Towards the Design of Matrix Metalloproteinases (MMP) Antibody Sequences

Matrix metalloproteinases (MMPs) are a family of enzymes dependent on metal ion cofactors that cleave the proteins found in the extracellular matrix and proteins involved in signalling processes in humans. These proteins play many important roles in normal physiology and development, but under pathological conditions, disregulated MMP activity can lead to cancer progression, undesired inflammation, and numerous other conditions [1]. Biological inhibitors of MMPs in the form of antibodies offer an appealing method for disrupting MMP enzyme activity, as antibodies are capable of potently, specifically inhibiting individual MMPs [2] in ways that traditional small molecules cannot [3]. The key challenge in establishing antibody-based inhibition as a general strategy for interfering with the pathologic functions of MMPs is the lack of fundamental information relating the amino acid sequences of inhibitory antibodies to their unique properties as inhibitors. In this poster, we present a method for analyzing protein sequences to uncover salient features of sequences of antibodies capable of binding to and inhibiting MMPs. The method consists of the following steps. First, we number antibody sequences [4] and divide them into frameworks [6], scaffolds that shape the antibodies' overall structures, and complementarity determining regions (CDRs) [5], peptide loops that directly contact the target MMP or other biological target. Second, we extract frameworks and CDRs [7] from the sequences. Third, we identify the most distinguishing features of antibody utilizing feature selection and machine learning algorithms. We expect to find distinct features of antibodies that bind or inhibit MMPs with this method. Distinct features can be further experimentally evaluated by altering the features unique to MMP-binding or MMP-inhibiting antibodies to assess their importance.