Protein-protein interactions tend to feature discontinuous hot-spots, hence finding small molecules that will mimic or disrupt the docking of one protein with another is problematic. This project demonstrates approaches to small molecules that selectively bind a particular receptor tyrosine kinase, TrkA, and to correlate their binding with signaling and pharmacological effects. Nerve growth factor (NGF) is a natural ligand for TrkA. NGF-TrkA interactions are relevant to cancer and neurodegenerative diseases. However, NGF is inappropriate for clinical applications because of cost, proteolytic instability, immune responses, and, notably, its interactions with other receptors (p75, and other Trk's) causing undesirable side effects. Small molecules that selectively bind TrkA are therefore potentially useful as pharmacological and diagnostic probes for disease states. If they give agonistic or antagonists responses, they are also pharmaceutical leads. In the first grant cycle we identified the first small molecules that mimic hot-spots on NGF for its TrkA interaction, and proved they were functional in a variety of in vitro, ex vivo, and in vivo assays. The molecules were not selected using a primary screen that probed for affinities, consequently, the K-d values measured for the leads were ca 1 mu M. Conventional binding assays are not viable for selecting compounds that mimic or disrupt the NGF-TrkA interaction from large combinatorial libraries since solubilized TrkA is difficult and labor intensive to obtain, and labeled NGF is very expensive. We have shown that the affinities of our hot-spot mimics can be increased by combining them in bivalent molecules, and that binding assays are feasible as a primary screen if all the compounds assayed are themselves fluorescently labeled. These screens feature fluorescence activated cell sorting (FACS) of Trk-expressing cells; they do not require isolated Trk receptors or the NGF ligand. Further, we demonstrated a convenient combinatorial method by which monovalent compounds could be dimerized into fluorescently labeled bivalent materials simply by mixing components in aqueous base. This sets the stage for screening large libraries of fluorescently labeled bivalent molecules for binding to TrkA in a primary screen. The original hot-spot mimics took around 8 steps and expensive reagents to prepare, so large libraries of bivalent derivatives formed from these would be hard to produce. This proposal features more synthetically accessible hot-spot mimics that can be made in the quantities required. Innovative ways to join these monomers into fluorescently labeled bivalent molecules with systematically varied linker dimensions are outlined. Libraries of these materials will be tested in a series of biochemical and pharmacological assays to filter out the strongly binding, selective, and biologically potent TrkA ligands. These assays include FACS binding (primary); FACS binding with chimeric receptors, fluorescence polarization, cell survival (secondary); receptor tyrosine phosphorylation, neuritogenic assays, photoaffinity labeling experiments, blood brain barrier permeation, and in vivo studies for anti-cancer and antineurbdegenerative effects (tertiary end points). These assays will reveal how strongly the compounds bind, where some of the cell signals are induced, and their effects in animal models.
Showing the most recent 10 out of 11 publications
