Potent and Selective Inhibition of Membrane-Type Serine Protease 1 by Human Single-Chain Antibodies † Jeonghoon Sun, Jaume Pons, ‡ and Charles S. Craik* Department of Pharmaceutical Chemistry, UniVersity of California, San Francisco, 513 Parnassus, San Francisco, California 94143 ReceiVed September 20, 2002; ReVised Manuscript ReceiVed December 1, 2002 ABSTRACT: Specific human antibodies targeting proteases expressed on cancer cells can be valuable reagents for diagnosis, prognosis, and therapy of cancer. To this end, a phage-displayed antibody library was screened against a cancer-associated serine protease, MT-SP1. A protein inhibitor of serine proteases that binds to a defined surface of MT-SP1 was used in an affinity-based washing procedure. Six antibodies were selected on the basis of their ELISA profiles and ability to serve as useful immunological reagents. The apparent K i , indicative of the potency of the antibodies at inhibiting human MT-SP1 activity, ranged from 50 pM to 129 nM. Two of the antibodies had approximately 800-fold and 1500-fold selectivity when tested against the most homologous serine protease family member, mouse MT-SP1, that exhibits 86.6% sequence identity. Surface plasmon resonance was used as an independent means of determining the binding constants of the six antibodies. Association rates were as high as 1.15 × 10 7 s -1 M -1 , and dissociation rates were as low as 3.8 × 10 -4 s -1 . One antibody was shown to detect denatured MT-SP1 with no cross reactivity to other family members in HeLa or PC3 cells. Another antibody recognized the enzyme in human prostate tissue samples for immunohistochemistry analysis. The mode of binding among the six antibodies and the protease was analyzed by competition ELISA using three distinctly different inhibitors that mapped the enzyme surface. These antibodies constitute a new class of highly selective protease inhibitors that can be used to dissect the biological roles of proteolytic enzymes as well as to develop diagnostic and therapeutic reagents. Proteases are involved in all stages of cancer progression including growth, angiogenesis, invasion, migration, and metastasis (1-6). At the crux of cancer pathogenesis is metastasis during which unregulated proteases lead to invasive remodeling of the extracellular matrix (7-10). Therefore, inhibitors targeting these proteases have been developed. Although partially effective as tools to study cancer progression and metastasis, small-molecule inhibitors are plagued by problems with resistance and toxicity, while macromolecular inhibitors can suffer from promiscuity and low activity. Moreover, proteases are widely distributed in nature comprising approximately 2% of the entire human genome (11). This omnipresence of proteases with nearly identical active site elements creates the need to develop a general method for producing potent, highly selective inhibi- tors. Validating a protease as an appropriate therapeutic target for cancer is a challenging task. The recent discovery of the type II transmembrane serine proteases that have been detected at high levels in various cancer cell lines has underscored their potential role in cancer (4, 5, 8). One member of this family, membrane-type serine protease 1 (MT-SP1), 1 has been implicated as a key protease in triggering the plasminogen proteolytic cascade, a central pathway in cancer progression and metastasis. MT-SP1 is expressed in human tissues such as prostate, breast, and ovary (12-15). MT-SP1 is a mosaic protein of 855 amino acids and is composed of a short cytosolic N-terminal region, a transmembrane signal anchor, several protein-protein in- teraction domains, and a C-terminal protease domain. The serine protease domain at the C-terminus (hMT-SP1-P) is strategically positioned on the cell surface to allow for efficient interaction with other proteins, thus making it a potentially recognizable epitope and a direct target for regulation of the protease activity (12). Since MT-SP1 is expressed on the cell surface, where large molecules can easily access the protease domain, human monoclonal antibodies (MAbs) were explored for their application in inhibitor design. As therapeutic reagents, human MAbs can be far less toxic and more selective inhibitors than both small molecule and other macromolecu- lar inhibitors. Furthermore, MAbs can easily be developed † This work was supported by a grant from NIH CA72006 (C.S.C.) and by postdoctoral training grants from the California Breast Cancer Research Program 7-FB-0053 (J.S.) and the U.S. Army Medical Research and Material Command DOD 2001 Prostate Cancer Research Program DAMD17-00-1-0611 (J.S.). * To whom correspondence should be addressed. Mailing address: 513 Parnassus Box 0446, University of California, San Francisco, San Francisco, CA 94143-0446. Telephone number: 415-476-8146. Fax: 415-502-8298. E-mail address: craik@cgl.ucsf.edu. ‡ Current address: Rinat Neuroscience corporation, 3155 Porter Drive, Palo Alto, CA 94304. 1 Abbreviations: 1 MT-SP1, membrane-type serine protease 1; scFv, single chain variable fragment; ELISA, enzyme-linked immunosorbent assay; hMT-SP1-P, protease domain of human MT-SP1; YT, yeast extract:tryptone. 892 Biochemistry 2003, 42, 892-900 10.1021/bi026878f CCC: $25.00 © 2003 American Chemical Society Published on Web 01/07/2003