Engineered antibodies as pharmacological tools

Abstract

The ability to engineer the antibody molecule has now progressed to the stage where one can realistically contemplate creating a pharmacologically useful targeting molecule comprising a single-chain, minimal antibody combining site together with a domain that imparts a second functionality. It will certainly be possible in the near future to adjust the fine specificity and affinity of the antibody domain, either by in vitro selection methods or by site-directed mutagenesis based on structural criteria. The functional domain could be an enzyme, a toxin, or any other protein that suits a pharmacological purpose. In this review I have illustrated these directions primarily by summarizing work from my laboratory and those of my collaborators. The review covers single point mutations in the heavy chains of digoxin-specific antibodies that either strikingly diminish or strikingly enhance recognition of a small feature of the antigen, the position 12 OH moiety in the steroid nucleus. The principles for constructing a minimal, single-chain antigen-binding domain based on one of the digoxin-specific antibodies are also outlined, as are the principles for incorporating such domains into fusion proteins. Finally, as a practical application of antibody-targeted enzymes, the construction and evaluation of an anti-fibrin antibody-single-chain urokinase fusion protein is examined. This protein has enhanced potency and specificity both in vivo and in vitro, and it offers the promise of increased therapeutic efficacy as well as diminished toxicity.