA). The hnRNA is processed with the ‘splicing out’ of introns to yield the mRNA encoding the polypeptide V and C domains within a continuous RNA sequence that is translated into protein.

The primary antibody repertoire is:

• generated through recombination of germline gene segments, with selection to eliminate B cells producing anti-self antibodies;

• present in the B cell population and expressed as membrane-bound IgM and IgD acting as antigen receptors.

Antibody genes undergo rearrangement during B cell development

Antibody gene rearrangement is initiated at the heavy chain locus. If a ‘productive’ rearrangement occurs (i.e. generating a functional gene product), then rearrangement at the κ locus is initiated:

• if the κ rearrangement is productive, the assembled antibody is expressed as a membrane receptor;

• if recombination at the κ locus is not productive, rearrangement at the λ locus proceeds.

A number of pseudogenes (non-functional genes with high homology to the heavy chain genes) and orphan genes (putative genes with no known homology to known functional genes) have been identified as related to functional heavy and light chain genes. For example, the heavy chain locus includes approximately 78 related pseudogenes, and orphan genes occur on chromosomes other than 14.

Rearrangement at the heavy chain VH locus precedes rearrangement at light chain loci

The germline human heavy chain locus, on chromosome 14 (Fig. 3.21), contains a library of 38 to 46 functional VH gene segments that encode the N-terminal 95 residues of the VH region.

Fig. 3.21 Heavy chain VDJ recombination in humans

The heavy chain gene loci recombine three segments to produce a VDJ gene, which encodes the VH domain. Of some 80 V genes, about 50 are functional and the others are pseudogenes. The V gene segment recombines with one of 23 DH segments and one of six JH segments to produce a functional VDJ gene in the B cell.

The C terminal residues of the VH region are encoded within 23 DH and six JH gene segments (see Fig. 3.21).

Productive rearrangement at the VH locus is an obligatory and early event in the generation of B cells that precedes rearrangement at light chain loci.

The first event is recombination between a JH gene segment and DH segments, followed by recombination with a VH gene segment

DH segments:

• are highly variable, both in the number of codons and in the nucleotide sequence;

• may be read in three possible reading frames without generating stop codons; and

• can be used singly or in combinations.

Productive recombination between DH and JH gene segments signals recombination of this DJ sequence to a VH gene segment, forming a contiguous DNA sequence encoding the entire VH protein sequence (see Fig. 3.21).

The recombined VH, DH, and JH gene segments generate widely diverse hypervariable HV3 (CDR3) sequences, which contribute greatly to the diversity of the primary antibody repertoire.

Rearrangement results in a VK gene segment becoming contiguous with a JK gene segment

The germline human κ light chain locus on chromosome 2 (Fig. 3.22) contains a library of 31–35 functional Vκ gene segments that encode the N terminal 95 residues of the Vκ region.

Fig. 3.22 κ chain production in humans

During differentiation of the pre-B cell, one of several Vκ genes on the germline DNA (V1–Vn) is recombined and apposed to a Jκ segment (Jκ1–Jκ5). The B cell transcribes a segment of DNA into a primary RNA transcript that contains a long intervening sequence of additional J segments and introns. This transcript is processed into mRNA by splicing the exons together, and is translated by ribosomes into κ chains. The rearrangement illustrated is only one of the many possible recombinations. (B cell DNA is colored light brown; RNA is colored green; and immunoglobulin peptides are colored yellow.)

The C terminal residues of the Vκ region are encoded within five Jκ gene segments (see Fig. 3.22).

During B cell development, rearrangement of the DNA occurs such that one of the Vκ genes becomes contiguous with one of five Jκ genes.

Q. If there are 31 Vκ segments and five Jκ segments, what is the theoretical maximum number of rearrangements that could occur at this locus?

A. 155 (i.e. 31 × 5), but imprecise joining introduces additional diversity (see below).

The κ locus also includes over 30 related pseudogenes, and orphan genes are present on other chromosomes.

A leader or signal sequence (a short hydrophobic segment responsible for targeting the chain to the endoplasmic reticulum) precedes each Vκ segment. The leader sequence is cleaved in the endoplasmic reticulum, and the antibody molecule is then processed through the intracellular secretory pathway.

Recombination results in a Vλ gene segment becoming contiguous with a functional Jλ gene segment

The germline human λ light chain locus (Fig. 3.23) on chromosome 22 contains a library of 29–33 functional Vλ gene segments that encode the N terminal 95 residues of the Vλ region.