When monoclonal antibodies were first described by Kohler and Milstein in1975 [1] it was immediately recognised that they had great potential as powerful and specific therapeutic agents. In attempts to exploit that potential, large numbers of clinically-relevant monoclonal antibodies have been derived and many of these have been tested in-vivo. However, with a few notable exceptions such as the licensing of OKT3 for use in immunosuppression [2],[3] the results of the great majority of clinical trials have been disappointing, and data on the effectiveness of the antibodies have been collected on a largely empirical basis. Because this initial "suck it and see" approach has been generally unsuccessful, a more considered and rational approach to the application of monoclonal antibodies in therapy is now required.
The design of a therapeutic strategy requires consideration of a number of different factors including short- and long-term side-effects as well as benefits. Particular attention needs to be paid to deciding which therapeutic effect is most desirable and how it is expected that the antibody may produce that effect (reviewed by Waldmann [4]). For many forms of therapy such as in the treatment of malignancies, the requirement is that the abnormal target cells should be destroyed. Similarly, in the cases of immunosuppression to prevent organ graft rejection or to treat autoimmune disease, or of prevention of graft-versus-host disease in bone marrow transplantation one could envisage using antibodies to destroy the normal haemopoietic cells responsible for the damage [4]. Alternatively, in such cases, antibodies may be used to block the activity of cells temporarily, allowing useful cellular functions to return once tolerance has become established [4]. In other examples such as in the use of monoclonal antibodies to provide passive immunity to toxins, drugs and pathogens it is necessary that the antibody activates the appropriate effector mechanisms.
In order to make better predictions of the therapeutic value of monoclonal antibodies it is necessary to understand how different antibodies interact with Fc receptors and complement to activate the cellular and humoral effector systems. Initial studies involved analysing the interactions of rodent monoclonal antibodies with human effector systems. Clearly rat immunoglobulin has evolved to work in-vivo in rats, and mouse immunoglobulin has evolved to work in-vivo in mice, and it should be remembered that the ability of these antibodies to activate human effector mechanisms is a result of the conservation of some of the key features. Conservation of effector functions between species is not absolute however, and this has muddied the interpretation of many of the studies in this area. Thus, traditionally, workers studying complement activation have often used heterologous systems, for example human target cells, rat or mouse monoclonal antibodies and rabbit or guinea pig complement. In such classical studies certain combinations of complement and target cells were found to be very effective with a large range of antibodies. In contrast, when experiments were carried out with human complement and human targets, mimicking more closely the therapeutic scenario, far fewer of the monoclonal antibodies gave good lysis [5]. This is now known to be due to the presence of a number of homologous complement restriction factors. These are cell surface molecules which act to restrict the toxicity of the animals own complement to the animals own cells and they include such molecules as "homologous restriction factor", 'decay accelerating factor, CD59 and 'membrane cofactor protein' (reviewed by Lachmann [6]). Hence, many monoclonal antibodies which appear to be potent mediators of complement lysis when measured in heterologous, in-vitro systems are completely ineffective in-vivo.
Experiments using complement and target cells derived from the same donor revealed that some antigens were consistently poor targets for activation of complement regardless of the isotype of the antibody, whilst others were consistently good targets providing that the antibody had even a modest ability to bind complement [7]. An exceptionally good target for cell lysis is the human antigen CDw52 [8], recognised by the CAMPATH-1 family of antibodies. Experiments with this system will be described later in this Chapter.
The Kohler and Milstein hybridoma technology facilitated the preparation of panels of monoclonal antibodies, mainly of rodent origin, for use in comparative studies [7],[9]. However, the antibodies in such panels differed in terms of their fine specificities and avidities so it was difficult to reach absolute conclusions about their properties. In addition, it has always proved difficult to make large numbers of human antibodies by this route (in most early studies the human immunoglobulins were myeloma proteins). Advances in recombinant DNA technology were necessary before further progress could be made. The cloning of immunoglobulin genes was a particular turning point and the ability to manipulate the genes and then express them in transfected cells, has made a very valuable contribution to the field [10],[11],[12],[13],[14]. Antibody V-genes can be expressed with different constant region genes, permitting production of matched sets of antibodies of different isotypes but identical specificity and single-site affinity, allowing a direct comparison of properties [15],[16],[17]. There is no restriction on the species from which the V-regions and C-regions can be derived, so that antibodies may be chimeric (eg. mouse or rat variable regions with human constant regions) or purely from one species (eg. fully reshaped antibodies as described in greater detail in other chapters of this book).
In this chapter we describe some of the information learned from studying two matched panels of recombinant antibody molecules [15],[18]. In the first panel all of the antibodies have specificity for a hapten molecule, 4-hydroxy-3-nitrophenylacetyl (NP) [15]. This hapten can conveniently be coupled to carriers such as soluble proteins, cell membrane proteins or to membrane soluble lipid anchors. The antibodies can therefore be purified on affinity columns made from the antigen, they can be quantified and compared in enzyme linked immunoadsorption assays and they can be targeted to hapten modified cells. Reviewed elsewhere [19] and this volume (Morrison et al., Chapter 6) are studies using a different matched set of antibodies to the hapten dansyl.
The second matched set of antibodies discussed here [18] is to a human cell surface antigen, CAMPATH-1 (CDw52) which is expressed on human lymphocytes and monocytes [8],[20]. The CAMPATH-1 antigen was first investigated because rat monoclonal antibodies of this specificity were extremely cytotoxic to human lymphocytes in the presence of autologous complement [21]. It has become a very important specificity for a number of clinical investigations [22],[23],[24],[25],[26],[27].
It is hoped that much of the information gained from these two panels of antibodies will be widely applicable and will help in directing the general use of monoclonal antibodies in therapy as well as extending our understanding of the normal functions of antibodies.