Table of Contents

Antibody Therapy of Carcinomas and Malignant Lymphomas

Since the early days of Paul Ehrlich who considered antibodies as "magic bullets" immunologists were attracted by the idea to destroy tumor cells with antibody molecules alone or conjugates made thereof. With the advent of monoclonal antibody technology 30 years ago new enthusiasm came up to develop a modern immunotherapy of cancer. This promise was fulfilled, however, only in rare cases namely in some leukemias and malignant lymphomas. The application of unmodified mouse monoclonal antibodies has yielded overall disappointing results with low response rates. One major drawback was for instance the fact that mouse antibodies are not able to activate human effector systems. Furthermore, the vast majority of antibodies used for tumor therapy recognize differentiation antigens that (although at a lower density) are also expressed on normal cells. A break through represented the approval by the American Food and Drug Administration of the chimeric (mouse/human) Rituximab antibody for the treatment of relapsed Non-Hodgkin´s lymphoma. Previously, humanized antibodies Herceptin, Erbitux and Avastin were also approved for therapy of certain types of mammary carcinoma and colon cancer, respectively. Thus, antibody engineering finally has led to a renaissance of antibody therapy, which now is evolving as a promising option to synergize with or even replace conventional chemotherapeutic regimens.

The main objective of our research program is to develop antibody-based strategies for the treatment both of carcinomas and malignant lymphomas. To this end we have constructed several bispecific monoclonal antibodies and recombinant single-chain Fv constructs in order to focus and redirect cytotoxic T cells against various tumor targets. Using bispecific antibodies, cytotoxic T lymphocytes from a patient can be activated and directed to the respective tumor targets inducing their destruction. The specificity of this reaction is solely determined by the bispecific antibody and is independent from the genetically fixed T cell receptor specificity. In essence, hybrid monoclonal antibodies are employed with one binding site being directed against the CD3/TCR complex on T cells and the other one against a tumor-associated antigen as the pan carcinoma marker EpCAM. Together with our colleagues at the University Hospital we have conducted a clinical trial for patients with advanced ovarian cancer and malignant ascites refractory to standard therapy that showed promising results.

In NOD SCID mice, EpCAMxCD3 had a long serum half-life (t1/2 ∼ 7 days). It significantly reduced growth of BxPC-3 pancreatic cancinoma xenografts. For mimicking the pancreatic cancer microenvironment in vitro we developed a three-dimensional tumor reconstruct system, in which lymphocytes were co-cultured with tumor cells and fibroblasts in a collagen matrix. In this in-vivo-like system EpCAMxCD3 potently stimulated production of the effector cytokines IFN-γ and TNF-α by pre-activated lymphocytes. Moreover, EpCAMxCD3 induced the production of TNF-α, IFN-γ and IL-2 by non-stimulated PBMCs more effectively than a bivalent anti-CD3 antibody. Most excitingly, we demonstrate for the first time that EpCAMxCD3 induces prolonged contacts between lymphocytes and tumor cells, which may be the main reason for the observed anti-tumor effects. Besides T lymphocytes there are other potent cytotoxic cells that can be targeted to the tumor site as natural killer cells, macrophages and polymorphonuclear neutrophils. Especially monocytes, macrophages, dendritic cells and activated neutophils carry high affinity Fc receptors (CD64) on their surface that mediate antibody-dependent cellular cytotoxicity and release of several cytokines. To recruit those effectors we succeeded in establishing a bispecific antibody with the specificity EpCAMxCD64. This reagent has underwent preclinical evaluation and proved promising; it will proceed to clinical testing soon.

For treating malignancies derived from B lymphocytes we have obtained hybrid-hybridomas of the specificity CD19xCD3 and CD19xCD5. The CD19 antigen was chosen as target molecule since it represents a pan-B cell antigen, stably expressed on most if not all leukemias and lymphomas of B cell origin. The bispecific antibodies were subjected to an extensive preclinical evaluation in order to demonstrate its efficacy in vitro. Ex vivo expanded cytokine-induced killer cells (CIK cells), targeted to tumors by the addition of CD19xCD3 or CD19xCD5 bispecific antibody could efficiently lyse the respective autologous tumor cells derived from patients with common acute lymphoblastic leukemia (cALL) and B-chronic lymphocytic leukemia (B-CLL).

To overcome human anti-mouse antibody (HAMA) responses together with Frank Breitling we have developed a novel method for the chimerization of mouse monoclonals taking advantage of homologous recombination of immunoglobulin genes. The technique permits the exchange of mouse heavy and light chain constant domains by human IgG1 homologues within the hybridoma cell line. The feasibility of this new approach was demonstrated for the murine hybridoma cell line HEA125, which produces an anti-EpCAM Ab of IgG1 isotype with kappa light chains. In the resulting chimeric mAb chiHEA125 no heavy or light chain constant region protein of murine origin was detectable either by Western blot, ELISA or FACS analysis. Sequencing of chiHEA125 cDNA revealed the exact human constant region sequences spliced together with the murine HEA125 variable domains, as expected after homologous recombination. The chiHEA125 Ab has retained its high affinity for EpCAM, demonstrating the advantage of this in vivo homologous recombination approach. Importantly, preclinical testing has revealed efficient induction of complement-mediated lysis (CDC) and activation of tumor-specific cytotoxicity by NK cells (ADCC) in vitro in a concentration dependent manner. We anticipate that chiHEA125 will develop into a valuable reagent for tumour therapy e.g. in breast and ovarian cancer. Furthermore, our method was successfully applied to the chimerization of an additional murine hybridoma secreting antibodies specific for L1CAM. The cell adhesion molecule L1CAM (CD171) represents a promising therapeutic target for ovarian and pancreatic cancer (collaboration with Peter Altevogt).

One chimerized hybridoma line was further “reconstructed” by homologous recombination in order to flank the expressed variable antibody genes with loxP- and FRT-sites, respectively. This allows for a Cre- (or Flp-) mediated exchange of variable light and heavy genes against a library of different antibody genes cloned from human B lymphocytes. Since each cell of this library displays huge amounts of its specific antibody molecule on the cell surface, individual hybridoma cells carrying the desired antibody can be selected by fluorescence-activated cell sorting using for instance a biotinylated antigen of choice. This hybridoma cell and its progeny secret fully human antibodies. Exploiting the technique we expect to obtain a whole series of potentially therapeutic human antibodies.

In collaboration with: Frank Breitling, DKFZ; Frederik Marmé, Frauenklinik der Universität Heidelberg; Martin Kornacker, Medizinische Klinik V der Universität Heidelberg; Alexander Marmé, Frauenklinik der Universität Tübingen; Antonio Pezzutto, Max-Delbrück-Centrum für Molekulare Medizin, Berlin-Buch; Peter Möller, Pathologisches Institut der Universität Ulm.