Engineered vascular-targeting antibody-interferon-c fusion protein for cancer therapy Christina Ebbinghaus 1 , Roberto Ronca 1 , Manuela Kaspar 1 , Dragan Grabulovski 1 , Alexander Berndt 2 , Hartwig Kosmehl 3 , Luciano Zardi 4 and Dario Neri 1 * 1 Department of Chemistry and Applied Biosciences, Swiss Federal Institute of Technology, Zu¨rich, Switzerland 2 Institute of Pathology, Friedrich Schiller University, Jena, Germany 3 Institute of Pathology, Helios Klinikum Erfurt, Erfurt, Germany 4 Department of Experimental and Clinical Immunology, Istituto Giannina Gaslini, Advanced Biotechnology Center, Genoa, Italy A number of cytokines are either approved drugs or are in advanced clinical trials, yet these biopharmaceuticals do not typi- cally localize efficiently in solid tumors and manifest their thera- peutic potential at the expense of severe side effects. The targeted delivery of cytokines to solid tumors is a promising avenue for increasing the therapeutic index of these biopharmaceuticals. We engineered a fusion protein between scFv(L19), a human antibody fragment specific to the EDB domain of fibronectin, and a cystei- ne-free mutant of murine interferon-g. The resulting fusion pro- tein was capable of targeting new blood vessels in solid tumors, and the targeting efficiency was strikingly increased in tumor- bearing knockout mice lacking the interferon-g receptor. ScFv(L19)- interferon-g displayed a strong antitumor effect in both subcuta- neous and metastatic murine F9 teratocarcinomas, but was not efficacious as single agent when used to treat C51 and CT26 tumors. The potency of this fusion protein could be substantially enhanced by combination with doxorubicin and other immuno- cytokines. These findings are of clinical relevance, as the EDB domain is a marker of angiogenesis, with identical sequence in mouse and man, which is abundantly expressed in a variety of aggressive solid tumors but is undetectable in most normal tis- sues. Supplementary material for this article can be found on the International Journal of Cancer website at http://www.interscience. wiley.com/jpages/0020-7136/suppmat/index.html. ' 2005 Wiley-Liss, Inc. Key words: tumor targeting; tumor therapy; scFv antibody fragment; immunocytokine; interferon-g; mice A quarter of a century after their discovery, 1 monoclonal antibod- ies have become the most rapidly expanding class of pharmaceuti- cals for the treatment of many human diseases, including cancer. Since 1997, with the FDA approval of the first anticancer therapeu- tic antibody, a chimeric anti-CD20 antibody for the treatment of non-Hodgkin’s lymphoma 2 (Rituxan), several antibodies have been approved for the treatment of cancer. The last FDA-approved anti- bodies in February 2004 are Erbitux and Avastin, the latter being the first approved therapy designed to inhibit angiogenesis. Despite these promising developments, the efficacy of the approved clinical anticancer antibodies needs to be enhanced since cure is rare. 3,4 One avenue for the improvement of the therapeutic performance of anticancer antibodies relies on the engineering of IgG binding affinity to Fc receptors, either by mutation or by gly- cosylation engineering. 5–7 Another avenue, which represents the focus of our laboratory, consists in the development of antibody derivatives, which can deliver a bioactive agent to the tumor envi- ronment (e.g., drugs, toxins, radionuclides, photosensitizers, pro- coagulant factors, cytokines). The targeted delivery of anticancer cytokines to the tumor envi- ronment appears to be a particularly promising strategy. A number of cytokines are either approved drugs or are in advanced clinical trials, yet these biopharmaceuticals do not typically localize effi- ciently in solid tumors and manifest their therapeutic potential at the expense of severe side effects. As cancer is often a disseminated disease; it appears that the most straightforward avenue for the targeted delivery of cytokines to tumors should rely on the fusion of cytokines to a suitable bind- ing molecule (in most cases, an antibody or an antibody fragment). These so-called immunocytokines are designed to create a prefer- ential accumulation of the cytokines at the tumor site, thus improving the therapeutic index of these biopharmaceuticals. 8–10 In turn, this may lead to a more efficient activation of a specific antitumor response or allow the administration of lower doses and limit toxicities to normal organs. In the past few years, our laboratory has focused on the devel- opment of immunocytokines specific to markers of tissue remodel- ing and angiogenesis. 11 The rationale of this approach relies on the observation that abundant, stable and accessible markers of tumor neovasculature may facilitate the targeted delivery of thera- peutic agents to solid tumors, overcoming problems associated with the high interstitial pressure in perivascular tumor struc- tures. 12 One of the best-characterized markers of angiogenesis known so far is the EDB domain of fibronectin. EDB is expressed in the extracellular matrix around newly formed blood vessels in the majority of aggressive solid tumors, but is undetectable in nor- mal vessels and tissues. 13–15 In our laboratories, we have generated a number of good-quality anti-EDB antibodies using human anti- body libraries. 16–18 In particular, the human antibody fragment scFv(L19) 19–23 has been shown to target both tumor and nontumor angiogenesis in animal models. Recently, primary tumors and meta- static lesions in cancer patients were successfully localized through immunoscintigraphy using 123 I-labeled dimeric scFv(L19). 24 The ability of scFv(L19) to target tumors in patients provides the foun- dations for new therapeutic applications in which the L19 antibody is engineered to deliver bioactive molecules selectively. The EDB domain of fibronectin is identical in sequence in mouse and man, thus facilitating investigations in immunocompetent rodent models of cancer. We have previously reported the anticancer properties of scFv(L19) fused to interleukin 2 (IL2), 25 to IL12 26 and to TNF- a, 27 as well as the combination of L19-IL12 with L19-TNF-a 28 and of L19-TNF-a with cytotoxic agents. 27 These fusion proteins were shown to retain both antibody and cytokine functions and displayed superior anticancer activities compared to equivalent amounts of free cytokines (and antibody). IFN-g is a pleiotropic cytokine that plays a central role in pro- moting innate and adaptive mechanisms of host defense. 29,30 IFN-g is able to inhibit tumor cell growth directly, to generate an Dr. Ronca’s current address is: Department of Biomedical Sciences and Biotechnology, University of Brescia, Italy. Grant sponsor: the Swiss National Science Foundation; Grant number: 31-64912; Grant sponsor: the Krebsforschung Schweiz; Grant sponsor: the Bundesamt fu ¨r Bildung und Wissenschaft/EU (Therapeutic Antibodies and STROMA Projects). *Correspondence to: Institute of Pharmaceutical Sciences, Department of Chemistry and Applied Biosciences, Swiss Federal Institute of Technol- ogy Zu ¨rich, Wolfgang-Pauli-Strasse 10, ETH Ho ¨nggerberg, HCI G396, CH-8093 Zu ¨ rich, Switzerland. Fax: þ41-44-633-13-58. E-mail: neri@pharma.ethz.ch Received 24 October 2004; Accepted after revision 14 December 2004 DOI 10.1002/ijc.20952 Published online 30 March 2005 in Wiley InterScience (www.interscience. wiley.com). Int. J. Cancer: 116, 304–313 (2005) ' 2005 Wiley-Liss, Inc. Publication of the International Union Against Cancer