THE USE OF A COMBINATORIAL LIBRARY METHOD TO ISOLATE HUMAN TUMOR-CELLADHESION PEPTIDES

Tumor cell progression is dependent in part on the successful adhesive interactions of the cells with the extracellular matrix. In this stud y, a new approach is described to isolate linear peptide ligand candid ates involved in cellular adhesion. A synthetic combinatorial peptide library based on the 'one-bead-one-peptide' concept was incubated with live human prostate cancer cells for 90 min at 37 degrees C. The pept ide bead coated with a monolayer of cells was then isolated for micros equencing. The DU145 (DU-H) cells were chosen since they have been pre viously characterized as containing elevated levels of a laminin recep tor for cell adhesion, the alpha(6) beta(1) integrin on the cell surfa ce. The use of a function-blocking antibody (GoH3) allows for the dete ction of peptides which are alpha(6)-specific ligand candidates. From two different libraries (linear 9-mer and 11-mer) of a total of 1 500 000 beads, 68 peptide beads containing attached cells were isolated. T hese positive beads were then retested to determine the ability of the GoH3 antibody to block binding of the cells to the peptide beads. The alpha(6) integrin candidate peptide beads(five in total) were recover ed and two of the beads were microsequenced. These two peptides, RU-1 (LNIVS-VNGRHX) and RX-1 (DNRIRLQAKXX), resemble the previously reporte d active peptide sequences (GD-2 and AG-73) from native laminin. The R U-1, RX-1 and AG-73 peptides were tested for their ability to support cell attachment and to bind the cell surface of DU-H prostate carcinom a cells in suspension using fluorescence-activated cell-sorting (FAGS) analysis. Both RU-1 and AG-73 peptides supported cellular attachment within 1 h. In contrast, after 1 h, EHS laminin supported both cellula r attachment and spreading. The RX-1 peptide exhibited only weak bindi ng to the DU-H prostate carcinoma cells. FAGS analysis indicated that AG-73 peptide attached to tumor cell surfaces over a range of concentr ations, whereas the RU-1 peptide showed a homogeneous concentration re quired for attachment. The described strategy for screening a random p eptide library offers three advantages: (i) ligands for conformational ly sensitive receptors of adhesion can be isolated using live cells; ( ii) specific binding can be selected for using function-blocking antib odies; and (iii) peptides supporting adhesion independent of spreading properties can be distinguished. In principle, specific adhesive pept ides without prior knowledge of the sequence could be isolated for any epithelial cell surface receptor for which a function-blocking reagen t is available.