Disregulated inflammation underlies or exacerbates the pathology of many human diseases, and an effective anti-inflammatory strategy involves stoichiometric antagonism of one or more components of the large network of pro-inflammatory receptor-ligand interactions. Monoclonal antibodies (mAb) have proven effective in the clinic against inflammatory targets. However, stoichiometric activity has often limited their effectiveness by necessitating the use of large doses to achieve adequate efficacy. Also, the potential immunogenicity of unnaturally high titers of single anti-self idiotypes often leads to a loss of efficacy. Such limitations could be overcome if the catalytic power of enzymes could be harnessed for the treatment of inflammatory disease. Custom design of enzymatic activities for desired applications has long been a major goal of protein engineering. However, current computational methods for the rational alteration of enzyme structures to enable desired reactions still leave prohibitively large amounts of protein sequence space to be searched for the required mutations. As a result, progress has been hampered by a lack of robust systems capable of efficient identification and recovery of rare variants possessing the sought activity from large libraries. To facilitate the engineering of target-specific proteolytic enzymes, a proprietary proteolytic activity sensor has been developed, which when co-expressed in cells with mutagenic libraries of a proteolytic scaffold, confers a selectable phenotype on cells expressing target-cleaving variants. The chymotrypsin family of serine proteases includes many of the most intensively studied of all enzymes. Thus, the chymotrypsin family structure/activity knowledge base permits the mutagenic targeting of residues lining the substrate binding sites to create a broad range of substrate specificities that can be efficiently searched with the aforementioned proteolytic activity biosensor. The goal of the proposed program is two-fold: (1) validate and optimize the proteolytic activity biosensor and protease library design and construction for selection and activity maturation of a human chymotrypsin-like protease against a human inflammatory target with a preliminary kcat/Km value of at least 102 M-1sec-1, and (2) test the protease for bioactivity in vitro. Human TNF1 will be used as a model target for development of the system because it is clinically proven for the treatment of inflammatory diseases, against which TNF1-neutralizing proteases can be tested and compared with approved anti-TNF1 antibodies and other stoichiometric therapeutics. The sensor will incorporate full-length human TNF1 for selections from a mutagenic library based on the proteolytic domain of HtrA1, a human chymotrypsin family protease with unique properties for the engineering of target-specific activities. The proposed program is expected to take 16 months, at the end of which we hope to have established (1) optimum selection conditions, (2) some indication of the range of proteolytic activities achievable with this system, and (3) how optimized target-cleaving proteases compare with high-affinity antibodies against the same target with respect to target neutralization in vitro.
The proposed project aims to develop a new system to solve a long-standing problem in protein engineering, namely to harness the catalytic power of enzymes for human therapy by engineering target-specific proteolytic activities into existing human protease scaffolds, which promises to provide improved efficacy and safety for the treatment of a broad range of diseases such as cancer, cardiovascular, inflammatory, and infectious diseases.