Production of humanized monoclonal antibodies for in vivo imaging and therapy.

Introduction Human immunoglobulins as intravenous immune globulin (IVIG) have important uses in immunodeficiency disorders and hyperimmune sera are used to a small extent for prophylaxis of severely immunocompromised patients against hospital-acquired infections [ 13 and neonatal infections. Such sera, however, have to be derived from volunteers who have high titres of effective antibodies acquired through either natural infection or vaccination. This is clearly limiting and, coupled with the cost of collection and problem of cross-contamination with hepatitis B and human immunodeficiency virus (HIV), severely limits the use of human materials. Rodent monoclonal antibodies offer a plentiful and pure source of effective antibody, and during the 1980s they were evaluated for in vivo human use, particularly aimed at imaging (tumours, blood clots) and therapy (tumours, blood clots, immunosuppression) [2]. While the data have shown that rodent monoclonals can indeed target and deliver an imaging signal or a therapeutic activity, they have not proved clinically successful except in a very few specialized cases [2-41. Clinical use has been limited by (a) the short half-life, (b) the poor recognition of human effector functions by murine immunoglobulin constant regions and (c) human antimouse response (HAMA). Genetic and protein engineering techniques have, however, now provided the tools to overcome these limitations by allowing the ‘humanization’ of antibodies while still retaining the same binding affinity and specificity of the original rodent monoclonal [ 5 ] . Antibody structure An important facet is that the immunoglobulin molecule consists of an assembly of discrete domains such that the structure of the antigen binding site (variable domains V,, and V,) is essentially the same in an intact immunoglobulin as in Fab or Fv fragments. The variable region consists of a framework of anti-parallel P-pleated sheets from which emanate six loops (three loops from the heavy chain and three loops from the light chain). These loops display high variability in amino acid sequence from one antibody to another compared with the P-pleated sheet framework regions. The different conformations of loops and their interactions provide an abundance of surfaces with binding domains for antigen, and hence they are commonly termed the complementarity-determining regions (CDRs). It is clear from structural analyses and threedimensional computer graphic analyses that a CDR must assume the correct conformation to bind antigen, and the correct conformation is dependent upon interaction between specific residues within the CDR and within the framework. The humanization of a rodent antibody involves the transfer of the CDRs from the rodent antibody to a human framework (Figure 1). The prime aim of such humanization is a significant reduction in, if not elimination of, the remaining HAMA response, particularly that part of the response that interferes with antigen-antibody binding domains, since it is this that renders repeated dosing ineffective.