Identification of a small peptide that inhibits the phosphorylation of ErbB2 and proliferation of ErbB2 overexpressing breast cancer cells.

ErbB2 is overexpressed in approximately 30% of breast cancer patients with a correlation to poor prognosis. ErbB2 has been identified as a useful receptor for molecular targeting. A cyclic 20 amino acid phage display random peptide library was constructed using the fUSE5 gene III system. The library was panned against 2 different purified forms of the external domain of ErbB2. This resulted in the identification of several ErbB2-binding phage clones with variable binding to different ErbB2 preparations. One clone (EC-1) bound all preparations of ErbB2 including live cells and fresh frozen human breast cancer specimens. The synthetic peptide based on the deduced sequence of the EC-1 clone and its biotin-conjugated form retained binding affinity for purified ErbB2 and ErbB2 overexpressing cell lysates. EC-1 peptide was able to effectively inhibit the phosphorylation of ErbB2 on residues Y1248 and Y877 in a dose- and time-dependent manner. Furthermore, EC-1 peptide selectively inhibits the proliferation of ErbB2 overexpressing breast cancer cells. The linear portion of the cyclic EC-1 peptide was shown to be essential for binding ErbB2. In addition, 4 biased phage libraries were constructed allowing 4 different regions of the EC-1 peptide to have random sequence. Screening these EC-1 biased libraries did not result in higher affinity peptides but did demonstrate the importance of amino acids at position 1-4 on the N-terminal flanking arm and 11-15 within the cyclic ring. Interestingly, EC-1 contains homologous motifs with known ErbB receptor family ligands. We have identified a small peptide that binds to the extracellular domain of ErbB2, inhibits ErbB2 autophosphorylation and inhibits the proliferation of ErbB2 overexpressing cells. This supports the notion that small peptides can bind to targets important in cancer therapy even if a target does not have a natural ligand. Continuing research with this peptide includes increasing its affinity to ErbB2, evaluation of pharmacokinetics and evaluation of anti-proliferative effects with conjugate anti-cancer agents.

Phage-displayed random peptide libraries in mice: toxicity after serial panning.

PURPOSE: In vivo screening of phage-displayed random peptide libraries (RPLs) has been used to identify peptide ligands to targets found on endothelial cells of blood vessels supplying specific tissues such as brain, kidney, and tumor tissue. Peptides that bind specifically to blood vessels supplying tumor tissue have been conjugated to cytotoxic agents and used to successfully eradicate tumors in a mouse model. With the ultimate goal of developing similar methods for treating human cancer, we describe an in vivo RPL screening process that, unlike previous in vivo experiments, does not harm the animal being screened. METHODS: RPLs were administered to FVB, BalbC, and tumor-bearing MRL/MpJ-fas(LPR) mice in a variety of dosing formats. Tumor nodules were excised 10 min following infusion and phage were amplified from the specimens. Phage were reinjected into the same animal within 48 h. This process was repeated twice for a total of three in vivo screens of mouse tumor tissue within the same animal. Mice were observed for systemic side effects, histopathologic damage, and presence of phage in organs. Peptide sequences were determined from several third-pan phage clones. RESULTS: Overall there was minimal toxicity from administration of single or repeat doses of RPLs. Amino acid consensus sequences were identified and some of the sequences were similar to those of peptide ligands that bind matrix metalloproteinases. CONCLUSIONS: Serial administration of an RPL is well tolerated and serial panning in individual mice leading to consensus sequence motifs is possible. Based on these preclinical data the Food and Drug Administration has approved the implementation of human clinical trials with this technique.