Engineering scFvs for Improved Stability

The high binding specificity and affinity of antibody molecules is a biological wonder. Because of these properties they have been used as prophylactic, diagnostic, and analytical reagents for almost a century now. Their use as “magic bullets” for targeted therapy was conceived by Ehrlich in the late 19th century. With the advent first of the hybridoma technology (1 ) and then of recombinant DNA technology dawned the era of antibody-based therapies. Not long after the beginning of this new field of medicine, the hurdles associated with the pharmakokinetic properties of antibodies became clear. It was realized that the large size of the antibody molecules precluded their proper penetration into masses of biological tissues such as solid tumors (2 ). It was also realized that additional effector functions can be hooked on to antibody molecules. There was the urge to widen the spectrum of effector functions of antibodies by substituting the constant domains with enzymes (to make prodrug), toxins (immunotoxins) and radionucleides (3 ). These requirements led to the development of the recombinant Fab and the Fv technology. The Fv technology greatly aided in solving not only the problems associated with the large size of the antibody but also in reducing the immunogenecity of mouse monoclonal antibodies (MAbs) in humans. However, it had its own drawbacks, one of which was the loss of stability of many Fvs as compared to the corresponding IgGs (4 ). This is a serious drawback in view of the fact that for a molecule to be therapeutically successful it has to be stable at 37�C. It is therefore conceivable that some Fvs that have the potential for therapeutic use because of their antigen binding specificity can face the severe limitations arising out of their inherent instability. It is generally believed that the hydrophobic interaction between the V_H and V_L strongly contribute to the stability of a Fv. But in many Fvs this interaction is not strong enough. Because of this Fvs are modified into scFvs where a flexible peptide linker, typically a (Gly4Ser)₃ repeat is used to hold the two V domains together (5 ). With the application of the scFv technology to many different Fvs, it was realized that although the linker was effective in covalently holding the two V-domains together, it was not always effective in keeping a scFv active. In other words the chains could still unfold and lead to aggregation. This had led to the hypothesis that increasing the interaction between the two V-domains may be important to enhancing the stability of a biologically active Fv (being capable of remaining as a monomer and bind antigen). Several approaches trying to increase the stability of Fvs have been described. One of the earlier approaches involved engineering disulphide bonds into the framework regions of Fvs such that they are covalently held together (6 ). Other technologies involve either decreasing hydrophobic patches on scFvs to minimize aggregation (7 ,8 ) and to increase the strength of non-covalent interaction between the V domains (8 ). This involves identifying the potentially “problematic” residues that either reduce the interdomain interaction and/or causes aggregation. This chapter is meant to help the reader in identifying the potentially problematic residues and in making a judgement call on substitution(s) that is (are) likely to stabilize the Fv. The discussion that follows and the strategy that is presented is largely based on information that was gathered from the Kabat database (Kabat database maintained by Johnson and Wu, http://immuno.bme.nwu.edu ). Similar types of information is also available in the Martin database (http://www.biochem.ucl.ac.uk/~martin/abs/abs.info ). But what follows below is a simple way to help molecular biologists to identify with reasonable accuracy in the absence of crystal structures amino acids in Fvs that can be mutated to other residues in order to make the Fv more stable. In the discussion to follow we will cite as an example the scFv derived from a MAb called K1. K1 scFv was very unstable, with a half life of only 4 h at 37�C. By following the strategy discussed here, K1 scFv was found to remain fully stable even after 48 h of incubation at 37�C.