(626c) Rational Methods for Optimizing Antibody Specificity By Controlling the Net Charge of the Complementarity-Determining Regions

The intense interest in using antibodies as therapeutics is largely due to their attractive properties, including their high affinity, specificity, potency, stability, manufacturability and low toxicity. Nevertheless, the affinities of lead antibodies identified via in vitro or in vivo methods are typically not high enough for therapeutic applications and must be further enhanced via in vitro affinity maturation. A key challenge during in vitro affinity maturation is that gains in antibody affinity can come at the cost of reduced specificity. To understand the sequence determinants of antibody specificity, we selected a large number of antibody variants with common frameworks using multiple in vitro antibody libraries and different selection methods that enable identification of antibodies with low and high specificity. Sequence-based analysis of these variants reveals that one of the most important determinants of antibody specificity is the average net charge per residue in the complementarity-determining regions (CDRs). Increased negative charge in the CDRs – at least over moderate ranges of CDR net charge explored in this work – is linked to reduced non-specific binding. Conversely, increased positive charge – especially due to arginine mutations – is linked to increased non-specific binding. Interestingly, several other sequence-based properties (fraction of hydrophobic and aromatic residues) are weakly correlated with antibody specificity for the antibodies studied in this work. Moreover, we find that net charge per CDR residue is a useful metric for identifying clinical-stage antibodies with increased risk for non-specific interactions despite that these antibodies have significant sequence differences outside of the CDRs. We are currently using these insights to design antibody libraries in which the net charge of the CDRs is fixed within specific ranges in order to promote high specificity and reduce affinity/specificity trade-offs during affinity maturation. Our long-term goal is to use these and related findings to develop systematic and robust design methods for rapidly generating and optimizing antibodies for use in a range of diagnostic and therapeutic applications.