Bioconjugate Chem. 2007, 18, 912−921
912
Synthesis and Radiolabeling of Selective High-Affinity Ligands Designed to Target Non-Hodgkin’s Lymphoma and Leukemia Saphon Hok,† Arutselvan Natarajan,‡ Rod Balhorn,† Sally J. DeNardo,‡ Gerald L. DeNardo,‡ and Julie Perkins*,† Chemistry, Materials and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, California 94551, and Radiodiagnosis and Therapy Laboratory, University of California, Davis, Sacramento, California 95816. Received October 2, 2006; Revised Manuscript Received January 25, 2007
Selective high-affinity ligands (SHALs) were synthesized as molecular targeting agents for HLA-DR10, a cell surface receptor upregulated on malignant B-cell lymphocytes in non-Hodgkin’s lymphoma and leukemia. SHALs are designed to mimic the affinity and selectivity of Lym-1, an antibody that binds to the β-subunit of HLADR10. To bind selectively to HLA-DR10, SHALs were constructed to bind to two adjacent pockets on the surface of the β-subunit of HLA-DR10 located within an epitope recognized by the Lym-1 antibody. A series of multivalent SHALs with molecular masses of 1500-3000 Da were synthesized using solid/polymer-supported synthesis on chlorotrityl chloride resin in 50-80% yield. To enable their use as radionuclide carriers in mouse studies, SHALs were conjugated to DOTA in a solution-phase reaction with 70-100% yield. 57Co/CoCl2 titrations revealed that 50-60% of the DOTA in the DOTA-conjugated SHALs was available for radiometal chelation. These DOTASHALs were labeled with 111In and used to carry out pharmacokinetic studies in mice. Radiolabeling reactions of DOTA-SHALs, with exactly one DOTA entity per targeting SHAL molecule, yielded products with greater than 90% radiochemical purity and specific activities ranging from 97 to 150 µCi/µg.
INTRODUCTION Antibodies have evolved to be one of nature’s most selective molecular targeting agents. These large proteins bind their target (small and large biological molecules) through two Fab arms, effectively a bivalent system. They typically have affinity constants ranging from 10-5 to 10-12 M for their target and are often used as therapeutic agents (1, 2) among many other uses. The IgG2a antibody, Lym-1, binds to the β-subunit of HLADR10, a surface-bound/trans-membrane protein found on lymphocytes (3-5). HLA-DR10 is more abundant on malignant lymphocytes than their normal precursor cells, and therefore, Lym-1 has been employed as a radionuclide carrier for the treatment of non-Hodgkin’s lymphoma (6-10). While minimizing exposure to healthy cells (11, 12), however, there are some disadvantages to using antibodies for treatment and imaging. As macromolecules, they have long residence time in the circulation, which limits the target contrast that can be achieved in tumor imaging (thus decreasing tumor detection). The long residence time also decreases the therapeutic index (ratio of damage to cancer cells relative to damage to normal cells) that can be obtained. Therefore, the design and realization of an alternative ligand that mimics the binding affinity and specificity of an antibody but with improved pharmacokinetic properties is an important contribution to future imaging and radiotherapeutic applications (13). Although small molecules typically have low to moderate affinity for protein targets, linking two or more small molecules generates ligands that can exhibit a polyvalent effect (14-19). Affinity is a qualitative term commonly used to describe a binding constant. However, avidity refers to the association constant of a polyvalent interaction (19). Establishing the * Julie Perkins, Lawrence Livermore National Laboratory, 7000 East Avenue, California 94551. Tel: (925) 422-7319. FAX: (925) 4223570. E-mail:
[email protected]. † Lawrence Livermore National Laboratory. ‡ University of California, Davis.
magnitude and individual contribution of the thermodynamic parameters of polyvalent interactions is complicated, however; it is generally accepted that multiple interactions or contacts on the surface of protein targets generally increases the avidity of the polyvalent ligand for the target (19). In many cases, the avidity of the polyvalent ligand is greater than the sum of the individual affinities of the component small molecules. Small (molecular weight 6.5 cm, respectively (Table 3). No radioactivity peak was observed at the origin of the CAE, indicating that no aggregation had occurred. The same trend was observed in the TLC assay; the 111In-DOTA-SHALs showed very little migration from the point of spotting (RF ) 0.25-0.3), whereas the 111In-EDTA complex moved further toward the solvent front (RF ) 0.6). Conversely, HPLC analysis showed the 111In-EDTA at 2.53.0 min and the 111In-DOTA-SHALs at 9.5-10 min. This distinct resolution facilitated the HPLC purification of the radiolabeled products. CAE, TLC, and HPLC were also used to determine the stability of the radiolabeled conjugate, and the 111In-DOTA-SHALs were stable up to 72 h at room temperature with radiochemical purity of >85%. Radiolabeled product yields are shown in Table 3 and are similar to those obtained when generating antibody agents, and those needed for practicality, but can be further optimized for large-scale use where 80-90% yield is desirable. If radiochemical yield is greater than 90%, then the elimination of a final purification step can be considered with additional improvements in product yield. In comparison to antibody agents, SHALs have a 1:1 ratio of protein binding sites to DOTA. Generally, radiolabeled monoclonal antibodies are modified to optimally contain an average of two DOTAs in order to preserve immunoreactivity, but can have up to a maximum of four DOTAs per macromol-
Bioconjugate Chem., Vol. 18, No. 3, 2007 919
Figure 2. Live cell ELISA. (a) Buffer blank. (b) 10 µM biotinylated SHAL 8b. (c) 10 µM biotinylated SHAL 9b. (d) 0.5 µg/mL Lym 1. Cells: white, Raji; black, Nu-DHL-1; gray, MCF7; hashed, HBT-3477. Raji and Nu-DHL-1 are malignant lymphoma HLA-DR10 expressing cell lines, MCF7 and HBT-3477 are HLA-DR10-negative human breast epithelial cancer cell lines.
ecule. One DOTA per SHAL with a molecular weight of ∼3000 Da compared to 2 DOTAs per monoclonal antibody with a molecular weight of ∼150 000 Da translates to a potentially higher specific activity (µCi/µg) for SHALs, although µCi/µmol may be less. Considering their relative mass differences of a factor of 50-100, and DOTA per molecule of about 2 in the opposite direction, then SHALs with specific activities 25-50 times greater than those for antibodies is theoretically feasible. Whole-cell binding experiments conducted in vitro with 111Inlabeled SHALs 4b and 7b showed that both SHALs bind Raji HLA-DR10 positive non-Hodgkin’s lymphoma cells. Linking the two molecules dabsyl-L-valine and N-benzoyl-L-arginyl-4aminobenzoic acid together using a single lysine and poly(ethylene glycol) spacer resulted in the generation of a SHAL (4b) that binds to HLA-DR10-containing Raji cells with a Kd of ∼80 nM. Moreover, the bis-version of 4b, SHAL 7b, exhibited an avidity of 93 pM for Raji lymphoma cells (21). Biotinylated derivatives 8b and 9b of the two DOTA-SHALs 4b and 7b (in which biotin was conjugated to each SHAL in place of DOTA) were assessed using a live-cell ELISA, and the results are shown in Figure 2. The biotinylated SHALs were assayed against four cell types, two malignant lymphoma cell lines expressing HLA-DR10 (Raji and Nu-DHL-1) and two human breast epithelial cancer cell lines, known to not express HLA-DR10 (MCF7 and HBT-3477). Clearly, SHAL 9b exhibits a much higher avidity than 8b, again indicating that increased multivalency increases binding interactions. Morover, the SHAL 9b exhibits a similar pattern of specificity to Lym1, more specifically binding HLA-DR10-expressing cell lines. The results presented here demonstrate that it is possible to create small-molecule-based targeting agents with an overall relatively small molecular weight (
